Diffusion strategies have advanced visual motor control by progressively denoising high-dimensional action sequences, providing a promising method for robot manipulation. However, as task complexity increases, the success rate of existing baseline models decreases considerably. Analysis indicates that current diffusion strategies are confronted with two limitations. First, these strategies only rely on short-term observations as conditions. Second, the training objective remains limited to a single denoising loss, which leads to error accumulation and causes grasping deviations. To address these limitations, this paper proposes Foresight-Conditioned Diffusion (ForeDiffusion), by injecting the predicted future view representation into the diffusion process. As a result, the policy is guided to be forward-looking, enabling it to correct trajectory deviations. Following this design, ForeDiffusion employs a dual loss mechanism, combining the traditional denoising loss and the consistency loss of future observations, to achieve the unified optimization. Extensive evaluation on the Adroit suite and the MetaWorld benchmark demonstrates that ForeDiffusion achieves an average success rate of 80% for the overall task, significantly outperforming the existing mainstream diffusion methods by approximately 20% in high difficulty tasks, while maintaining more stable performance across the entire tasks.

The Mamba architecture has been widely applied to various low-level vision tasks due to its exceptional adaptability and strong performance. Although the Mamba architecture has been adopted for spectral reconstruction, it still faces the following two challenges: (1) Single spatial perception limits the ability to fully understand and analyze hyperspectral images; (2) Single-scale feature extraction struggles to capture the complex structures and fine details present in hyperspectral images. To address these issues, we propose a multi-scale, multi-perceptual Mamba architecture for the spectral reconstruction task, called M3SR. Specifically, we design a multi-perceptual fusion block to enhance the ability of the model to comprehensively understand and analyze the input features. By integrating the multi-perceptual fusion block into a U-Net structure, M3SR can effectively extract and fuse global, intermediate, and local features, thereby enabling accurate reconstruction of hyperspectral images at multiple scales. Extensive quantitative and qualitative experiments demonstrate that the proposed M3SR outperforms existing state-of-the-art methods while incurring a lower computational cost.

Reinforcement Fine-tuning (RFT) methods such as Group Relative Policy Optimization (GRPO) have demonstrated strong capabilities in aligning Large Language Models with human preferences. However, these approaches often suffer from limited data efficiency, necessitating extensive on-policy rollouts to maintain competitive performance. We propose PSPO (Prompt-Level Prioritization and Experience-Weighted Smoothing for Efficient Policy Optimization), a lightweight yet effective enhancement to GRPO that improves training stability and sample efficiency through two complementary techniques. First, we introduce an experience-weighted reward smoothing mechanism, which uses exponential moving averages to track group-level reward statistics for each prompt. This enables more stable advantage estimation across training steps without storing entire trajectories, allowing the model to capture historical reward trends in a lightweight and memory-efficient manner. Second, we adopt a prompt-level prioritized sampling strategy, which is an online data selection method inspired by prioritized experience replay. It dynamically emphasizes higher-impact prompts based on their relative advantages, thereby improving data efficiency. Experiments on multiple mathematical reasoning benchmarks and models show that PSPO achieves comparable or better accuracy than GRPO, while significantly accelerating convergence, and maintaining low computational and memory overhead.

As an essential component of the Industrial Internet of Things (IIoT), lights-out manufacturing (LoM) relies heavily on the rapid detection of anomalies. However, LoM anomalies often arise from complex, cross-modal correlations, and traditional detection models struggle with the dual challenges of limited local data and stringent data privacy requirements, leading to poor generalization. To address these challenges, this paper introduces a novel Federated Graph Process Distillation framework (Fed-GPD) for multi-modal anomaly detection. Our approach first represents heterogeneous industrial data as unified graph structures to effectively model the underlying device relationships. We then propose a new graph knowledge distillation paradigm designed for Graph Neural Networks (GNNs) in a federated setting. Instead of merely distilling final predictions, we introduce two novel distillation mechanisms: 1) Neighborhood Aggregation Process Distillation (NAPD), which transfers the knowledge of how a model processes local neighborhood information at each GNN layer, and 2) Relational Knowledge Matrix Distillation (RKMD), which aligns the global understanding of node-to-node relationships learned by the models. These mechanisms are integrated into an asynchronous mentor-mentee architecture, enabling efficient and deep knowledge transfer from powerful, private mentor models to a lightweight, global mentee model. Simulation results on multiple real-world datasets demonstrate that Fed-GPD significantly outperforms existing federated and graph-based anomaly detection methods. Notably, our in-depth ablation studies validate the effectiveness of the proposed process distillation mechanisms, showing substantial improvements in model accuracy and communication efficiency.

Mobile edge computing (MEC) is a promising paradigm that provides abundant computation and storage resources at the edge close to mobile devices (MDs). In MEC networks, MDs offload compute-heavy tasks to nearby edge servers (ESs) for delay-sensitive processing, where relevant services are stored to support task execution. However, the limited computation and storage capacities of ESs make joint optimization of service caching and computation offloading challenging due to coupled decisions, a large solution space, and dynamic environments. In this paper, we investigate the joint optimization of service caching and computation offloading in MEC networks, aiming to maximize the cache hit ratio and minimize the average service latency. To tackle this problem, the original formulation is decomposed into two hierarchical subproblems, namely high-level service caching and low-level computation offloading. We propose a novel hierarchical deep reinforcement learning (DRL) algorithm with active inference, termed HADRL. At the high-level, we adopt a deep deterministic policy gradient (DDPG) based DRL approach to maximize the cache hit ratio. At the low-level, we employ an active inference based DRL approach to minimize the average service latency. Unlike conventional DRL, the active inference based DRL approach selects policies by minimizing expected free energy instead of relying only on explicit rewards, making it well suited for highly dynamic low-level computation offloading. According to the simulation outcomes, the HADRL scheme surpasses the benchmark algorithms with respect to cache hit ratio as well as average service latency.

Underground fires pose a significant threat to production safety in coal mines, and existing detection methods suffer from drawbacks such as poor adaptability to complex subterranean environments and excessive model parameters. To address the need for deploying object detection models on resource-constrained devices, this paper proposes a novel and efficient algorithm forUndergroundFireYOLOdetection, named UF-YOLO. The core innovation of this method is threefold: first, the StarNet module is introduced into the backbone to significantly reduce model parameters and computational complexity without sacrificing accuracy; second, the Cross-scale Context Fusion Module (CCFM) is integrated into the neck to enhance the model’s detection capability for fires of various scales, particularly small targets; and finally, Partial Convolution (PConv) is integrated to extract spatial features more efficiently, further reducing redundant computations and memory access. On our self-built Mine Fire Image Dataset (MFID), compared to the baseline model YOLOv11m, UF-YOLO reduces parameters by 77.1%, increases inference speed by 60.6%. Experimental results on the public COCO val 2017 dataset demonstrate that the proposed method outperforms state-of-the-art (SOTA) models such as YOLOv12. The results confirm that UF-YOLO can be efficiently deployed on the edge-side of coal mine IoVT monitoring systems to performe accurate and real-time fire detection. This work provides a new intelligent paradigm for the real-time monitoring of underground fires.

The security of voice control in the Voice Internet of Things (Voice IoT) heavily relies on the fast and accurate authentication of the command issuer. In this work, we focus on the critical application scenario of Voice IoT in underground coal mines, where voice commands typically last 4–10 seconds. Speech in this scenario typically consists of short, imperative utterances and faces challenges from environmental noise and device heterogeneity. The limitations of traditional speaker verification models in temporal modeling restrict their performance in such scenarios. To address this, this paper proposes a three-dimensional attention module (Branch-MFA) designed for Voice IoT. This module employs a dual-parallel branch architecture: the MFA branch is responsible for extracting attention in the frequency and channel dimensions, and its multi-scale nature enables it to effectively focus on speaker-discriminative frequency bands that remain stable under noise and different collection devices, thereby enhancing the model’s environmental robustness; the GLTA branch, through its innovative grouped variable-length attention mechanism, specifically models the temporal structure of these short voice commands, addressing the challenge of sparse temporal information in short utterances. By integrating the dual-branch outputs through a fusion module, we construct the Branch-MFA-TDNN model. Experiments on the Cn-Celeb dataset show that this model significantly outperforms baseline models in short-utterance verification tasks, particularly for the challenging 4–10 second duration relevant to mine communications, providing an identity authentication solution for Voice IoT that combines high security and real-time performance. We have also released the code1for future comparison.

The Internet of Nano Things (IoNT) refers to an interconnected network of nanoscale components engineered to perform tasks such as data processing, storage, and actuation. IoNT has broad applications, including environmental monitoring and pollution source localization. In order to achieve monitoring and localization for multiple releasing sources (RSs), the deployment of nanosensor networks is indispensable. However, constrained by spatial limitations and high costs, sensors can only be sparsely deployed, resulting in severe degradation in localization performance. In this paper, we consider a turbulent diffusion molecular communication scenario and the objective is to enable multi-source localization and obstacle perception with sparse nanosensors. For sparse signal recovery, we first propose a real-symmetric based on Truncated Nuclear Norm Regularization with Alternating Direction Method of Multipliers (RS-TNNR) matrix completion algorithm, which utilizes the spatial symmetry of molecular diffusion to achieve precise data recovery under high missing ratios. Furthermore, for multi-source localization and obstacle perception, we also propose an Adaptive Iterative Grid based on Sparse Bayesian Learning (AIG-SBL) algorithm, which enhances the localization accuracy with SBL, mitigates off-grid errors via the proposed adaptive iterative grid, and simultaneously estimates obstacle position and radii. Simulation results demonstrate the effectiveness of the proposed algorithms for RS-TNNR and AIG-SBL.

This paper addresses the challenges of automatic monitoring and restoration in ancient mural conservation, aiming to enhance the efficiency and quality of heritage preservation. Traditional manual inspection is time-consuming and often misses early damage, while existing digital restoration models struggle with consistent restoration, especially for large-scale damage. To address these issues, we propose an Internet of things (IoT)-based solution combining event-driven edge intelligence and artificial intelligence generated content (AIGC) techniques. A fine-tuned EdgeSAM model, using a Conv-adapter, enables efficient damage segmentation at the edge; an event-driven mechanism reduces resource consumption; and a LoRA-tuned PowerPaint model, aided by Blip2 and Qwen, provides effective restoration of large damaged areas. Cloud-side processing utilizes AIGC techniques to restore damaged mural areas, ensuring high-quality restoration while minimizing communication demands. Experimental results demonstrate that the proposed method achieves accurate damage monitoring on resource-constrained edge devices and generates diverse, contextually appropriate restoration results on cloud servers, providing a deployment-oriented feasibility validation under simulated temporal degradation and real hardware constraints.

Due to the triple mobility of mobile users (MUs), unmanned aerial vehicle (UAV) relays, and low Earth orbit (LEO) satellites, handover becomes a critical and challenging issue for maintaining the continuity and quality of communication services in UAV-assisted LEO satellite networks. This paper proposes a distributed handover decision-making process aimed at improving scalability and reducing communication overhead. The handover problem is modeled as a decentralized Markov decision process (DEC-MDP) with the objective of maximizing the total end-to-end (E2E) throughput. We design an independent proximal policy optimization-based distributed intelligent handover (IPPO-DIH) algorithm within a centralized training with decentralized execution framework to solve the DEC-MDP. To analyze the theoretical optimal E2E throughput, we eliminate the correlation between handover decisions at different time steps. A three-sided matching algorithm with theoretical convergence guarantees is designed to obtain a stable matching among MUs, UAV relays, and LEO satellites at each time step. These stable matchings are combined to provide a theoretical performance benchmark for the handover algorithms. Simulation results validate the convergence of the proposed IPPO-DIH and three-sided matching algorithms. Additionally, the total E2E throughput achieved by the IPPO-DIH algorithm approaches the theoretical performance benchmark and outperforms typical handover algorithms.

Perceiving scene depth and 3D structure is fundamental for environmental comprehension and interaction in Internet of Video Things (IoVT) devices. Self-supervised monocular depth estimation, which leverages photometric consistency across sequential video frames without requiring ground-truth labels from active sensors, has emerged as a compelling paradigm. Nevertheless, this paradigm suffers from inherent limitations in low-texture or occluded regions, where photometric supervision becomes ambiguous or invalid, typically leading to structural degradation and indistinct object boundaries when inferring depth. To mitigate this limitation, we propose a dual-branch local distillation framework that harnesses priors from depth foundation models (DFMs) to alleviate depth ambiguity and improve fine-grained estimation accuracy. Specifically, guided by a frozen DFM, the framework synergistically employs local-context and cross-context supervision to optimize a student network, facilitating robust perception of depth discontinuities and precise boundary modeling. Furthermore, to enhance the spatial representation within a lightweight architecture, we design DE-LiteMono as the student model, which recovers geometric details through a detail enhancement block and fusion modules. Moreover, a spatial distance consistency loss is introduced to explicitly model the relative geometric topology in the image plane, providing robust structural guidance. Extensive experiments on standard benchmarks, including KITTI, Cityscapes, and Make3D, demonstrate that the proposed method outperforms state-of-the-art self-supervised methods, achieving sharper depth boundaries, reduced ambiguity, and superior quantitative performance.

Industrial IoVT (Internet of Video Things) still faces the dual bottleneck of insufficient accuracy and poor real-time performance in circuit board tiny defect detection. To this end, we propose RGM-YOLO (RefConv–GhostNet–CBAM-enhanced YOLOv8 ), which introduces deformable convolution and channel attention via RefConv and GhostNet modules, and experimentally validates it on the BDL-PCB (Bare Die on Laminate–Printed Circuit Board) large-scale dataset. Experimental results show that RGM-YOLO achieves 94.2% in mAP50 and 67.3% in mAP90–95, representing improvements of 2.4% and 11.2% over the baseline model, YOLOv8. The number of parameters and GFLOPs is reduced by 4.2M and 2.5G, respectively, while the FPS increases from 78 to 102. This approach offers a high-precision, low-latency defect detection paradigm for edge IoVT devices targeting small defects and can be generalized to other industrial quality-inspection scenarios.

Integrated Sensing, Communication, and Computation (ISCC) has the potential to meet diverse requirements of Internet of Vehicles (IoV), such as high reliability and low power consumption. However, existing works have not fully considered the problems of unreliable communication links and inefficient data processing under resource constraints in non-ideal environments. To address these issues, this paper proposes a predictive Integrated Sensing, Communication and Computation Over-the-Air (ISCCO) approach based on Orthogonal Time Frequency Space (OTFS) modulation. It takes high Doppler shifts, network dynamics, and resource constraints into account. In particular, the Road Side Unit (RSU) performs target tracking while communicating with the downlink users through Space Division Multiplexing (SDM), and receives the transmission results of the uplink. For the downlink, a predictive beamforming approach based on Extended Kalman Filtering (EKF) is employed, while Over-the-Air computation (AirComp) is utilized for the uplink. The transmit power and receive beamformer at the RSU, along with the transmit power of the uplink users, are jointly optimized through two formulated optimization problems: sensing performance maximization and power consumption minimization. To solve these problems, we adopt an Alternating Optimization (AO)-based algorithm for finding the local optimal solution. Simulation results validate the effectiveness of the AO-based algorithm, and the analysis of the trade-offs between multi-dimensional performance of ISCC and power consumption is conducted.

Federated Learning (FL) is a promising learning paradigm that allows for training a shared model by coordinating multiple distributed devices, namely, clients, without exposing their raw data. To mitigate excessive communication overhead and enhance practicality, a variant known as Hierarchical FL (HFL) has been introduced, which integrates edge servers between the cloud server and clients. In HFL, the number of potential clients is typically large, making full client participation impractical due to various resource constraints. As a result, it is essential to develop a sampling strategy that effectively selects suitable clients for federated optimization. While several methods have been proposed to protect the privacy of communicated models, we argue that the outcomes of client sampling are closely tied to the local data of clients, thereby raising privacy concerns, like the risk of differential attacks. Motivated by this, we propose a Two-step Privacy-Preserving client Sampling framework (TPPS) designed to protect against both attacks on communicated models and potential vulnerabilities in client sampling outcomes. Initially, we consider the diverse privacy requirements of clients by presenting a double-layer noise mechanism. We then conduct a thorough analysis of the impact of this noise mechanism, proposing a novel client sampling strategy that seeks to balance the trade-off between privacy and training performance. The insight lies in maintaining a real-time sampling probability for each client, which can be acutely tuned based on personalized privacy needs and previous training feedback. We provably show that TPPS achieves local differential privacy, a bounded sampling regret, and a privacy-related convergence rate. Furthermore, we conduct extensive simulations based on open datasets, showing the robustness and applicability of TPPS in enhancing privacy while optimizing HFL performance.

The covert communication scenario with multiple users involved needs to take into account multiple conflicting optimization goals, such as maximizing the number of served users, minimizing service energy consumption, and balancing resource allocation. And the UAV-assisted multi-user communication is a typical knapsack traveling salesman problem, which is difficult to solve because of its large solution space and NP-hard nature. Therefore, to address this, a novel UAV-assisted multi-user downlink covert communication scenario is considered. To find an optimal balance of the maximum transmission rate and the minimum energy consumption, a low-complexity solution is proposed. First, a closed-form solution for the hovering position and transmission power of the UAV is deduced. Then, the user selection and trajectory planning results from the heuristic algorithm are used to initialize the deep reinforcement learning (DRL) agent, providing a superior starting point for the subsequent policy search. This strategy combines the fast convergence of heuristics with the dynamic adaptability of DRL, resulting in a final design that is more efficient and lightweight than conventional, pure DRL-based methods. The simulation results demonstrate that the proposed scheme improves energy efficiency by at least 40% compared with the benchmark scheme.

Band selection (BS) is a key method in Hyperspectral image (HSI) classification that helps to reduce the computational burden and improve the class separability. However, with the emerging of unmanned aerial vehicle (UAV)-borne HSI datasets, their attributes such as high spatial and spectral resolution as well as large-scale samples pose serious challenges to the existing BS methods, making them inefficient. In addition, the efficient utilization of the prior knowledge from the data collected by fixed UAV-borne sensors in different regions is often easily overlooked. In view of these issues, this paper proposes a neural network-assisted particle swarm optimization (PSO) algorithm for cross-domain BS of UAV-borne HSIs. First, a knowledge learning strategy is designed for the source domain, which applies a neural network model to learn the useful prior knowledge in labeled source domain data. Then, a network-assisted PSO algorithm is introduced to search for the optimal subset of bands in the target domain under the guidance of the valid prior knowledge captured from the source domain by the network model. Moreover, a similarity-based grouping strategy is designed to group similar bands and then select bands from each group with the aims of reducing the redundant information in the subset of bands. Finally, experimental results on three common UAV-borne HSI datasets show that our proposed method can efficiently handle UAV-borne HSI data with large samples, as it is able to find a subset of bands with higher quality compared to several state-of-the-art BS methods.

Recently, edge computing has emerged as a prevailing paradigm in applying deep learning-based object detection models, offering a promising solution for time-sensitive tasks. However, existing edge object detection faces several challenges: 1) These methods struggle to balance detection precision and model lightweightness. 2) Existing generalized edge-deployment designs offer limited adaptability for object detection. 3) Current works lack real-world evaluation and validation. To address these challenges, we propose theEdgeDetectionToolbox(ED-TOOLBOX), which leverages generalizable plug-and-play components to enable edge-site adaptation of object detection models. Specifically, we propose a lightweightReparameterized Dynamic Convolutional Network(Rep-DConvNet) that employs a weighted multi-shape convolutional branch structure to enhance detection performance. Furthermore, ED-TOOLBOX includes aSparse Cross-Attention(SC-A) network that adopts a localized-mapping-assisted self-attention mechanism to facilitate a well-craftedJoint Modulein adaptively transferring features for further performance improvement. Moreover, we propose anEfficient Headfor the classification and location modules to achieve more efficient prediction. Additionally, in practical industrial scenarios, we identify that helmet detection-one of the most representative edge object detection tasks-overlooks band fastening, which introduces potential safety hazards. To address this, we build aHelmet Band Detection Dataset(HBDD) and apply an edge object detection model optimized by the ED-TOOLBOX to tackle this real-world task. Extensive experiments validate the effectiveness of components in ED-TOOLBOX. In visual surveillance simulations, ED-TOOLBOX-assisted edge detection models outperform sixstate-of-the-artmethods, enabling real-time and accurate detection. These results demonstrate that our approach offers a superior solution for edge object detection.

The rapid adoption of 5G technology and Internet of things (IoT) devices has fueled significant growth in intelligent applications, increasing their complexity beyond simple task definitions. Scheduling intelligent applications modeled as directed acyclic graphs (DAGs) has thus emerged as a crucial challenge. Our proposed solution is a deep reinforcement learning (DRL) framework that uniquely integrates proximal policy optimization (PPO) with a transformer-based module for scheduling DAG applications. Unlike other approaches that rely on predefined priorities or static optimization algorithms, our approach enables agents to autonomously explore task execution orders and dynamically adapt to changing network resource conditions, learning optimal scheduling strategies. The algorithm leverages transformers to handle complex task dependencies, minimizing application duration and user energy consumption by jointly optimizing application processing order, task priorities, transmit power, offloading decisions, and computational frequency. Through a series of simulations, we prove the effectiveness of the proposed algorithm and demonstrate the performance comparison under different settings, providing a more flexible and robust solution for DAG scheduling in edge networks.

The Internet of Things (IoT) employing symbiotic radio (SR) technology encounters challenges such as low throughput and susceptibility to double fading. To address these challenges, this paper integrates non-orthogonal multiple access (NOMA) with simultaneous transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) technology in an SR system, introducing a novel transmission model termed STAR-RIS-assisted NOMA-SR with hybrid active-passive communication. The proposed model operates in three phases. In the first two phases, when the primary system’s licensed spectrum is occupied, the backscatter devices (BDs) utilize backscatter communication (BC) to establish a symbiotic relationship with the primary system. Specifically, in Phase 1, STAR-RIS enhances the energy harvesting (EH) of BDs via the reflection mode, while in Phase 2, it aids both the primary and secondary systems via the transmission mode. In Phase 3, when the licensed spectrum is idle, STAR-RIS facilitates the active communication (AC) of BDs via the transmission mode. To maximize the total throughput of BDs while guaranteeing the primary system’s target throughput, we formulate a non-convex optimization problem and develop a block coordinate descent (BCD)-based resource allocation scheme. The problem is decomposed into subproblems and solved using successive convex approximation (SCA), variable substitution, and semi-definite relaxation (SDR) to jointly optimize transmission time, beamforming, STAR-RIS reflection and transmission coefficients, as well as BDs’ power allocation and reflection coefficients. Numerical results show that the proposed scheme enhances the total throughput of BDs by 14.36%, 43.43%, 67.78%, and 439.69% compared to four baseline schemes.

Graph Neural Networks (GNNs) have shown efficacy in graph node classification, but face computational challenges on large-scale graphs. Although existing graph reduction methods address these issues, they still require high computational resources and fail to prioritize robust performance on out-of-distribution data. To tackle these challenges, we introduce the subgraph invariant learning paradigm, inspired by the small-world phenomenon. This approach enables models trained on specific subgraphs to generalize across diverse subgraphs, reducing computational demands, and enhancing scalability. To promote generalization, we maximize the invariance log-likelihood by deriving a theoretical lower bound of it and formulating the InVar loss. This loss minimizes the discrepancy between node representations and their corresponding invariance representations while maximizing the entropy of the node representation. In response to InVar loss, we propose the Invariance Facilitation Model (IFM), comprising the Invariance Representation Encoder (IRE) and Node Representation Encoder (NRE). IRE, capturing the invariance representations, utilizes Invariance ATTention (InvarATT) to compress long-range dependencies, while NRE learns the node representation, by integrating invariance representations via Telematic ATTention (TeleATT) and exchanging local information within each subgraph through GNNs. Evaluations on four large-scale graph datasets demonstrate the effectiveness, computational efficiency, and interpretability of IFM for large-scale graph node classification.

Pose-controlled human video generation is of significant interest and finds extensive applications in areas such as automated advertising and content creation on social media platforms. While existing methods employing pose sequences and reference images for human image animation have exhibited notable performance, they tend to encounter issues such as specific region blurring, background sharpening, and decreased identity consistency. In this paper, we introduce ReMask-Animate, which utilizes masks as additional priors to guide the model's local visual attention to specific areas, thereby alleviating feature confusion between different regions of the image. Three distinct mask-guided adapters are designed for cross-condition regional fusion of hand and face pose features, mitigating feature confusion between the foreground and background, and enhancing the visual consistency of character identity. Moreover, these lightweight adapters introduce minimal computational overhead and can be seamlessly integrated into specific layers of the backbone architecture. Extensive experiments show that our method outperforms state-of-the-art methods on five metrics in public datasets. Additionally, qualitative evaluations highlight a significant improvement in the quality of generated videos, demonstrating our approach's superiority.

Talking head synthesis with arbitrary speech audio is a crucial challenge in the field of digital humans. Recently, methods based on radiance fields have received increasing attention due to their ability to synthesize high-fidelity and identity-consistent talking heads from just a few minutes of training video. However, due to the limited scale of the training data, these methods often exhibit poor performance in audio-lip synchronization and visual quality. In this paper, we propose a novel 3D Gaussian-based method called PointTalk, which constructs a static 3D Gaussian field of the head and deforms it in sync with the audio. It also incorporates an audio-driven dynamic lip point cloud as a critical component of the conditional information, thereby facilitating the effective synthesis of talking heads. Specifically, the initial step involves generating the corresponding lip point cloud from the audio signal and capturing its topological structure. The design of the dynamic difference encoder aims to capture the subtle nuances inherent in dynamic lip movements more effectively. Furthermore, we integrate the audio-point enhancement module, which not only ensures the synchronization of the audio signal with the corresponding lip point cloud within the feature space, but also facilitates a deeper understanding of the interrelations among cross-modal conditional features. Extensive experiments demonstrate that our method achieves superior high-fidelity and audio-lip synchronization in talking head synthesis compared to previous methods.

Visual Language Models (VLMs) have rapidly progressed with the recent success of large language models. However, there have been few attempts to incorporate efficient linear Recurrent Neural Networks (RNNs) architectures into VLMs. In this study, we introduce VisualRWKV, the first application of a linear RNN model to multimodal learning tasks, leveraging the pre-trained RWKV language model. We propose a data-dependent recurrence and sandwich prompts to enhance our modeling capabilities, along with a 2D image scanning mechanism to enrich the processing of visual sequences. Extensive experiments demonstrate that VisualRWKV achieves competitive performance compared to Transformer-based models like LLaVA-1.5 on various benchmarks. Compared to LLaVA-1.5, VisualRWKV has a speed advantage of 3.98 times and can save 54% of GPU memory when reaching an inference length of 24K tokens. To facilitate further research and analysis, we have made the checkpoints and the associated code publicly accessible at the following GitHub repository: https://github.com/howard-hou/VisualRWKV.

Event cameras capture visual information as asynchronous pixel change streams, excelling in challenging lighting and high-dynamic scenarios. Existing multimodal large language models (MLLMs) concentrate on natural RGB images, failing in scenarios where event data fits better. In this paper, we introduce EventGPT, the first MLLM for event stream understanding, pioneering the integration of large language models (LLMs) with event-based vision. To bridge the huge domain gap, we propose a three-stage optimization paradigm to progressively equip a pre-trained LLM with event understanding. Our EventGPT consists of an event encoder, a spatio-temporal aggregator, a linear projector, an event-language adapter, and an LLM. Firstly, GPT-generated RGB image-text pairs warm up the linear projector, following LLaVA, as the gap between natural images and language is smaller. Secondly, we construct N-ImageNet-Chat, a large synthetic dataset of event data and corresponding texts to enable the use of the spatio-temporal aggregator and to train the event-language adapter, thereby aligning event features more closely with the language space. Finally, we gather an instruction dataset, EventChat, which contains extensive real-world data to fine-tune the entire model, further enhancing its generalization ability. We construct a comprehensive benchmark, and experiments show that EventGPT surpasses previous state-of-the-art MLLMs in generation quality, descriptive accuracy, and reasoning capability. Code: EventGPT

Achieving robust generalization across diverse scenarios is crucial for advancing the practical application of robotics. Existing approaches typically rely on interaction object masks as visual inputs to predict subsequent actions, gaining a certain degree of generalization capability. However, these methods primarily map visual inputs and task-relevant object masks to expert actions, overlooking the environmental dynamics that govern physical interactions among objects during manipulation. To overcome these limitations, we introduce ABM++, a novel framework that leverages pre-trained VLMs to build a mask-guided world model (MGWM) within an imitation learning paradigm for generalized robotic manipulation. Specifically, we extend a world model into a coarse-to-fine imitation learning framework to reconstruct future mask-dominated visual features, which enables the model to capture state transitions between the current and next states based on predicted actions, effectively modeling environmental dynamics. Comprehensive experiments demonstrate that ABM++ significantly surpasses established baselines in both simulation and real-world environments, achieving a relative improvement of 12.2% across 8 complex tasks, which underscores the superiority of our method.

Fine-tuning large language models (LLMs) often causes overfitting to specific prompt wording, where minor phrasing variations drastically reduce performance.To address this, we propose Prompt-Agnostic Fine-Tuning (PAFT), a method that enhances robustness through dynamic prompt variation during training.PAFT first generates diverse synthetic prompts, then continuously samples from this set to construct training instances, forcing models to learn fundamental task principles rather than surface-level patterns.Across systematic evaluations using both supervised fine-tuning (SFT) and reinforcement learning fine-tuning (RLFT), PAFT demonstrates substantially improved prompt robustness, achieving 7% higher generalization accuracy on unseen prompts than standard methods.In addition to enhanced robustness, PAFT consistently yields superior overall performance on established benchmarks for question answering, mathematical reasoning, and tool use.Notably, models trained with PAFT attain 3.2× faster inference speeds due to reduced prompt sensitivity.Ablation studies further validate effectiveness of PAFT, while theoretical analysis reveals that PAFT can effectively enhance the cross-domain generalization ability of LLM.

Increasing attention has been garnered by LiDAR points and multi-view images fusion based on Transformer for supplementing another modality in 3D object detection. However, challenges persist for cross-modal fusion methods due to the heterogeneity of these two modalities, leading to issues such as inaccurate detection results encountered by Transformer-based methods. In this work, a one-way mid-level fusion based framework for 3D object detection named Modal Fusion Transformer (MFT) using LiDAR points and multi-view images is introduced. It comprises a Depth-Guided Generation(DGG) module, Position Encoding Generation (PEG) module and Cross Modal Fusion(CMF) module. Specifically, depth information from point cloud is utilized for both gathering the image depth map and initializing object queries in DGG. PEG unifies the form of position encoding from LiDAR features and multi-view image features. Depth and position information of object from images is aggregated to point clouds by CMF, which fully explores dual-modal information. Furthermore, a Modal Fusion Network with deformable attention named fast-MFT is introduced to reduce the relatively large computational cost associated with global attention. Our MFT and fast-MFT achieve competitive performance while maintaining a faster inference speed than other models.

Inderscience is a global company, a dynamic leading independent journal publisher disseminates the latest research across the broad fields of science, engineering and technology; management, public and business administration; environment, ecological economics and sustainable development; computing, ICT and internet/web services, and related areas.

The latest Segment Anything Model enables realtime scene annotation and understanding for autonomous driving systems, enhancing driving safety. However, effectively allocating resources for real-time performance and accuracy remains challenging in edge-cloud architectures. Traditional reinforcement learning struggles with poor generalization and the explorationexploitation dilemma, making it difficult to define clear reward functions. To address this, we propose a likelihood active inference approach to optimize resource allocation and improve system resource utilization. We use "intelligence" as a high-level indicator to quantify the efficiency of cognition in active inference, evaluating the difference between predicted and actual states during policy exploration. Experimental results show our algorithm outperforms mainstream deep reinforcement learning algorithms, improving sample efficiency and suitability for dynamically changing task workloads.

With the widespread adoption of edge computing, service migration is critical for meeting real-time computing demands and ensuring service continuity. However, the dynamic and uncertain nature of edge computing-enabled networks, characterized by fluctuating topologies, bandwidth, and resources, significantly complicates migration decisions. Existing strategies rely on precise analytical models and reward functions but struggle with generalization and adaptability. This paper proposes a novel service migration strategy driven by active inference for edge computing-enabled networks. Unlike traditional approaches, it eliminates the need for explicit reward functions, instead leveraging a cognitive optimization mechanism where decisions are guided by minimizing free energy. This allows the system to maintain efficient service migration across a wider range of edge scenarios, with enhanced generalization and flexibility. Simulation results show that the proposed strategy outperforms existing approaches by reducing latency and improving adaptability to varying environments, highlighting its superiority in service migration for edge computing-enabled networks.

The management of computing resources through the computing power network (CPN) has gradually become a focal point of research. With the development of the 6th generation (6G) mobile networks, some promising technologies such as satellite-terrestrial integrated network (STIN) and smart endogenous network driven by artificial intelligence (AI) are increasingly being applied in Industrial Internet of Things (IIoT). However, several issues in current studies are worthy of attention: 1) the large number of devices powered by battery in IIoT, 2) the complex environments of communication, 3) the finite computing resources for task data processing. To cope with these challenges, a satellite-terrestrial integrated computing power network (STICPN) framework is introduced in this article. Within this framework, a task offloading link selection scheme is proposed, which minimizes the delay and the consumption of energy. The task offloading optimization problem is modeled as a Markov decision process (MDP). Meanwhile, deep reinforcement learning (DRL) algorithm is employed to adapt to the dynamic states of environment. Specifically, a dueling double deep Q network (D3QN) is used to make optimal decisions and delay as well as energy consumption can be reduced significantly. Moreover, the D3QN-based scheme extends the usage time of IIoT devices. The simulation results indicate that the proposed scheme outperforms comparison schemes significantly.

Web 4.0 is characterized by decentralized intelligence and blockchain integration, which introduces significant challenges in congestion management for software-defined networking (SDN). Traditional reinforcement learning (RL)-based approaches encounter inefficiencies due to limited adaptability to decentralized and delayed online learning capabilities. To address these issues, we propose an Active Inference-based Reinforcement Learning (AIRL) framework that integrates generative modelling with RL for enhanced decision-making in congestion control. By leveraging blockchain-enabled secure model trading and predictive intelligence, AIRL ensures adaptive policy optimization while maintaining transparency and trust in decentralized network environments. The proposed method demonstrates substantial improvements in delay reduction, packet loss, and efficient utilization of network resources under various dynamic scenarios.

Recent interest in large-scale text-driven diffusion models has highlighted their ability to generate diverse images from textual prompts, with style transfer being a significant application. However, existing text-guided stylization methods are constrained by the reliance on manually created masks for localized transformations, which limits scalability and automation. In this work, a novel framework, OmniStyle, is introduced for text-driven global and localized image style transfer, leveraging diffusion model inversion and attention-based mask generation. Input images are mapped into latent noise representations using DDIM inversion, while cross-attention maps are utilized to automatically generate precise semantic masks, eliminating the need for manual annotations. Latent features are dynamically optimized with mask-guided constraints and attention manipulations, enabling fine-grained style transfer confined to target regions while maintaining the integrity of global content. Extensive experiments demonstrate the scalability and effectiveness of OmniStyle.

Precise audio-visual synchronization in speech videos is crucial for content quality and viewer comprehension. Existing methods have made significant strides in addressing this challenge through rule-based approaches and end-to-end learning techniques. However, these methods often rely on limited audio-visual representations and suboptimal learning strategies, potentially constraining their effectiveness in more complex scenarios. To address these limitations, we present UniSync, a novel approach for evaluating audio-visual synchronization using embedding similarities. UniSync offers broad compatibility with various audio representations (e.g., Mel spectrograms, HuBERT) and visual representations (e.g., RGB images, face parsing maps, facial landmarks, 3DMM), effectively handling their significant dimensional differences. We enhance the contrastive learning framework with a margin-based loss component and cross-speaker unsynchronized pairs, improving discriminative capabilities. UniSync outperforms existing methods on standard datasets and demonstrates versatility across diverse audio-visual representations. Its integration into talking face generation frameworks enhances synchronization quality in both natural and AI-generated content.

Open-ended story visualization is a challenging task that involves generating coherent image sequences from a given storyline. One of the main difficulties is maintaining character consistency while creating natural and contextually fitting scenes—an area where many existing methods struggle. In this paper, we propose an enhanced Transformer module that uses separate self attention and cross attention mechanisms, leveraging prior knowledge from pre-trained diffusion models to ensure logical scene creation. The isolated self attention mechanism improves character consistency by refining attention maps to reduce focus on irrelevant areas and highlight key features of the same character. Meanwhile, the isolated cross attention mechanism independently processes each character’s features, avoiding feature fusion and further strengthening consistency. Notably, our method is training-free, allowing the continuous generation of new characters and storylines without re-tuning. Both qualitative and quantitative evaluations show that our approach outperforms current methods, demonstrating its effectiveness.

Face editing modifies the appearance of face, which plays a key role in customization and enhancement of personal images. Although much work have achieved remarkable success in text-driven face editing, they still face significant challenges as none of them simultaneously fulfill the characteristics of diversity, controllability and flexibility. To address this challenge, we propose MuseFace, a text-driven face editing framework, which relies solely on text prompt to enable face editing. Specifically, MuseFace integrates a Text-to-Mask diffusion model and a semantic-aware face editing model, capable of directly generating fine-grained semantic masks from text and performing face editing. The Text-to-Mask diffusion model provides diversity and flexibility to the framework, while the semantic-aware face editing model ensures controllability of the framework. Our framework can create fine-grained semantic masks, making precise face editing possible, and significantly enhancing the controllability and flexibility of face editing models. Extensive experiments demonstrate that MuseFace achieves superior high-fidelity performance.

Large Language Models (LLMs) have become indispensable in numerous real-world applications. However, fine-tuning these models at scale, especially in federated settings where data privacy and communication efficiency are critical, presents significant challenges. Existing approaches often resort to parameter-efficient fine-tuning (PEFT) to mitigate communication overhead, but this typically comes at the cost of model accuracy. To this end, we propose *federated full-parameter tuning at scale for LLMs* (Ferret), **the first first-order method with shared randomness** to enable scalable full-parameter tuning of LLMs across decentralized data sources while maintaining competitive model accuracy. Ferret accomplishes this through three aspects: **(i)** it employs widely used first-order methods for efficient local updates; **(ii)** it projects these updates into a low-dimensional space to considerably reduce communication overhead; and **(iii)** it reconstructs local updates from this low-dimensional space with shared randomness to facilitate effective full-parameter global aggregation, ensuring fast convergence and competitive final performance. Our rigorous theoretical analyses and insights along with extensive experiments, show that Ferret significantly enhances the scalability of existing federated full-parameter tuning approaches by achieving high computational efficiency, reduced communication overhead, and fast convergence, all while maintaining competitive model accuracy. Our implementation is available at [https://github.com/allen4747/Ferret](https://github.com/allen4747/Ferret).

Invisible watermarking is widely used to protect digital images from unauthorized use. Accurate assessment of watermarking efficacy is crucial for advancing algorithmic development. However, existing statistical metrics, such as PSNR, rely on access to original images, which are often unavailable in text-driven generative watermarking and fail to capture critical aspects of watermarking, particularly visibility. More importantly, these metrics fail to account for potential corruption of image content. To address these limitations, we propose WMarkGPT, the first multimodal large language model (MLLM) specifically designed for comprehensive watermarked image understanding, without accessing original images. WMarkGPT not only predicts watermark visibility but also generates detailed textual descriptions of its location, content, and impact on image semantics, enabling a more nuanced interpretation of watermarked images. Tackling the challenge of precise location description and understanding images with vastly different content, we construct three visual question-answering (VQA) datasets: an object location-aware dataset, a synthetic watermarking dataset, and a real watermarking dataset. We introduce a meticulously designed three-stage learning pipeline to progressively equip WMarkGPT with the necessary abilities. Extensive experiments on synthetic and real watermarking QA datasets demonstrate that WMarkGPT outperforms existing MLLMs, achieving significant improvements in visibility prediction and content description. The datasets and code are released at https://github.com/TanSongBai/WMarkGPT.

Recent advancements in implicit 3D reconstruction methods, e.g., neural rendering fields and Gaussian splatting, have primarily focused on novel view synthesis of static or dynamic objects with continuous motion states. However, these approaches struggle to efficiently model a human-interactive object with n movable parts, requiring 2^n separate models to represent all discrete states. To overcome this limitation, we propose Inter3D, a new benchmark and approach for novel state synthesis of human-interactive objects. We introduce a self-collected dataset featuring commonly encountered interactive objects and a new evaluation pipeline, where only individual part states are observed during training, while part combination states remain unseen. We also propose a strong baseline approach that leverages Space Discrepancy Tensors to efficiently modelling all states of an object. To alleviate the impractical constraints on camera trajectories across training states, we propose a Mutual State Regularization mechanism to enhance the spatial density consistency of movable parts. In addition, we explore two occupancy grid sampling strategies to facilitate training efficiency. We conduct extensive experiments on the proposed benchmark, showcasing the challenges of the task and the superiority of our approach. The code and data are publicly available at https://github.com/Inter3D-ui/Inter3D.

We propose VideoHumanMIB, a novel framework for Video Human Motion In-betweening that enables seamless transitions between different motion video clips, facilitating the generation of longer and more natural digital human videos. While existing video frame interpolation methods work well for similar motions in adjacent frames, they often struggle with complex human movements, resulting in artifacts and unrealistic transitions. To address these challenges, we introduce a two-stage approach: First, we design an Appearance Reconstruction AutoEncoder to decouple appearance and motion information, extracting robust appearance-invariant features. Second, we develop an enhanced diffusion pretrained network that leverages both motion optical flow and human pose as guidance conditions, enabling the model to learn comprehensive latent distributions of possible motions. Rather than operating directly in pixel space, our model works in a learned latent space, allowing it to better capture the underlying motion dynamics. The framework is optimized with a dual-frame constraint loss and a motion flow loss to ensure temporal consistency and natural movement transitions. Extensive experiments demonstrate that our approach generates highly realistic transition sequences that significantly outperform existing methods, particularly in challenging scenarios with large motion variations. The proposed VideoHumanMIB establishes a new baseline for human motion synthesis and enables more natural and controllable digital human animation.

This research aims to enhance the interaction between humans and robots, especially in environments with multiple similar objects or semantic ambiguities. Traditional command-based interactions typically require users to provide precise descriptions, which often poses a significant challenge. To address this issue, we propose a framework named Tag-GuideBot, which leverages Visual Language Models (VLMs) and utilizes object markers to help locate and identify objects in the environment. By integrating positional point prompts of the target objects with robot motion planning models, we aim to achieve a more accurate understanding and execution of complex commands, thus improving the efficiency and naturalness of interactions. Experimental results demonstrate that TagGuideBot effectively addresses the challenges posed by complex commands and environmental complexities, achieving an accuracy of 66.3% on user instructions extended beyond the training set, providing solid support for further optimization of human-robot interaction.

Dense colored point clouds enhance visual perception and are of significant value in various robotic applications. However, existing learning-based point cloud upsampling methods are constrained by computational resources and batch processing strategies, which often require subdividing point clouds into smaller patches, leading to distortions that degrade perceptual quality. To address this challenge, we propose a novel 2D-3D hybrid colored point cloud upsampling framework (GaussianPU) based on 3D Gaussian Splatting (3DGS) for robotic perception. This approach leverages 3DGS to bridge 3D point clouds with their 2D rendered images in robot vision systems. A dual scale rendered image restoration network transforms sparse point cloud renderings into dense representations, which are then input into 3DGS along with precise robot camera poses and interpolated sparse point clouds to reconstruct dense 3D point clouds. We have made a series of enhancements to the vanilla 3DGS, enabling precise control over the number of points and significantly boosting the quality of the upsampled point cloud for robotic scene understanding. Our framework supports processing entire point clouds on a single consumer-grade GPU, eliminating the need for segmentation and thus producing high-quality, dense colored point clouds with millions of points for robot navigation and manipulation tasks. Extensive experimental results on generating million-level point cloud data validate the effectiveness of our method, substantially improving the quality of colored point clouds and demonstrating significant potential for applications involving large-scale point clouds in autonomous robotics and human-robot interaction scenarios.

Land-air bimodal robots (LABR) are gaining attention for autonomous navigation, combining high mobility from aerial vehicles with long endurance from ground vehicles. However, existing LABR navigation methods are limited by suboptimal trajectories from mapping-based approaches and the excessive computational demands of learning-based methods. To address this, we propose a two-stage lightweight framework that integrates global key points prediction with local trajectory refinement to generate efficient and reachable trajectories. In the first stage, the Global Key points Prediction Network (GKPN) was used to generate a hybrid land-air keypoint path. The GKPN includes a Sobel Perception Network (SPN) for improved obstacle detection and a Lightweight Attention Planning Network (LAPN) to improves predictive ability by capturing contextual information. In the second stage, the global path is segmented based on predicted key points and refined using a mapping-based planner to create smooth, collision-free trajectories. Experiments conducted on our LABR platform show that our framework reduces network parameters by 14% and energy consumption during land-air transitions by 35% compared to existing approaches. The framework achieves real-time navigation without GPU acceleration and enables zero-shot transfer from simulation to reality during deployment.

Predicting the future motion of road participants is a critical task in autonomous driving. In this work, we address the challenge of low-quality generation of low-probability modes in multi-agent joint prediction. To tackle this issue, we propose a two-stage multi-agent interactive prediction framework named keypoint-guided joint prediction after classification-aware marginal proposal (JAM). The first stage is modeled as a marginal prediction process, which classifies queries by trajectory type to encourage the model to learn all categories of trajectories, providing comprehensive mode information for the joint prediction module. The second stage is modeled as a joint prediction process, which takes the scene context and the marginal proposals from the first stage as inputs to learn the final joint distribution. We explicitly introduce key waypoints to guide the joint prediction module in better capturing and leveraging the critical information from the initial predicted trajectories. We conduct extensive experiments on the real-world Waymo Open Motion Dataset interactive prediction benchmark. The results show that our approach achieves competitive performance. In particular, in the framework comparison experiments, the proposed JAM outperforms other prediction frameworks and achieves state-of-the-art performance in interactive trajectory prediction. The code is available at https://github.com/LinFunster/JAM to facilitate future research.

This paper presents a lightweight bidirectional cable-driven ankle exoskeleton system (total mass: 2.6 kg) based on a series elastic actuation architecture (actuator module mass: 1.05 kg). The system utilizes a waist-mounted drive unit and Bowden cables to deliver bidirectional assistance to the ankle joint (nominal force: 460 N, peak force: 680 N). By integrating a dynamically coupled adaptive oscillator (AO), the system achieves robust gait synchronization across a range of walking speeds (0.6 – 1.8 m/s, phase estimation RMSE <2.48%, stride frequency estimation RMSE <0.1 Hz). This is complemented by a Gaussian Process (GP)-based torque planner and a cascaded torque control framework, ensuring seamless coordination with natural gait. Experimental characterization of the actuator demonstrates its high dynamic performance (torque bandwidth: 12.5 Hz) and low-impedance characteristics (peak passive backdrive torque: 0.97 N · m). Human trials involving five participants show that the system significantly expands the ankle joint range of motion (up to [-15.64 °, 20.67 ° ] at high speeds) while reducing peak muscle activation levels in the tibialis anterior (18.54%–30.21%) and gastrocnemius (19.34%– 25.45%). This design, combining lightweight construction with adaptive control strategies, provides a highly effective solution for daily mobility assistance and rehabilitation applications.

Talking face video generation with arbitrary speech audio is a significant challenge within the realm of digital human technology. The previous studies have emphasized the significance of audio-lip synchronization and visual quality. Currently, limited attention has been given to the learning of visual uncertainty, which creates several issues in existing systems, including inconsistent visual quality and unreliable performance across different input conditions. To address the problem, we propose a Joint Uncertainty Learning Network (JULNet) for high-quality talking face video generation, which incorporates a representation of uncertainty that is directly related to visual error. Specifically, we first design an uncertainty module to individually predict the error map and uncertainty map after obtaining the generated image. The error map represents the difference between the generated image and the ground truth image, while the uncertainty map is used to predict the probability of incorrect estimates. Furthermore, to match the uncertainty distribution with the error distribution through a KL divergence term, we introduce a histogram technique to approximate the distributions. By jointly optimizing error and uncertainty, the performance and robustness of our model can be enhanced. Extensive experiments demonstrate that our method achieves superior high-fidelity and audio-lip synchronization in talking face video generation compared to previous methods.

The perception and generation of Human-Object Interaction (HOI) are crucial for fields such as robotics, AR/VR, and human behavior understanding. However, current approaches model this task in an offline setting, where information at each time step can be drawn from the entire interaction sequence. In contrast, in real-world scenarios, the information available at each time step comes only from the current moment and historical data, i.e., an online setting. We find that offline methods perform poorly in an online context. Based on this observation, we propose two new tasks: Online HOI Generation and Perception. To address this task, we introduce the OnlineHOI framework, a network architecture based on the Mamba framework that employs a memory mechanism. By leveraging Mamba's powerful modeling capabilities for streaming data and the Memory mechanism's efficient integration of historical information, we achieve state-of-the-art results on the Core4D and OAKINK2 online generation tasks, as well as the online HOI4D perception task.

The explosive growth of generative video models has amplified the demand for reliable copyright preservation of AI-generated content. Despite its popularity in image synthesis, invisible generative watermarking remains largely underexplored in video generation. To address this gap, we propose Safe-Sora, the first framework to embed graphical watermarks directly into the video generation process. Motivated by the observation that watermarking performance is closely tied to the visual similarity between the watermark and cover content, we introduce a hierarchical coarse-to-fine adaptive matching mechanism. Specifically, the watermark image is divided into patches, each assigned to the most visually similar video frame, and further localized to the optimal spatial region for seamless embedding. To enable spatiotemporal fusion of watermark patches across video frames, we develop a 3D wavelet transform-enhanced Mamba architecture with a novel scanning strategy, effectively modeling long-range dependencies during watermark embedding and retrieval. To the best of our knowledge, this is the first attempt to apply state space models to watermarking, opening new avenues for efficient and robust watermark protection. Extensive experiments demonstrate that Safe-Sora achieves state-of-the- art performance in terms of video quality, watermark fidelity, and robustness, which is largely attributed to our proposals. Code and additional supporting materials are provided in the supplementary.

DeepSeek-R1 has successfully enhanced Large Language Model (LLM) reasoning capabilities through its rule-based reward system. While it's a ''perfect'' reward system that effectively mitigates reward hacking, such reward functions are often discrete. Our experimental observations suggest that discrete rewards can lead to gradient anomaly, unstable optimization, and slow convergence. To address this issue, we propose ReDit (Reward Dithering), a method that dithers the discrete reward signal by adding simple random noise. With this perturbed reward, exploratory gradients are continuously provided throughout the learning process, enabling smoother gradient updates and accelerating convergence. The injected noise also introduces stochasticity into flat reward regions, encouraging the model to explore novel policies and escape local optima. Experiments across diverse tasks demonstrate the effectiveness and efficiency of ReDit. On average, ReDit achieves performance comparable to vanilla GRPO with only approximately 10% the training steps, and furthermore, still exhibits a 4% performance improvement over vanilla GRPO when trained for a similar duration. Visualizations confirm significant mitigation of gradient issues with ReDit. Moreover, theoretical analyses are provided to further validate these advantages.

Tactile perception is essential for embodied agents to understand the physical attributes of objects that cannot be determined through visual inspection alone. While existing methods have made progress in visual and language modalities for physical understanding, they fail to effectively incorporate tactile information that provides crucial haptic feedback for real-world interaction. In this paper, we present VTV-LLM, the first multi-modal large language model that enables universal Visuo-Tactile Video (VTV) understanding, bridging the gap between tactile perception and natural language. To address the challenges of cross-sensor and cross-modal integration, we contribute VTV150K, a comprehensive dataset comprising 150,000 video frames from 100 diverse objects captured across three different tactile sensors (GelSight Mini, DIGIT, and Tac3D), annotated with four fundamental tactile attributes (hardness, protrusion, elasticity, and friction). We develop a novel three-stage training paradigm that includes VTV enhancement for robust visuo-tactile representation, VTV-text alignment for cross-modal correspondence, and text prompt finetuning for natural language generation. Our framework enables sophisticated tactile reasoning capabilities including feature assessment, comparative analysis, and scenario-based decision-making. Extensive experimental evaluations demonstrate that VTV-LLM achieves superior performance in tactile reasoning tasks, establishing a foundation for more intuitive human-machine interaction in tactile domains.

Preserving intellectual property (IP) within a pre-trained diffusion model is critical for protecting the model's copyright and preventing unauthorized model deployment. In this regard, model watermarking is a common practice for IP protection that embeds traceable information within models and allows for further verification. Nevertheless, existing watermarking schemes often face challenges due to their vulnerability to fine-tuning, limiting their practical application in general pre-training and fine-tuning paradigms. Inspired by using mode connectivity to analyze model performance between a pair of connected models, we investigate watermark vulnerability by leveraging Linear Mode Connectivity (LMC) as a proxy to analyze the fine-tuning dynamics of watermark performance. Our results show that existing watermarked models tend to converge to sharp minima in the loss landscape, thus making them vulnerable to fine-tuning. To tackle this challenge, we propose **RoMa**, a **Ro**bust **M**odel w**a**termarking scheme that improves the robustness of watermarks against fine-tuning. Specifically, RoMa decomposes watermarking into two components, including *Embedding Functionality*, which preserves reliable watermark detection capability, and *Path-specific Smoothness*, which enhances the smoothness along the watermark-connected path to improve robustness. Extensive experiments on benchmark datasets MS-COCO-2017 and CUB-200-2011 demonstrate that RoMa significantly improves watermark robustness against fine-tuning while maintaining generation quality, outperforming baselines. The code is available at [https://github.com/xiekks/RoMa](https://github.com/xiekks/RoMa).

Increasing attention has been garnered by LiDAR points and multi-view images fusion based on Transformer to supplement another modality in 3D object detection. However, most current methods perform data fusion based on the entire scene, which entails substantial redundant background information and lacks fine-grained local details of the foreground objects to be detected. Furthermore, global scene fusion results in coarse fusion granularity, and the excessive redundancy leads to slow convergence and reduced accuracy. In this work, a novel Dual Scenes Transformer pipeline (DSTR), which comprises a Global-Scene Integration (GSI) module, Local-Scene Integration (LSI) module and Dual Scenes Fusion (DSF) module, is presented to tackle the above challenge. Concretely, features from point clouds and images are utilized for gathering the global scene information in GSI. The insufficiency issues of global scene fusion are addressed by extracting local instance features for both modalities in LSI, supplementing GSI in a more fine-grained way. Furthermore, DSF is proposed to aggregate the local scene to the global scene, which fully explores dual-modal information. Experiments on the nuScenes dataset show that our DSTR has state of the art (SOTA) performance in certain 3D object detection benchmark categories on validation and test sets.

In the paper, we consider the general semantic transmission in wireless networks based on probability distribution. Firstly, we extract a multidimensional semantic probability distribution function, independent of any a specific wireless channel model, by using the variational inference technique. Secondly, we propose a new semantic similarity metric for measuring the difference between the received semantics and the expected semantics based on Kullback-Leibler divergence. Then, we formulate the semantic transmission problem as an optimization problem of transmission symbol adjustment with the aim to maximize the semantic similarity. Finally, we develop an optimal semantic transformation and transmission (STT) algorithm to obtain the optimal transmission symbol adjustment decision. This decision makes the closed-form expression of semantic transmission symbol available, which can realize lossless semantic transmission with energy constraint. Simulation results verify the effectiveness and robustness of the proposed STT algorithm.

Federated Learning (FL) enables the collaborative training of models by users distributed across various locations, transforming traditional data sharing into model sharing. This paradigm holds the promise of facilitating the development of safe, reliable, and accurate driving models within Internet of Vehicles (IoV), with its performance contingent upon the stability of the training process. However, traditional FL relies on a central server for aggregation, which is susceptible to malicious attacks. Moreover, limited communication resources prevent the inclusion of all users in the training process. To resolve issues related to reliability and resource utilization, this paper proposes a reliable Rotating Server Federated Learning (RSFL) algorithm to enhance the security and efficiency of FL. Specifically, we first consider the vehicular topology and participation in FL during their transition, and introduce a server rotation algorithm that incorporates a weighted sum of multiple factors including model training activity, vehicle credibility, speed stability, and distance to augment system security. Additionally, addressing the limitation of server channel resources that can impede FL efficiency, this paper proposes a method to select high-quality users for channel resource allocation by comprehensively considering participation latency, contribution, energy, and channel state during the FL process. This optimizes resource usage at the FL server side and constructs an efficiency-maximization problem for FL to improve the convergence rate. Simulation results confirm that the proposed RSFL algorithm can significantly enhance the security and system efficiency of FL.

In Internet of things (IoT) systems enabled by 6G, unmanned aerial vehicles (UAVs), acting as communication nodes, have the advantages of flexible deployment and wide-area coverage. The channel prediction capability of UAVs for ground communication is of great significance for improving the reliability of IoT communication systems. We propose an innovative and interpretable paradigm for channel prediction based on “physical feature extraction + machine learning”. Specificallywe proposes a real-time UAV-to-ground channel prediction method that leverages propagation environment sensing data, aiming to enhance prediction accuracy and generalization by deeply integrating environmental and communication information. Firstly, we construct the first UAV sensing-communication integrated dataset featuring multi-band, multi-dimensional channel parameters, including UAV-to-ground RGB images, depth maps, and channel data. We then extract multimodal features with clear physical significance relevant to wireless propagation, such as relative position, relative altitude, relative volume, and transmitter-receiver distance. Finally, this paper designs a fusion architecture based on convolutional neural network (CNN) and multilayer perceptron (MLP). This architecture takes multimodal feature data as input, utilizes CNN to extract local features of multi-modal features, and models the fusion of multi-modal features through MLP. Experimental results demonstrate that our model consistently outperforms comparative model. Importantly, our feature analysis quantitatively reveals—for the first time—that building volume is the most influential factor in channel behavior, and that prediction accuracy degrades with increasing flight altitude. Furthermore, system-level simulations confirm that channel prediction leads to substantial improvements in network performance. This work presents a robust and interpretable framework for environment-aware channel characterization, laying a foundation for future 6G intelligent communication systems.

Federated learning (FL), as a typical distributed machine learning framework, has been effectively applied to traffic flow optimization, driving behavior analysis, and other areas. However, the stability and efficiency of FL systems heavily rely on the quality and cooperation of participants. If participants find no profit in the FL process, they may reduce their willingness to participate due to energy consumption and limited resources. To bridge these gaps, this article proposes a demand-balanced incentive mechanism for multiple FL tasks. First, considering the time-varying channel characteristics in the Internet of Vehicles (IoV) scenario, two optimization problems are constructed: 1) maximizing task-matching satisfaction and 2) minimizing communication energy consumption. Second, these problems are transformed into a distributed incentive mechanism based on a multileader-multifollower (MLMF) Stackelberg game, and a bi-level alternating direction method of multipliers (ADMM) algorithm is proposed to solve for the optimal resource allocation and reward schemes that balance the demands of all parties. Furthermore, this article designs a multiagent deep reinforcement learning-based method to solve the incentive problem, thereby avoiding the impact of information asymmetry. Simulation results verify that the proposed scheme not only balances the demands of all parties but also enhances user participation without being affected by the number of participants, making it suitable for IoV environments.

Federated learning (FL), as a distributed machine learning paradigm, facilitates collaborative training without sharing raw data and holds promise for effective application in the Internet of Vehicles (IoV) for tasks, such as traffic flow prediction and driving behavior analysis. However, the efficiency of FL systems relies on the integrity of the local dataset and the level of user contribution. Vulnerabilities to attacks by malicious users and suboptimal aggregation methods can compromise system performance. To address these issues, this article proposes a blockchain-based FL secure aggregation algorithm to bolster FL robustness. Specifically, in the absence of a centralized trust authority in the IoV, we establish a hierarchical blockchain-empowered IoV reputation management framework that leverages smart contracts to create a trustworthy environment for reputation sharing. Additionally, a lightweight consensus protocol tailored for blockchain efficiency is proposed, thus facilitating a flexible and effective implementation of FL in the IoV. Furthermore, we introduce a reputation-based model selection evaluation scheme and, based on this, a robust FL secure aggregation algorithm. This novel reputation assessment strategy mitigates the effects of interaction uncertainties and integrates a broader spectrum of IoV-specific reputation determinants, thereby enhancing the precision of model selection. The simulation results validate the proposed framework’s superiority in terms of robustness, adaptability, and security.

For the Internet of things (IoT) network, the integrated localization and communication (ILAC) is expected to provide high localization and communication performance simultaneously. However, the existing research to evaluate the performance of ILAC systems fails to reveal the fundamental performance of ILAC systems in practical IoT network topology analytically. In this paper, we develop a unified analytical ILAC framework using stochastic geometry. We then validate the theoretical results obtained from the proposed analytical framework with the simulation results via extensive Monte Carlo simulations. We further analyse the communication coverage and localization coverage probability with respect to the network density, time-frequency-power domain resource allocation, and communication throughout and localization threshold. Finally, based on the ILAC simulation results, we reveal design guidance for ILAC systems. Specifically, we observe the fundamental trade-off between localization and communication performance attributed to the time-frequency-power domain resource allocation. Network density positively affects the ILAC performance, while power control is much less effective due to the dense network topology. The major observations are that time-domain (TD) resource allocation is preferred in dense networks with low localization CRB thresholds, while frequency-domain (FD) resource allocation dominates in sparse networks with large localization CRB thresholds.

Artificial intelligence generated content (AIGC) technologies, with a predominance of large language models (LLMs), have demonstrated remarkable performance improvements in various applications, which have attracted great interests from both academia and industry. Although some noteworthy advancements have been made in this area, a comprehensive exploration of the intricate relationship between AIGC and communication networks remains relatively limited. To address this issue, this article conducts an exhaustive survey from dual standpoints: first, it scrutinizes the integration of LLMs and AIGC technologies within the domain of communication networks and second, it investigates how the communication networks can further bolster the capabilities of LLMs and AIGC. Additionally, this research explores the promising applications along with the challenges encountered during the incorporation of these AI technologies into communication networks. Through these detailed analyses, our work aims to deepen the understanding of how LLMs and AIGC can synergize with and enhance the development of advanced intelligent communication networks, contributing to a more profound comprehension of next-generation intelligent communication networks.

The Industrial Internet of Things (IIoT), a core driver of Industrial 4.0, is considered as one of the most promising revolutionary technologies propelling the evolution of smart manufacturing towards Specialization, Reinforcement, Distinctiveness, and Innovation. The security and efficiency of smart manufacturing depend on the secure and efficient interaction of massive production data among entities. Yet, as a crucial measure of securing entity interactions, current authentication mechanisms overlook the single-point-of-failure issue and lightweight design. Moreover, interaction efficiency is rarely optimized and enhanced from the perspective of communication-supporting nodes. Paramountly, the assurance and optimization of entity interaction security and efficiency are strongly coupled, which is not considered in existing interaction frameworks. This paper designs a three-layer entity interaction framework based on mobile edge computing (MEC) and blockchain technology. Specifically, the double-layer blockchain and MEC-cluster assisted lightweight authentication (BCLA) mechanism is proposed under the three-layer framework to achieve lightweight entity authentication in a weakly centralized manner. To optimize the entity interaction efficiency from joint authentication and transmission, this paper further proposes an industrial edge server (IES) deployment optimization scheme and the proximity policy optimization based IES deployment (PAID) algorithm. The security features and efficiency of the three-layer framework are demonstrated by carrying out security analysis and performance evaluation, which is based on the Hyperledger Fabric platform.

The management of computing resources through the computing power network (CPN) has gradually become a focal point of research. With the development of the 6th generation (6G) mobile networks, some promising technologies, such as satellite-terrestrial integrated network (STIN) and smart endogenous network driven by artificial intelligence (AI) are increasingly being applied in Industrial Internet of Things (IIoT). However, several issues in current studies are worthy of attention: 1) the large number of devices powered by battery in IIoT; 2) the complex communication environments; and 3) the finite computing resources for task data processing. To cope with these challenges, a satellite-terrestrial integrated CPN (STICPN) framework is introduced in this article. Within this framework, a task offloading link selection scheme is proposed, which minimizes the delay and the consumption of energy. The task offloading optimization problem is modeled as a markov decision process (MDP). Meanwhile, deep reinforcement learning (DRL) algorithm is employed to adapt to the dynamic states of environment. Specifically, a Dueling Double Deep Q Network (D3QN) is used to make optimal decisions and delay as well as energy consumption can be reduced significantly. Moreover, the D3QN-based scheme extends the usage time of IIoT devices. The simulation results indicate that the proposed scheme outperforms the comparison schemes significantly.

Traditional visual-inertial Simultaneous Localization and Mapping (SLAM) systems predominantly rely on feature point matching from a single robot to realize the robot pose estimation and environment map construction. However, in complex scenarios, these traditional systems struggle with issues, such as tracking failures due to illumination changes, rapid movements, and low-texture environments, and they perform poorly in terms of mapping efficiency and global consistency. To address these challenges, we propose a centralized collaborative SLAM system that employs both point and line features for tracking in the robot and map fusion in the cloud. The proposed system leverages the fusion of point and line features across all instances in the process, which allows our method to achieve higher localization accuracy in structured, low-texture scenes. With the aid of classifying gravity-aligned vertical lines and spatial parallel lines, the proposed system can deliver faster and more accurate odometry in complex scenes. Furthermore, we developed intrarobot and interrobot loop closure detection methods based on point and line features, generating a globally consistent sparse point cloud and structured scene map in the cloud. Our method is able to build richer maps while improving accuracy compared to existing methods. Experimental results on public datasets and in real-world environments show that, compared to existing advanced methods, our approach demonstrates better performance.

With the rapid advancement of technologies such as artificial intelligence and the Internet of Things, video surveillance—being a core component of video IoT systems—has been widely adopted for underground coal mine safety monitoring. However, the dim lighting and heavy coal dust in underground mines result in poor visibility and significant detail loss in monitoring images, posing a major challenge to coal mine safety management. To address these issues, we propose a low-light image enhancement method tailored for underground coal mine environments, based on Zero-DCE. In our method, traditional convolutions are replaced with Ghost modules to reduce computational cost while maintaining feature extraction capability. Additionally, we incorporate global context blocks and a Vision Transformer branch to integrate more global information into the model. Specifically, the global context blocks improve the model’s ability to correct uneven illumination and prevent overexposure. Meanwhile, the Vision Transformer branch captures long-range dependencies and fuses local and global features to enhance brightness while mitigating color distortion. Furthermore, we replace the original quadratic iterative function with a reciprocal illumination mapping function, enabling more stable and perceptually aligned brightness adjustments. Experimental results on the coal mine underground personnel dataset demonstrate that our method outperforms several state-of-the-art low-light enhancement techniques, achieving superior results in both qualitative and quantitative evaluations. These findings indicate that our approach significantly improves the visibility and overall quality of underground coal mine monitoring images.

With the rapid development of Visual Internet of Things (VIoT) and text detection technology, they have been widely combined and applied to many industrial production sites, such as label text detection, achieving impressive results. However, there are still many shortcomings in the text detection technology: 1) the existing VIoT system has very limited detection precision for text with large scale changes, especially for some small-scale text detection; 2) the existing text detection algorithms cannot meet the actual situation, as the labels often contain handwritten texts, and the text to be detected is arbitrary shape; and 3) in the actual detection, there are many creases or defects on the text label. To solve the above problems, this article designs a text detection method based on a multiscale selection fusion feature pyramid and multisemantic spatial network (MSNet) to assist the VIoT system in detecting label text. First, a multiscale selective fusion feature pyramid is designed, which not only uses the texture extraction module to effectively improve the text texture feature and multiscale feature extraction ability, but also uses the cross-scale selective fusion block to selectively fuse the features of different stages to reduce the influence of pollution on detection. In addition, a MSNet is designed to capture the multisemantic spatial information of each feature channel by using the multiscale deep shared 1-D convolution, which effectively integrates global context dependence and multisemantic spatial prior. Experimental results show that the comprehensive index F-measure on the public datasets ICDAR2015, total-text, and CTW1500 is increased by 5.7%, 3.3%, and 3.8%, respectively. Furthermore, the precision, recall, and F-measure on the dataset label-text are 94.6%, 90.7%, and 92.6%, respectively. The label text detection VIoT system we designed has been deployed in the field and achieved excellent performance. The code of our proposed method can be found in:https://github.com/rebornone1/MSNet

The explosive growth in video services has significantly strained current mobile network infrastructure, leading to spectrum scarcity, backhaul congestion, and degraded quality of experience. While edge caching has emerged as a promising solution to address these challenges and deliver seamless video playback experience, multiversion edge caching for heterogeneous clients remains challenging due to varying client requirements and network conditions. This article proposes a transcoding-enabled edge caching framework for mobile edge-cloud computing networks. Specifically, we combine video transcoding with edge caching to support both “direct cache hits” and “soft cache hits” for reducing transmission latency. We model this joint caching and resource allocation problem as a Stackelberg game to minimize video transmission latency. To solve this problem, we develop a novel dual timescale model agnostic meta-learning (MAML)-based Stackelberg game (DTMSG) optimization approach that determines the delay-optimal Stackelberg equilibrium (SE) and accelerates convergence. Simulation results demonstrate that our DTMSG optimization algorithm efficiently converges to the SE point, maximizing the utility function of the MVNO and BSs while reducing the average video transmission delay.

Clear monitoring images are crucial for the safe operation of coal mine Internet of Video Things (IoVT) systems. However, low illumination and uneven brightness in underground environments significantly degrade image quality, posing challenges for enhancement methods that often rely on difficult-to-obtain paired reference images. Additionally, there is a tradeoff between enhancement performance and computational efficiency on edge devices within IoVT systems.To address these issues, we propose a multimodal image enhancement method tailored for coal mine IoVT, utilizing an ISP operations within a differentiable CNN framework fusion architecture optimized for uneven illumination. This two-stage strategy combines global enhancement with detail optimization, effectively improving image quality, especially in poorly lit areas. A contrastive language-image pretraining (CLIP)-based multimodal iterative optimization allows for unsupervised training of the enhancement algorithm. By integrating traditional image signal processing (ISP) with convolutional neural networks (CNN), our approach reduces computational complexity while maintaining high performance, making it suitable for real-time deployment on edge devices. Experimental results demonstrate that our method effectively mitigates uneven brightness and enhances key image quality metrics, with preservation of original visual information (PSNR) improvements of 2.9%–4.9%, structural similarity (SSIM) by 4.3%–11.4%, and visual information fidelity (VIF) by 4.9%–17.8% compared to seven state-of-the-art algorithms. Simulated coal mine monitoring scenarios validate our method’s ability to balance performance and computational demands, facilitating real-time enhancement and supporting safer mining operations.

Wearable device-based Human Activity Recognition (HAR) has attracted considerable interest with the rapid development of the Internet of things (IoT), and Federated Learning (FL) has been widely adopted in this domain for its ability to collaboratively train models across decentralized devices while preserving privacy. However, its performance is hindered by data heterogeneity arising from variations in the placement of the wearable devices, user behaviors, and physiological characteristics. In this work, we present LDA-FedHAR, a federated HAR framework designed for wearable devices by capturing more common knowledge from aligned client HAR data. It performs Local HAR Data Alignment (LDA) on each client, which is an entirely on-device alignment method that operates independently on local HAR data. By computing the transformation matrix solely from local HAR data and applying it to the data itself, LDA projects heterogeneous client data into a unified space, thereby reducing inter-client discrepancies at the source. To further enhance efficiency and robustness, we propose two IMU-specific variants, LDA(S-IMU) and LDA(C-IMU), which explore intra-and inter-IMU correlations based on practical placements of wearable devices. Experiments are conducted on 4 public HAR datasets: HHAR, Shoaib2014, OPPORTUNITY++, and PAMAP2. The results show that LDA effectively reduces inter-client discrepancies, and LDA-FedHAR along with its variants consistently outperforms state-of-the-art FL methods. Moreover, the improvements achieved by integrating LDA into other FL methods highlight its applicability.

Perceiving scene depth and 3-D structure is one of the key tasks for Internet of Video Things (IoVT) devices to understand and interact with the environment. Self-supervised monocular depth estimation has demonstrated significant potential in leveraging large-scale unlabeled datasets to achieve competitive performance, thereby playing an increasingly important role in depth estimation. Despite recent methods providing additional supervisory signals through self-distillation strategies to improve depth estimation, an effective method for generating pseudo-depth labels suitable for addressing occlusion issues among elements far from the camera remains unexplored. To address this limitation, we propose a patch-based self-distillation learning framework to exploit the potential of self-supervised monocular depth estimation in recovering fine-grained scene depth. In the proposed framework, elements far from the camera within the input image are enlarged by enlarging and cropping operations in the patch-based self-distillation branch. Guided by photometric consistency, the model learns the detailed occlusion relationships among elements from the enlarged patches, producing patch depth maps with fine structures. In the main branch, which takes full-scale images as input, patch depth maps serve as pseudo-depth labels through self-distillation loss to provide additional supervisory signals for regions where photometric consistency fails to offer effective supervision. This forces the depth estimation network to recover fine structures of elements far from the camera in full-scale input images. Regarding the architecture of the depth estimation network, we introduce a bin-center prediction. In this prediction, a global aggregator based on self-attention provides additional scene structure queries for adaptive scene depth discretization. Finally, to encourage the model to explore more general cues for depth inference beyond road plane cues, we propose a PatchMix data augmentation method to enhance the model’s generalization ability to unseen scenes. Extensive experiments on the KITTI dataset show that the proposed method significantly improves performance over the baseline, particularly in fine-grained scene depth estimation. Moreover, the model also exhibits good generalization performance when transferred to the Make3D and Cityscapes datasets.

Speech Emotion Recognition (SER) is important in improving human-computer interaction. Cross-Linguistic SER (CLSER) has been a challenging research problem due to significant variability in the linguistic and acoustic features of different languages. In this study, we propose a novel approach,HuMP-CAT, which combines HuBERT (Hidden Unit BERT), MFCC (Mel-Frequency Cepstral Coefficients), and Prosodic characteristics. These features are fused using a cross-attention transformer (CAT) mechanism during feature extraction. Transfer learning is applied to gain from a source emotional speech dataset to the target corpus for emotion recognition. We use IEMOCAP as the source data set to train the source model and evaluate the proposed method on seven data sets in five languages (i.e., English, German, Spanish, Italian, and Chinese). We show that, by fine-tuning the source model with a small portion of speech from the target datasets,HuMP-CATachieves an average accuracy of 78.75% across the seven datasets, with notable performance of 88.69% on EMODB in German language and 79.48% on EMOVO in Italian language. Our extensive evaluation demonstrates thatHuMP-CAToutperforms existing methods across multiple target languages.

Effective sensing capabilities are crucial for the safe and reliable operation of Connected and autonomous vehicles (CAVs). While traditional approaches focus on enhancing onboard sensors, the integration of road sensor networks (RSNs) into the CAV ecosystem presents a promising solution to improve sensing performance, but also introduces significant challenges, including heterogeneous sensing requirements, inconsistencies in multisource sensor data, and efficient resource utilization. To address these challenges, this article proposes a novel cooperative sensing framework that leverages multisource and multilevel sensing information from RSNs to optimize CAV sensing performance in resource-constrained scenarios. We develop an improved decision transformer (DT)-based approach that dynamically adapts to diverse driving conditions and efficiently fuses sensor data at various abstraction levels. To tackle the issue of long-delayed rewards, we introduce a reshaped reward function and a bi-level optimization framework that enables effective propagation of rewards along decision sequences. An advanced gradient approximation technique is employed to efficiently solve the optimization problem. Extensive simulations demonstrate the superior performance of our improved DT approach compared to state-of-the-art reinforcement learning (RL) methods in terms of sensing accuracy, coverage, and data efficiency under various traffic conditions.

With the rapid development of the Internet of Things (IoT), train control systems have emerged as a successful application scenario. The virtually coupled train set (VCTS), as a new paradigm for train control, relies on more efficient vehicle-to-vehicle and vehicle-to-ground communication to achieve closer train spacing. This enhanced communication allows trains to capture more complex and detailed state information. However, traditional train control algorithms, limited by their data processing capabilities, often cannot fully utilize this additional information, leading to conservative control strategies to ensure safety and stability. Generative Artificial Intelligence (GAI), particularly generative diffusion models, has recently shown great potential in optimizing IoT scenarios by handling more complex environments. This article proposes a GAI-based control algorithm framework that leverages diffusion models to optimize train trajectories. By integrating the extensive real-time data generated by IoT systems, the GAI-driven approach enhances decision-making processes, offering more precise and adaptive control strategies tailored to the demands of VCTS. This framework demonstrates the potential of combining IoT data with GAI to achieve higher control accuracy, ensuring safety and performance in dynamic and complex urban rail transit scenarios. Experimental results validate the effectiveness of the proposed method, highlighting its robustness and adaptability across various conditions.

Human motion video generation has garnered significant research interest due to its broad applications, enabling innovations such as photorealistic singing heads or dynamic avatars that seamlessly dance to music. However, existing surveys in this field focus on individual methods, lacking a comprehensive overview of the entire generative process. This paper addresses this gap by providing an in-depth survey of human motion video generation, encompassing over ten sub-tasks, and detailing the five key phases of the generation process: input, motion planning, motion video generation, refinement, and output. Notably, this is the first survey that discusses the potential of large language models in enhancing human motion video generation. Our survey reviews the latest developments and technological trends in human motion video generation across three primary modalities: vision, text, and audio. By covering over two hundred papers, we offer a thorough overview of the field and highlight milestone works that have driven significant technological breakthroughs. Our goal for this survey is to unveil the prospects of human motion video generation and serve as a valuable resource for advancing the comprehensive applications of digital humans.

In untrimmed video tasks, identifying temporal boundaries in videos is crucial for temporal video grounding. With the emergence of multimodal large language models (MLLMs), recent studies have focused on endowing these models with the capability of temporal perception in untrimmed videos. To address the challenge, in this paper, we introduce a multimodal large language model named MLLM-TA with precise temporal perception to obtain temporal attention. Unlike the traditional MLLMs, answering temporal questions through one or two words related to temporal information, we leverage the text description proficiency of MLLMs to acquire video temporal attention with description. Specifically, we design a dual temporal-aware generative branches aimed at the visual space of the entire video and the textual space of global descriptions, simultaneously generating mutually supervised consistent temporal attention, thereby enhancing the video temporal perception capabilities of MLLMs. Finally, we evaluate our approach on both video grounding task and highlight detection task on three popular benchmarks, including Charades-STA, ActivityNet Captions and QVHighlights. The extensive results show that our MLLM-TA significantly outperforms previous approaches both on zero-shot and supervised setting, achieving state-of-the-art performance.

Human emotion synthesis is a crucial aspect of affective computing. It involves using computational methods to mimic and convey human emotions through various modalities, with the goal of enabling more natural and effective human-computer interactions. Recent advancements in generative models, such as Autoencoders, Generative Adversarial Networks, Diffusion Models, Large Language Models, and Sequence-to-Sequence Models, have significantly contributed to the development of this field. However, there is a notable lack of comprehensive reviews in this field. To address this problem, this paper aims to address this gap by providing a thorough and systematic overview of recent advancements in human emotion synthesis based on generative models. Specifically, this review will first present the review methodology, the emotion models involved, the mathematical principles of generative models, and the datasets used. Then, the review covers the application of different generative models to emotion synthesis based on a variety of modalities, including facial images, speech, and text. It also examines mainstream evaluation metrics. Additionally, the review presents some major findings and suggests future research directions, providing a comprehensive understanding of the role of generative technology in the nuanced domain of emotion synthesis.

In this paper, we address the problem of the operator’s economic profit maximization in a multi-access edge computing (MEC)-enabled time division multiple access (TDMA)-based air-ground integrated networking (AGIN) network. We consider to optimize task placement and replacement, unmanned aerial vehicle (UAV) placement, UAV flight time, access control, and task offloading ratios in user devices (UDs) and the UAV. The optimization is constrained by storage capacity, task processing quality of service (QoS) requirements, and TDMA requirements, etc. Our optimization is conducted in two time scales. Task placement and replacement are performed in a coarse-grained time scale (frame), while other optimizations are conducted in a fine-grained time scale (time slot). Due to the high dynamics of the environment, finding a solution is challenging. To address this problem, we present a hierarchical deep reinforcement learning (DRL) algorithm. The high-level component is a deep Q network (DQN) agent responsible for obtaining task placement and replacement solutions within a frame. The low-level component is an improved deep deterministic policy gradient (IDDPG) agent, which is used to address task processing-related issues within a time slot. Our simulations illustrate that the proposed algorithm has good performance in economic profit maximization compared with other algorithms.

The communication network in a tunnel construction site facilitates real-time data exchange, and serves as a backbone for successfully executing construction projects. However, the long and closed spaces, irregular surfaces, and variable topology as tunnel excavation impose rigorous limitations on signal propagation, communication quality and coverage. To alleviate the realistic issues, we introduce a holistic three-phase cooperative relay scheme based on 5G New Radio (NR) vehicle-to-everything (V2X) architecture, which can extend the communication range and enhance network throughput. We theoretically derive the outage probability of the entire cooperative relaying process from source to destination, and quantify the impact of relaying on construction workflow with relay cost. To minimize the outage probability and relay cost, we formulate a cooperative relay strategies optimization problem and transform the solving procedure into a Markov decision process (MDP). We design a hybrid hierarchical proximal policy optimization (HH-PPO) reinforcement learning method to solve the MDP, which consists of two discrete actor networks, two continuous actor networks, and two critic networks. The hybrid structure enables HH-PPO to tackle the mixed action space, and the hierarchical structure enables adaptive and contextual actions generation by integrating the discrete network outputs into the continuous actor network. Simulation results validate the effectiveness of the HH-PPO algorithm with faster convergence speed, and show superior performance in terms of lower, stable outage probability and relay cost satisfaction compared with another benchmark.

In this paper, we investigate the fundamental tradeoff between computation and communication for the full-duplex (FD) wireless MapReduce distributed computing network. Specifically, a coded interference alignment and neutralization (CIAN) scheme is proposed to significantly reduce the achievable normalized delivery time (NDT) for any given computation load, which jointly exploits both the coding and interference management technologies. In particular, a novel coding strategy is designed to create the coded message desired by multiple nodes, thereby providing the coded multicasting gain. Furthermore, the Shuffle phase is molded as a special cooperative X-multicast network. For this network, a novel IAN scheme is proposed to improve the achievable sum degree of freedom (SDoF), thereby providing the IAN gain. In the proposed CIAN scheme, the fundamental tradeoff between the coded multicasting gain and IAN gain is characterized, and the achievable NDT is minimized by carefully optimizing these two gains. Furthermore, a tight information-theoretic lower bound on the NDT is derived, demonstrating the optimality of the CIAN scheme in some cases. In other cases, the achievable NDT of the CIAN scheme and the lower bound are within a multiplicative gap of 2. Theoretical analysis and numerical results indicate the superior performance of the CIAN scheme compared to existing schemes, particularly by providing additional coded multicasting gain and improved IAN gain.

Existing multi-view classification and clustering methods typically improve task accuracy by leveraging and fusing information from different views. However, ensuring the reliability of multi-view integration and final decisions is crucial, particularly when dealing with noisy or corrupted data. Current methods often rely on Kullback-Leibler (KL) divergence to estimate uncertainty of network predictions, ignoring domain gaps between different modalities. To address this issue, KPHD-Net, based on Hölder divergence, is proposed for multi-view classification and clustering tasks. Generally, our KPHD-Net employs a variational Dirichlet distribution to represent class probability distributions, models evidences from different views, and then integrates it with Dempster-Shafer evidence theory (DST) to improve uncertainty estimation effects. Our theoretical analysis demonstrates that Proper Hölder divergence offers a more effective measure of distribution discrepancies, ensuring enhanced performance in multi-view learning. Moreover, Dempster-Shafer evidence theory, recognized for its superior performance in multi-view fusion tasks, is introduced and combined with the Kalman filter to provide future state estimations. This integration further enhances the reliability of the final fusion results. Extensive experiments show that the proposed KPHD-Net outperforms the current state-of-the-art methods in both classification and clustering tasks regarding accuracy, robustness, and reliability, with theoretical guarantees.

With the rapid evolution of vehicular network technology, the integration of Machine Learning (ML) with Connected and Autonomous Vehicles (CAVs) presents both remarkable opportunities and formidable challenges. This paper addresses the crucial need for efficient ML model training in the context of Federated Learning (FL) within vehicular networks. Recognizing the limitations imposed by the tradeoff between the high energy cost at the local level with the performance problem at the global level, we propose an innovative approach that harmonizes cost optimization with strategic model selection. Our strategy primarily focuses on optimizing energy consumption during model training and updating at the vehicle end, thereby resolving the prevalent issue of limited end-user participation in FL due to high energy demands. Additionally, we introduce an advanced model selection method, prioritizing local model uploads and adaptively allocating bandwidth to clients with more extensive training data. This method enhances the efficiency and reliability of model updates, ensuring robust global model performance. We validate our approach through extensive simulations, demonstrating not only improved learning performance but also a significant reduction in energy consumption among participating clients.

Vehicular cooperative perception (VCP) facilitates the exchange of sensing data among vehicles through vehicle-to-everything (V2X) communication, significantly increasing the sensing range and precision of individual autonomous vehicles (AVs). However, efficiently managing the sharing and processing of large volumes of sensing data presents challenges due to restricted communication and computation resources. This study introduces an integrated sensing, communication, and computation (ISCC)-based task offloading and resource allocation (ITORA) framework, which optimizes cooperative perception by determining what data to share, which vehicles to involve, and how to process the data effectively. We develop an information value function to evaluate the data quality for each vehicle. Subsequently, we design strategies for sensing task allocation, task offloading, and resource allocation to enable value-driven data selection at a subregion level, facilitating collaborative computing among edge servers and vehicles. Additionally, we formulate an optimization problem aimed at maximizing information value while minimizing delay and energy consumption, subject to constraints on a full region of interest (RoI) coverage, delay, wireless bandwidth, and computational resources. We decompose the mixed-integer nonlinear programming (MINLP) problem into two subproblems, devising a sensing task allocation algorithm and a proximal policy optimization (PPO)-based task offloading and resource allocation (PTORA) algorithm to address them. Comprehensive simulations validate the effectiveness of the proposed PTORA in optimizing information value, reducing task execution delay, and minimizing energy consumption.

Connected and Autonomous Vehicles (CAVs) increasingly leverage sophisticated sensor systems integrated with emerging technologies like the sixth generation (6G) for enhancing driving safety and efficiency. Despite the potential of utilizing the advanced communication technology to enhance driving reliability, conficts between high-quality sensing needs of CAVs and insufficient sensing sharing wellness present significant challenges. To bridge the gaps, this paper proposes a novel cooperative sensing sharing framework that utilizes vehicle-to-vehicle (V2V) communication to extend the effective sensory range of CAVs, thereby reducing perception gaps and enhancing driving efficiency in complex driving environments. Aiming at incentivizing cooperative sensing sharing ecosystem within this decentralized framework, we first introduce a multi-tier blockchain architecture that ensures secure and transparent data sharing among CAVs. Concurrently, a custom-designed efficient consensus algorithm is proposed to minimize the overhead while guaranteeing transaction throughput. To tackle issues related to trust and motivate long-term cooperative behavior, we introduce a supervision-oriented model that utilizes evolutionary game to formulate incentive mechanisms discouraging malicious participation and promoting honest, active engagement. Finally, theoretical analysis and extensive simulations demonstrate that our system not only ensures healthy sensing sharing performance but also maintains system security.

With the advancement of urbanization, communication-based train control (CBTC) systems for urban rail transit and train-to-train (T2T) communication have garnered significant attention. T2T communication establishes mobile ad hoc networks (MANETs), similar to those in vehicular ad hoc networks (VANETs). Building upon this foundation, we propose an adaptive cooperative (ADCO) MAC protocol for T2T communication. The scheme introduces clustering and cooperative transmission mechanisms, which enhance the efficiency and reliability of safety packet transmission. Additionally, the protocol assigns different priorities to packets engaging in contention on the control channel (CCH) and enables trains to access service channels (SCHs) without contention through pre-reserved time slots. To analyze the transmission probabilities and success rates of packets with varying priorities, a Markov-based model is utilized, ultimately determining the optimal ratio of the CCH interval (CCHI) to the SCH interval (SCHI) for maximizing channel utilization. Theoretical analysis and simulation results demonstrate that the proposed MAC protocol ensures reliable transmission of safety packets while simultaneously optimizing the throughput on SCHs.

Currently, heterogeneous driving scenarios and complex traffic conditions challenge connected and autonomous vehicles (CAVs) to achieve accurate sensing of road conditions. Existing research on the sensing capabilities of vehicles only relys on adding more onboard sensors, which makes the driving safety unable to be guaranteed due to the installment and cost limit of various sensors. Therefore, this paper proposes an enhanced 3D sensor deployment method to break through the sensing capabilities of CAVs’ own equipment limitations. By efficiently utilizing road infrastructure, reasonable roadside sensor deployment will effectively assist CAVs to expand the sensing range and improve overall sensing accuracy. Firstly, in order to address the limitations of existing works that often rely on simplified sensor models and idealized road conditions, we propose a Bresenham-based sensor and environment model that can be used to construct realistic road environments. Secondly, a decision transformer (DT)-based method is adopted to solve the problem of optimal deployment of sensors in road environments. Our approach effectively addresses the limitations of traditional static deployment methods, which often fail to consider the complexities of real-world driving conditions and the diverse factors influencing optimal sensor deployment. Finally, in order to solve the problem of DT delayed rewards, we propose a two-layer optimization method to redistribute the reward function to solve the challenge of local optimization. A large number of simulations oriented to sensing effects not only verify the effectiveness of the sensor deployment method but also ensure the reliability of sensing assistance.

Collaborative resource allocation is crucial for reducing overhead and enhancing resource utilization in edge computing-enabled networks. To ensure a satisfactory user experience, we recognize the importance of considering information uncertainty in resource allocation. Therefore, we explore information uncertainty in edge computing-enabled networks, especially within the complex environment of resource coupling. However, existing methods lack a comprehensive and robust solution for coordinating wireless, transport, and computing resource under this information uncertainty. This paper addresses this gap by proposing a joint optimization of access point (AP) selection, computing node association, and traffic engineering, aiming to maximize network utility under the uncertain conditions of wireless status and application QoS requirements. The constraints under these uncertainties are modeled as chance constraints, complicating the problem's solvability. We adopt the Bernstein approximation to establish convex conservative approximations of the chance constraints. Given the problem's substantial size and computational complexity, the alternating direction method of multipliers is employed to solve the approximated problem in a distributed manner. We further derive the closed solutions of the corresponding sub-problems. Extensive simulations validate the superiority of our proposed scheme, demonstrating its ability to achieve a good trade-off between meeting user requirements and optimizing resource utilization.

Virtually Coupled Train Set (VCTS) is an advanced train control technology enabling multiple trains to operate closely through wireless communication, enhancing capacity and operational flexibility. Traditional VCTS control algorithms struggle with complex dynamic models and local optimality, hindering real-time, long-term optimization. This paper proposes an Edge Intelligence (EI) enhanced Monte Carlo Tree Search (MCTS) framework for VCTS Optimal Control (M-VOC). MCTS is a heuristic search algorithm that identifies optimal operational solutions efficiently, focusing on long-term stability over local optimums. EI supports MCTS for real-time decision-making, and we introduce a model-based reinforcement learning algorithm to manage VCTS's complex dynamics. Our framework addresses VCTS control issues in real-time while optimizing long-term benefits. To meet computational and real-time demands, we propose a train-to-edge cooperative computing strategy using multi-intelligence reinforcement learning. Simulations demonstrate that our EI-enhanced MCTS strategy effectively provides cooperative control, ensuring virtually coupled trains operate safely, stably, and punctually with reduced intervals.

Tile-based streaming and super resolution (SR) are two representative technologies adopted to improve bandwidth efficiency of 360° video streaming. The former allows selective downloading of contents in the user viewport by splitting the video into multiple independently decodable tiles. The latter leverages client-side computation to enhance the received video to higher quality using advanced neural network models. In this work, we propose a Collaborated Streaming and Enhancement (CSE) adaptation framework for mobile 360° videos, which integrates super resolution with tile-based streaming to optimize the user experience with dynamic bandwidth and limited computing capability. To effectively enhance the tile-based video streaming through SR, we propose to adaptively group the tiles for quality enhancement adapting to the content similarity. We also identify and address several key design issues to integrate SR into tile-based video streaming including unified video quality assessment, computational complexity model for super resolution, and buffer analysis considering the interplay between transmission and enhancement. We further formulate the quality-of-experience (QoE) maximization problem for mobile 360° video streaming and propose a rate adaptation algorithm to make the best decisions for download and for enhancement based on the Lyapunov optimization theory. Extensive evaluation results validate the superiority of our proposed approach, which demonstrates stable performance with considerable QoE improvement, while enabling a trade-off between playback smoothness and video quality.

Currently, beamforming antenna array technologies are of utmost importance in 5G communication systems. These technologies are essential for optimizing the coverage and signal quality of the cellular network. However, the optimization of broadcast beams presents significant challenges due to the complex strategy profile space. Each beam can be configured with different widths and heights, making it difficult for conventional algorithms to handle. To address this issue, we propose a novel approach called Multi-Agent Moth-Flame Reinforcement Learning (MAMF-RL) algorithm for broadcast beam optimization. MAMF-RL combines reinforcement learning and moth-flame optimization algorithms to interactively search for the optimal broadcast beams. By decomposing the problem into multiple single-sector antenna configuration problems, MAMF-RL effectively reduces the algorithm complexity. We conducted experiments utilizing real data in an 18-sector wireless coverage area. To evaluate the performance of our proposed method, we compared it with traditional methods such as the particle swarm algorithm. The results demonstrate that our MAMF-RL model achieves an average coverage rate of 1.82% higher and a 13.74% lower overlapping coverage rate compared to traditional methods.

Wireless Traffic Prediction (WTP) plays a significant role in achieving intelligent resource management for communication systems. However, WTP still faces challenges such as inaccurate prediction resulting from the complex spatial-temporal characteristics due to user mobility, high communication overhead caused by the complexity of the prediction model, and user privacy issues stemming from Centralized Learning (CL). To address the aforementioned issues, this paper proposes a WTP framework under the Federated Learning (FL) strategy called Federated Spatial-Temporal Dual-attention based Network (FedSTDN). Aiming at improving communication efficiency and simultaneously representing various wireless traffic patterns, a data augmentation-based clustering algorithm is adopted, which groups cells into different regions using a small augmented dataset, facilitating subsequent processing. To improve prediction performance, a local prediction model based on Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) is proposed to capture the short- and long-term dependencies of traffic. Additionally, a novel Kolmogorov-Arnold Network (KAN) layer is introduced to replace the traditional Multi-Layer Perceptron (MLP) layer, further enhancing prediction performance. Simulations on two different real-world datasets verify the effectiveness and efficiency of FedSTDN. Compared to the well-performing baseline, the proposed FedSTDN achieves up to 32.83% and 24.30% improvements in Mean Square Error (MSE) and Mean Absolute Error (MAE) on the Milan dataset, respectively. For the Trentino dataset, FedSTDN achieves up to 17.25% and 5.86% improvements in MSE and MAE, respectively.

Large Language Models (LLMs), as advanced AI technologies for processing and generating natural language text, bring substantial benefits to the Industrial Internet of Things (IIoT) by enhancing efficiency, decision-making, and automation. Nevertheless, their deployment faces significant obstacles due to high computational and energy demands, which often exceed the capabilities of many industrial devices. To overcome these challenges, edge-cloud collaboration has become increasingly essential, assisting in offloading LLMs tasks to reduce the computational load. However, traditional reinforcement learning (RL)-based strategies for LLMs task offloading encounter difficulties with generalization ability and defining explicit, appropriate reward functions. Therefore, in this paper, we propose a novel framework for offloading LLMs inference tasks in IIoT, utilizing a Decentralized Identifier (DID)-based identity management system for trusted task offloading. Furthermore, we introduce an intelligence-based RL (IRL) approach, which sidesteps the need for defining specific reward functions. Instead, it uses “intelligence” as a metric to evaluate cognitive improvements and adapt to varying environmental preferences, significantly improving generalizability. In our experiments, we employ the GPT-J-6B model and utilize the Human Eval dataset to assess its ability to tackle programming challenges, demonstrating the superior performance of our proposed solution compared to existing methods.

Intelligent transportation systems demand efficient resource allocation and task offloading to ensure low-latency, high-bandwidth vehicular services. The dynamic nature of vehicular environments, characterized by high mobility and extensive interactions among vehicles, necessitates considering time-varying statistical regularities, especially in scenarios with sharp variations. Despite the widespread use of traditional reinforcement learning for resource allocation, its limitations in generalization and interpretability are evident. To overcome these challenges, we propose an Intelligence-based Reinforcement Learning (IRL) algorithm. This algorithm utilizes active inference to infer the real world and maintain an internal model by minimizing free energy. Enhancing the efficiency of active inference, we incorporate prior knowledge as macro guidance, ensuring more accurate and efficient training. By constructing an intelligence-based model, we eliminate the need for designing reward functions, aligning better with human thinking, and providing a method to reflect the learning, information transmission and intelligence accumulation processes. This approach also allows for quantifying intelligence to a certain extent. Considering the dynamic and uncertain nature of vehicular scenarios, we apply the IRL algorithm to environments with constantly changing parameters. Extensive simulations confirm the effectiveness of IRL, significantly improving the generalization and interpretability of intelligent models in vehicular networks.

For visual disbalance defects (VDDs) in low-light images, such as brightness unevenness and color imbalance, existing enhancement methods struggle to extract defect features from local regions and apply adaptive enhancement based on varying degrees of these defects. To address these challenges, we propose an unsupervised multi-modal enhancement method based on a high-order adaptive curve, named CLIP-AE. Specifically, we introduce a multi-modal recurrent optimization approach utilizing contrastive language-image pre-training (CLIP). This method iteratively optimizes variable embedded prompts and an Adaptive Enhancement Module (AEM) to establish dependencies between the prompts and detailed style features in the images, guiding the AEM to perform adaptive image enhancement. Additionally, we implement a progressive feature alignment strategy to enhance the model's ability to perceive style features and improve optimization efficiency by using multiple enhanced images with identical content features and incremental style features. In the AEM, the optimized Hyperparameters Generative Network (HGN) generates the optimal hyperparameters, which drive a High-Dimensional Nested Gamma correction (HDN-Gamma) to perform pixel-wise adaptive enhancement for VDDs. HDN-Gamma further maps pixel values using specific enhancement curves to avoid artifacts. Extensive experiments demonstrate that our method effectively improves visual disbalance defects and reduces artifacts. Compared to seven state-of-the-art algorithms, our method shows significant improvements (PSNR: 16.46%, 16.89%, and 15.14%; SSIM: 9.26%, 8.02%, and 9.85%; MUSIQ: 6.37%, 6.54%, and 7.45%) on the LOL, SICE, and MIT-Adobe FiveK datasets. Our approach offers a novel solution for applying multimedia technology in low-light image enhancement tasks.

Evidence-based deep learning represents a burgeoning paradigm for uncertainty estimation, offering reliable predictions with negligible extra computational overheads. Existing methods usually adopt Kullback-Leibler divergence to estimate the uncertainty of network predictions, ignoring domain gaps among various modalities. To tackle this issue, this paper introduces a novel algorithm based on Hölder Divergence (HD) to enhance the reliability of multi-view learning by addressing inherent uncertainty challenges from incomplete or noisy data. Generally, our method extracts the representations of multiple modalities through parallel network branches, and then employs HD to estimate the prediction uncertainties. Through the Dempster-Shafer theory, integration of uncertainty from different modalities, thereby generating a comprehensive result that considers all available representations. Mathematically, HD proves to better measure the “distance” between real data distribution and predictive distribution of the model and improve the performances of multi-class recognition tasks. Specifically, our method surpasses the existing state-of-the-art counterparts on all evaluating benchmarks. We further conduct extensive experiments on different backbones to verify our superior robustness. It is demonstrated that our method successfully pushes the corresponding performance boundaries. Finally, we perform experiments on more challenging scenarios,i.e., learning with incomplete or noisy data, revealing that our method exhibits a high tolerance to such corrupted data.

In-band network telemetry (INT) is an emerging network measurement technique that offers real-time and fine-grained visualization capabilities for networks. However, the utilization of INT for network measurement introduces additional overheads to the network. The process of data collection consumes extra bandwidth resources, and adjustments to the data collection scheme can impact network stability. Additionally, the INT orchestration scheme requires adaptation to dynamics in the network to improve measurement accuracy. Therefore, striking a balance between accuracy and stability becomes a critical problem. In this paper, our focus lies in the trade-off between measurement accuracy and network stability. We consider the long-term orchestration of multiple telemetry tasks, rationally deploying distinct telemetry tasks to different application flows. To address the challenge, we propose a self-adaptive Dynamic INT Orchestration scheme, D-INTO. Specifically, we formulate a stochastic optimization problem for dynamic INT orchestration. Then we employ Lyapunov optimization to decouple the stochastic optimization problem and use surrogate Lagrangian relaxation to construct a polynomial-time approximation algorithm. Theoretical analysis and experimental results demonstrate that our proposed D-INTO outperforms existing schemes in terms of adaptability to the network dynamics.

Recently, Computing Force Networks (CFNs) have emerged to deeply integrate and flexibly schedule multi-layer, multi-domain, distributed, and heterogeneous computing force resources. CFNs build a resources trading platform between consumers and providers, facilitating efficient resource sharing. Therefore, resources trading is an important issue but it faces some challenges. Firstly, because all kinds of large-scale and small-scale resource providers are distributed in a wide area and the number of consumers is larger compared with edge/cloud computing scenarios, the credibility of consumers and providers is hard to guarantee. Secondly, due to market monopolies by large resource providers, fixed pricing strategies, and information asymmetry, both consumers and providers exhibit a low willingness to engage in resources trading. To solve these challenges, the paper proposes an incentive mechanism for trust-driven resources trading to guarantee trusted and efficient resources trading. We first design a trust guarantee scheme based on reputation evaluation, blockchain, and trust threshold setting. Then, the proposed incentive scheme can dynamically adjust prices and enable the platform to provide appropriate rewards based on providers’ classified types and contributions. We formulate an optimization problem aiming at maximizing the trading platform’s utility and obtaining an optimal contract based on individual rationality and incentive compatible constraints. Simulation results verify the feasibility and effectiveness of our scheme, highlighting its potential to reshape the future of computing resource management, increase overall economic efficiency, and foster innovation and competitiveness in the digital economy.

The Industrial Internet of Things (IIoT) has become a critical technology to accelerate the process of digital and intelligent transformation of industries. As the cooperative relationship between smart devices in IIoT becomes more complex, obtaining deterministic responses of IIoT periodic time-critical computing tasks becomes a crucial and nontrivial problem. However, few current works in cloud/edge/fog computing focus on this problem. This paper is a pioneer in exploring deterministic scheduling and network structural optimization problems for IIoT periodic time-critical computing tasks. We first formulate the two problems and derive theorems to help quickly identify computation and network resource sharing conflicts. Based on this, we propose a deterministic scheduling algorithm,IIoTBroker, which realizes a deterministic response for each IIoT task by optimizing the fine-grained computation and network resources, and a network optimization algorithm,IIoTDeployer, which provides a cost-effective structural upgrade solution for existing IIoT networks. Our methods are illustrated to be cost-friendly, scalable, and deterministic response guaranteed with low computation cost from our simulation results.

In recent years, there has been significant attention in UAV Named Data Networking (UNDN) from both industry and academia. This network paradigm adopts a “request-reply” communication model that allows UAVs to access desired content without the need for specific information regarding the geographical location or IP address of the content producer. This IP-independent design is well-suited for dynamic UAV swarms, but it presents challenges in establishing matching policies between content consumers and producers. This is because that during the distributed decision-making process in content sharing, consumers cannot possess private information regarding producers, and producers may lack the motivation to distribute content. As a result, a revelation and incentive mechanism is needed to be formulated in the system. In this paper, a resource-efficient content-sharing mechanism is proposed to address the aforementioned challenges. First, we propose a contract-based mechanism to incentivize content producers to share content and reveal their private information at the same time. The problem of obtaining the optimal contract is discussed in both cases of information asymmetry and complete information. Then, the Gale-Shapley (GS) algorithm is adopted to make a stable many-to-one matching between content consumers and content producers. The simulation results verify the feasibility, effectiveness and energy efficiency of the proposed mechanism.

In recent years, the integration of mobile edge computing (MEC) and non-orthogonal multiple access (NOMA) has gained significant attention for its potential to reduce energy consumption and offloading latency in future wireless networks. While NOMA can enhance system capacity, accommodating multiple users on the same channel may lead to decoding inaccuracies and reduced offloading accuracy. To tackle these problems, this paper proposes a multi-user offloading model that combines NOMA and orthogonal multiple access (NOMA-OMA) to optimize resource allocation. Users are divided into groups based on their geographical locations, with each group further divided into subgroups. OMA is used within each subgroup, while NOMA is employed between different subgroups to achieve joint multi-user offloading. We divide the optimization problem into two sub-problems, namely power and time allocation between different subgroups and delay allocation within the same subgroup. Closed-form expressions for the two sub-problems are derived. The proposed method achieves optimal system energy consumption while increasing the number of users and maintaining low system complexity. Simulation results demonstrate the effectiveness of the proposed method.

The Digital Twin (DT) represents a rapidly advancing technological innovation within the Industrial Internet of Things (IIoT) domain. DT leverages the power of simulation, machine learning, and data mining to facilitate optimal decision-making for physical objects. However, the creation of a dynamic and living digital counterpart comes at a considerable cost. It requires continuous massive data updating and processing every time the physical object changes. As most data collected by IIoT devices are in their original form, such as images and videos, transmitting such data to remote cloud computing will result in large delays. Furthermore, data processing is often a computationally intensive operation, such as image recognition and video coding, making it impractical to perform processing tasks directly in IIoT devices. To overcome this problem, we introduced the Multi-access/mobile Edge Computing (MEC) architecture to enhance capabilities of DT-enabled IIoT devices. IIoT devices can leverage the extra computing resources in MEC to process raw data, transmitting only the calculation results to update the digital counterpart. To efficiently allocate resources between IIoT devices and MEC, we propose a double auction-based resource allocation scheme. The IIoT devices can purchase computing power from MEC, and an iterative double auction scheme is applied to achieve system efficiency within this market. Furthermore, we propose the Win or Learn Fast Algorithm Policy Hill Climbing (Wolf-PHC) algorithm, which enables agents to improve their strategies continuously through participation in auctions. Simulation results demonstrate that this algorithm accelerates the process of market equilibrium convergence.

Rendering photo-realistic novel-view images of complex scenes has been a long-standing challenge in computer graphics. In recent years, great research progress has been made in enhancing rendering quality and accelerating rendering speed in the realm of view synthesis. However, when rendering complex dynamic scenes with sparse views, the rendering quality remains limited due to occlusion problems. Besides, for rendering high-resolution images on dynamic scenes, the rendering speed is still far from real-time. In this work, we propose a generalizable view synthesis method that can render high-resolution novel-view images of complex static and dynamic scenes in real-time from sparse views. To address the occlusion problems arising from the sparsity of input views and the complexity of captured scenes, we introduce an explicit 3D visibility reasoning approach that can efficiently estimate the visibility of sampled 3D points to the input views. The proposed visibility reasoning approach is fully differentiable and can gracefully fit inside the volume rendering pipeline, allowing us to train our networks with only multi-view images as supervision while refining geometry and texture simultaneously. Besides, each module in our pipeline is carefully designed to bypass the time-consuming MLP querying process and enhance the rendering quality of high-resolution images, enabling us to render high-resolution novel-view images in real-time. Experimental results show that our method outperforms previous view synthesis methods in both rendering quality and speed, particularly when dealing with complex dynamic scenes with sparse views.

With the extensive adoption of information technology, tunnel construction is experiencing a rapid digital transformation. Integrating powerful direct communication among construction equipment (CE) facilitates real-time data exchange, promoting collaborative operations among CE. Concurrent execution of multiple construction procedures leads to a significant rise in the amount of CE and communication links, resulting in resource competition. However, this competition is aimed at enhancing collaboration. To address this inherently contradictory issue, we propose a hierarchical resource management framework and align communication quality of service (QoS) to construction efficiency using construction procedure coherence degree (CPCD) based on age of information (AoI). By formulating resource management as a stochastic optimization problem, a suitable online two-level deep reinforcement learning algorithm referred to as diffusion based soft actor critic (DSAC)-QMIX is designed to derive the radio resource allocation strategies. DSAC is responsible for orchestrating spectrum inter-fleets at the high-level, and QMIX makes the resource management and power control decision for each CE at the low-level. Simulation results validate the effectiveness of the DSAC-QMIX algorithm with comparable transmission rate, and show superior performance in terms of CPCD satisfaction compared with other benchmarks.

With the growing popularity and demand for text-to-video generation applications on mobile devices, resource-constrained mobile terminals struggle to efficiently perform video diffusion inference tasks. Cloud-edge computing networks, with the enhanced computational capabilities, flexibility and spectrum utilization, offers a promising solution for improved performance. Traditional Deep Reinforcement Learning (DRL) based methods have been employed for video diffusion inference tasks. However, existing DRL solutions suffer from low data efficiency, insensitivity to latency, and inability to adapt to task load variations, which degrade the performance. In this paper, we propose a novel deep active inference approach for video diffusion inference task offloading and resource allocation in cloud-edge computing networks. Simulation results demonstrate that our method outperforms mainstream DRL in terms of data utilization efficiency and adaptability to cloud-edge networks, better coping with dynamic task load scenarios.

The cloud computing environment is highly dynamic due to a variety of external factors such as seasonal changes, market trends, and social events. In this case, tenant behavior patterns exhibit significant variability. Therefore, the cloud computing resource allocation model must continuously adapt to these changes. To this end, we propose an adaptive algorithm for dynamic cloud resource allocation based on meta-hierarchical active inference reinforcement learning (MHAIRL). The algorithm combines active inference with meta-hierarchical reinforcement learning. It improves the overall performance of the algorithm, as well as quickly adapts to environmental changes, and improves the generalization performance. In addition, we design a novel polling scheduling framework combined with long short-term memory (LSTM) network. The framework ensures scheduling fairness and flexibility while greatly reducing the state and action space dimensions of the agent. Extensive simulation results show that our method outperforms baseline algorithms in quality of service (QoS) metrics and significantly improves system performance in highly dynamic cloud resource allocation.

Based on quantum parallelism theory and quantum phenomena such as superposition and entanglement, quantum reinforcement learning (QRL) has the potential to surpass classical reinforcement learning (RL). Although some excellent works have been done on QRL, existing RL methods encounter chanllenges when performing in environments with sparse rewards. In this paper, we provide a new perspective on conducting temporal difference (TD) learning in quantum computing, which can eliminate redundant exploration steps compared to classical methods. Specifically, we first use environment information to construct a world model with quantum circuits, enabling it to interact in a quantum way. Then, we perform the learning process by using the quantum world model and Grover’s algorithm to query backwards how to reach the recorded states with high TD-errors. Simulation results show that our proposed method has superior performance compared to classical RL algorithms.

This paper establishes a dual-timescale framework for joint resource slicing and task scheduling in satellite edge computing (SEC) networks. Specifically, to capture network dynamics and task stochasticity at small timescales, we formulate the task scheduling problem as a Markov decision process (MDP) to minimize task delay, network energy consumption, and packet loss. We design a deep reinforcement learning-assisted task scheduling (DRTS) algorithm inspired by the soft actor-critic (SAC) algorithm to learn the scheduling policy. Task processing performance is affected by communication and computing re-sources allocated to respective resource slices. Thus, considering that frequent resource slicing has a significant management over-head, we further optimize resource slices on a larger timescale. To obtain a policy with low complexity, we propose a greedy-based heuristic algorithm. A hierarchical solution is constructed to find the optimal solution due to the correlation between the two timescale problems. Finally, to validate the effectiveness and superiority of the proposed scheme, extensive simulations are performed.

Service Customized Networking (SCN) is emerging as an escalating technological trend to address the personalized requirements of services, which are plagued in traditional “best-effort” transmission networks. Organically coordinating heterogeneous domains in Wide Area Networks (WANs) is essential for establishing end-to-end customized service delivery. However, the peer-to-peer centralized communication mode between Au-tonomous Domains (ADs) hinders their connectivity and makes it challenging to support diverse intra-domain routing protocols for the desired Quality of Service (QoS). To achieve customized service scheduling for predictable QoS in WANs, we propose a contract theory-based incentive mechanism. In specific, we first select trusted ADs with high service qualities by computing their reputations through a subjective logic model. The Service Provider (SP) decomposes the overall QoS requirements by domains logically from a global perspective. These decomposed QoS metrics will splice differentiated service capabilities from ADs to obtain the expected end-to-end connection. To address information asymmetry between the SP and ADs, we formulate contribution-reward contract items and devise an optimization problem of maximizing the whole system utility. The optimal contract problem is solved through constraints of individual rationality and incentive compatibility. Simulation results indi-cate the feasibility and effectiveness of our scheme on service customization and economic benefits.

The rapid growth of edge computing calls for fine-tuned deep neural network (DNN) deployment that emphasizes energy-efficient implementation, due to the resource constraints of edge devices. Traditional Federated Learning-based Neural Architecture Search (FL-based NAS) has been instrumental in the complexities of this deployment, particularly in addressing constraints posed by device heterogeneity, limited resources, and privacy preservation. However, it is hindered by issues such as suboptimal aggregation of homogeneous neural blocks, significant knowledge waste in disregarding heterogeneous neural blocks, and excessive communication energy consumption. This paper introduces F2NAS in green edge computing, a novel energy-efficient approach that addresses these limitations by ensuring flexible and energy-efficient model design and training for edge devices. Firstly, F2NAS introduces an innovative aggregation strategy that enhances the integration of homogeneous neural blocks by using inter-block distances to optimize weight allocation. Further, it employs a unique parameter extraction technique that recaptures valuable insights from previously overlooked heterogeneous neural blocks. Finally, F2NAS meticulously calibrates communication energy consumption by balancing loss function and model interaction, setting and refining an upper limit for model communication. Experimental results reveal F2NAS enhances model accuracy by 2.8% to 4.7%, simultaneously reducing the energy consumption by nearly 50% through optimizing the communication cost.

Personalized motion planning holds significant importance within urban automated driving, catering to the unique requirements of individual users. Nevertheless, prior endeavors have frequently encountered difficulties in simultaneously addressing two crucial aspects: personalized planning within intricate urban settings and enhancing planning performance through data utilization. The challenge arises from the expensive and limited nature of user data, coupled with the scene state space tending towards infinity. These factors contribute to overfitting and poor generalization problems during model training. Henceforth, we propose an instance-based transfer imitation learning approach. This method facilitates knowledge transfer from extensive expert domain data to the user domain, presenting a resolution to these issues. We initially train a pre-trained model using large-scale expert data. Subsequently, during the fine-tuning phase, we feed the batch data, which comprises expert and user data. Employing the inverse reinforcement learning technique, we extract the style feature distribution from user demonstrations, constructing the regularization term for the approximation of user style. In our experiments, we conducted extensive evaluations of the proposed method. Compared to the baseline methods, our approach mitigates the overfitting issue caused by sparse user data. Furthermore, we discovered that integrating the driving model with a differentiable nonlinear optimizer as a safety protection layer for end-to-end personalized fine-tuning results in superior planning performance. The code will be available at https://github.com/LinFunster/PP-TIL.

In 3D object manipulation, collecting expert data for end-to-end imitation learning becomes a mainstream method. Though successful, previous works neglect the guiding role of language in action execution. These methods lack the understanding of action semantics, in which multiple action sequences are guided by a category of instructions, resulting in overlearned object semantics and vague action semantics. To address the above limitation, we introduce a novel framework named LLaKey, which breaks down skill commands into more detailed action instructions based on key states for fine-grained action control. Specifically, LLaKey first leverages the knowledge encoded in pre-trained large-scale models to fine-tune an action instruction conductor. Then, these instructions are executed by a downstream action model. Comprehensive experiments show that LLaKey significantly surpasses baselines with a relative improvement of 15% in nine complex and varied skill tasks, demonstrating the superiority of our method.

Accurate perception of semantic and spatial information is crucial for robots performing language-driven navigation tasks. Existing approaches utilize visual-language models to extract semantic information from the environment and construct maps. However, constrained by the generalization and accuracy of these models themselves, the constructed maps may not be accurate and comprehensive, thereby affecting the accuracy of navigation tasks. Inspired by foundational models’ outstanding classification and segmentation capabilities, this study introduces a semantic map constructed using foundational models. We leverage a foundational model to semantically segment objects in the robot’s video stream and fuse semantics onto the map. Furthermore, this map is used in conjunction with large language models (LLMs) that receive natural language instructions to complete the navigation task. A substantial number of experiments in a simulated environment demonstrate that our method outperforms existing ones in language-driven navigation tasks.

Face swapping, the technique of transferring the identity from one face to another, merges as a field with significant practical applications. However, previous swapping methods often result in visible artifacts. To address this issue, in our paper, we propose CodeSwap, a symmetrical framework to achieve face swapping with high-fidelity and realism. Specifically, our method firstly utilizes a codebook that captures the knowledge of high quality facial features. Building on this foundation, the face swapping is then converted into the code manipulation task in a code space. To achieve this, we design a Transformer-based architecture to update each code independently, which enable more precise manipulations. Furthermore, we incorporate a mask generator to achieve seamless blending of the generated face with the background of target image. A distinctive characteristic of our method is its symmetrical approach to processing both target and source images, simultaneously extracting information from each to improve the quality of face swapping. This symmetry also simplifies the bidirectional exchange of faces in a singular operation. Through extensive experiments on ClelebA-HQ and FF++, our method is proven to not only achieve efficient identity transfer but also substantially reduce the visible artifacts.

First-order optimization (FOO) algorithms are pivotal in numerous computational domains, such as reinforcement learning and deep learning. However, their application to complex tasks often entails significant optimization inefficiency due to their need of many sequential iterations for convergence. In response, we introduce first-order optimization expedited with approximately parallelized iterations (OptEx), the first general framework that enhances the time efficiency of FOO by leveraging parallel computing to directly mitigate its requirement of many sequential iterations for convergence. To achieve this, OptEx utilizes a kernelized gradient estimation that is based on the history of evaluated gradients to predict the gradients required by the next few sequential iterations in FOO, which helps to break the inherent iterative dependency and hence enables the approximate parallelization of iterations in FOO. We further establish theoretical guarantees for the estimation error of our kernelized gradient estimation and the iteration complexity of SGD-based OptEx, confirming that the estimation error diminishes to zero as the history of gradients accumulates and that our SGD-based OptEx enjoys an effective acceleration rate of Θ(√N ) over standard SGD given parallelism of N, in terms of the sequential iterations required for convergence. Finally, we provide extensive empirical studies, including synthetic functions, reinforcement learning tasks, and neural network training on various datasets, to underscore the substantial efficiency improvements achieved by our OptEx in practice.

Urban rail transit is currently moving toward automated driving and mobile occlusion, which raises higher expectations for the train positioning system. Operating in diverse environments like open spaces and tunnels, a single positioning system falls short in meeting the accuracy and consistency requirements for train positioning along the entire rail line. In response to these stringent requirements, this research proposes an Error-State Kalman Filter (ESKF) based real-time fusion train positioning system for urban rail transit, incorporating Ultra-Wideband (UWB) and Inertial Measurement Unit (IMU) technologies. Through practical measurements, an UWB ranging error model is established. Based on this model, a simulation study is conducted on the proposed real-time fusion positioning algorithm under different UWB anchor deployment methods in the context of urban rail transit scenarios. The real-time performance and accuracy of the proposed algorithm were validated through in-tunnel testing.

The uncertain factors such as network status, measurement errors and quality of service (QoS) requirements of applications make it challenging to guarantee the performance of edge computing-enabled networks through resource allocation schemes modeled on accurate information. This paper investigates the impact of information uncertainty on resource allocation in edge computing-enabled networks. We model the resource constraints as chance constraints and jointly optimize wireless access point (AP) selection, computing node association, and traffic engineering to maximize the network utility. Since the problem contains uncertainty parameters and binary variables, it is intractable to solve. Therefore, we utilize the Bernstein approximation to derive convex conservative approximations for chance constraints. To address the unrealistic nature of the problem due to its large size and computing complexity, we employ the alternating direction method of multiplier to iterate wireless AP selection, computing node association, and bandwidth allocation in a distributed manner. Additionally, we use the convex optimization method to solve the corresponding sub-problems. Simulations are conducted to demonstrate that our proposed resource allocation scheme can satisfy more requirements and save more resources than other schemes.

The dynamic edge computing-enabled networks contain various resources, and network parameters and system models are subject to uncertainty. Despite this, there is still a lack of comprehensive online solutions for coordinating wireless, transport, and computing resources. This paper investigates the use of online convex optimization for resource allocation in edge computing-enabled networks with time-varying cost and time-varying constraint functions. Taking into account the uncertainty of wireless status, quality of service requirements, and cost function, the goal is to minimize the long-term cost by optimizing the selection of access points, association of computing nodes, allocation of computing resources, and bandwidth allocation. To address the proposed online resource allocation problem, the modified online saddle-point algorithm is employed and dynamic regret and accumulative constraint violation are defined to measure the performance of the algorithm. To reduce the computational complexity of the projection in the modified online saddle point algorithm, the projection is reformulated as quadratic programs, which can be solved efficiently by convex optimization. Finally, the effectiveness and superiority of the proposed solution are demonstrated through simulation analysis.

Visual simultaneous localization and mapping (vS-LAM) has attracted great attentions in mobile robots. Most vSLAM systems assume that the objects are stationary in static environments. However, in the real world, there are many objects that are non-stationary in dynamic environments, which will cause performance degradation of vSLAM systems. In this paper, we propose a novel vSLAM system suitable for dynamic environments, named as SPP-SLAM, which is based on semantic masks and probabilistic propagation. Prior motion probabilities of feature points are obtained using semantic mask constraints and multi-view geometric constraints. Then the dynamic probability of each feature point is obtained via the probability propagation model, and highly dynamic feature points are rejected. In addition, we propose a missed detection compensation module combined with inertial measurement unit (IMU) information to recover the semantic masks of the missed objects. Experimental results on the OpenLORIS-Scen and TUM RGB-D datasets demonstrate that the proposed approach can improve the performance of vSLAM systems in a variety of challenging scenarios.

The rapid advancement of intelligent transportation systems necessitates efficient resource allocation for low-latency and high-bandwidth vehicular services. While traditional reinforcement learning has been widely utilized for resource allocation, it suffers from limitations such as poor generalization and interpretability. To overcome these challenges, we propose a novel Intelligence-based Reinforcement Learning (IRL) al-gorithm, which uses active inference to infer the real world and maintain an internal model of the world by minimizing free energy. We address the inefficiency of active inference by incorporating prior knowledge as macro guidance, ensuring more accurate and efficient training. By constructing the intelligence-based model, we eliminate the need for designing reward functions, which aligns better with human thinking and provides a method to reflect the learning, information transmission, and intelligence accumulation processes. Considering the dynamic and uncertain nature of vehicular scenarios, we apply the IRL algorithm to continuously evolving environments where environmental parameters are not fixed. Extensive simulations confirm the effectiveness of IRL, significantly enhancing the generalization and interpretability of intelligent models.

The rapid advancement and wide application of the Internet of Things technology (IoT) have brought unprecedented convenience to people’s production and life. A great number of devices are connected to the IoT network to provide various services for people, which also makes the IoT more vulnerable to various cyber-attacks. This paper designs a novel IoT web attack detection architecture, which combines the powerful knowledge expression ability and high interpretability of symbolic artificial intelligence (AI) with the adaptive learning ability of connectionist AI to form a closed loop of knowledge embedding and extraction, effectively improve the detection ability of web attacks. The architecture solves the “black box” feature of deep learning models and can obtain knowledge from the trained detection model and add it to the training process of the new model to improve detection capabilities. It also uses the advantages of blockchain technology to realize intelligent sharing between different detection systems, solve the problem of difficult detection model updates and training data acquisition “bottlenecks”. To better detect web attacks, we propose a semi-supervised learning method based on an interpretable convolutional neural network (CNN) to reduce misjudgments during self-training and improve detection accuracy. Additionally, we propose a new feature method to extract the features of web logs in IoT devices, which can help the system to detect web attacks in IoT more quickly and accurately. Simulation results on two different datasets show that the proposed architecture and method can effectively detect web attacks in IoT and reduce the false positive rate.

The field of autonomous driving research has made significant strides towards achieving full automation, endowing vehicles with self-awareness and independent decision-making. However, integrating automation into vehicular operations presents formidable challenges, especially as these vehicles must seamlessly navigate public roads alongside other cars and pedestrians. An intriguing yet relatively underexplored domain within autonomous driving is overtaking. Overtaking involves a dynamic interplay of complex tasks, including precise steering and speed control, rendering it one of the most intricate operations for implementing augmented intelligence driving technologies. Surprisingly, the overtaking of autonomous vehicles remains largely uncharted territory in the context of augmented intelligence for autonomous systems. This void in knowledge beckons researchers to embark on explorations and investigations in this nascent field. Our review paper systematically synthesises overtaking methodologies hinging on computer vision techniques tailored for augmented intelligence autonomous driving scenarios in response to this pressing need. Our analysis encompasses an array of domains central to overtaking in augmented intelligence autonomous vehicles, encompassing Object Detection, Lane/Line Detection, Depth Estimation, Obstacle Detection, Segmentation, and Pedestrian Detection. We meticulously analyze each domain using well-established Multimodal datasets. We assess different models’ performance across various parameters by employing graphical structures, enabling visual comparative analyses. In object detection, YOLOv4 achieves a top performance with 0.90 mAP on the BDD100K dataset. For lane detection, CLRNET excels with the highest F1 score of around 0.96 on the LLAMAS dataset. ViT-Adapter-L leads in segmentation tasks, boasting an impressive mIoU score of 83 on Cityscapes. The Hierarchical Model achieves a superior mAP of 0.90 in road sign detection on the Tsinghua-Tencent Dataset. Steering angle computation sees InterFuser as the standout, achieving the highest driving score of approximately 74.0. This paper’s primary contributions include a comprehensive assessment of diverse models for each Multimodal dataset, aiding future research in this evolving domain.

Multiaccess edge computing (MEC) empowered air–ground integrated networks (AGINs) hold great promise in delivering accessible computing services for users and Internet of Things (IoT) applications, such as forest fire monitoring, emergency rescue operations, etc. In this article, we present a comprehensive air–ground integrated MEC framework, where edge servers carried by unmanned aerial vehicles (UAVs) will provide efficient computation services to IoT devices and user equipment (UE) (which are collectively referred to as UEs). We aim to minimize the long-term average weighted sum of task completion delay and economic expenditure for all the UEs. This objective is achieved through various strategies, including preinstalling new service instances into UAVs, removing idle service instances from UAVs, task offloading decision making, access control, selecting appropriate service instances for each offloaded service request, and resource allocation optimization. Considering the complexity of the problem and the dynamics of the system, we reformulate the problem as a Markov decision process (MDP) and present a multiagent deep deterministic policy gradient (MADDPG)-based algorithm to enable low-complexity and real-time adaptive decision-making. Since our problem contains integer, binary and continuous variables, it is not straightforward to apply the MADDPG algorithm. Specifically, we first normalize the continuous variables, and then convert the continuous output generated by MADDPG into discrete variables, while ensuring the coupling constraints between different variables are preserved. The simulation results demonstrate the fast convergence of our proposed algorithm and its superior performance in minimizing costs compared with the baseline algorithms.

FL enables collaborative training of autonomous driving models without sharing the original data. It enhances the model’s environmental adaptability and establishes an effective distributed paradigm for connected and autonomous vehicles (CAVs) to share driving experiences as well as make collaborative decisions. However, participants’ negative behavior, such as free riding due to selfishness, can significantly reduce federated learning (FL) training efficiency and model accuracy. Unlike previous studies that focused solely on a single FL task, this article proposes an incentive mechanism for long-term driving model training, which models the interactions between participants and the server during the long-term FL process as an infinitely repeated game. The incentive mechanism considers the relationship between participants’ historical behaviors and their future incomes, motivating participants to maintain positive behaviors throughout the long-term FL process and ensuring the efficient operation of the training process. Furthermore, in order to increase CAVs’ enthusiasm, we design reward rules that attract new participants and encourage sustained engagement. The simulation results demonstrate that the proposed incentive mechanism maximizes the profits of both CAVs and the server in long-term FL, which effectively reduces negative CAVs’ behaviors and improves the efficiency of FL training.

Multiedge cooperative computing that combines constrained resources of multiple edges into a powerful resource pool has the potential to deliver great benefits, such as a tremendous computing power, improved response time, and more diversified services. However, the mass heterogeneous resources composition and lack of scheduling strategies make the modeling and cooperating of multiedge computing system particularly complicated. This article first proposes a system-level state evaluation model to shield the complex hardware configurations and redefine the different service capabilities at heterogeneous edges. Second, an integer linear programming model is designed to cater for optimally dispatching the distributed arriving requests. Finally, a learning-based lightweight real-time scheduler, CoRaiS is proposed. CoRaiS embeds the real-time states of the multiedge system and requests information, and combines the embeddings with a policy network to schedule the requests, so that the response time of all requests can be minimized. Evaluation results verify that the CoRaiS can make a high-quality scheduling decision in real-time, and can be generalized to other multiedge computing system, regardless of the system scales. Characteristic validation also demonstrates that the CoRaiS successfully learns to balance loads, perceive real-time state and recognize heterogeneity while scheduling.

The fiercely competitive business environment and increasingly personalized customization needs are driving the digital transformation and upgrading of the manufacturing industry. IIoT intelligence, which can provide innovative and efficient solutions for various aspects of the manufacturing value chain, illuminates the path of transformation for the manufacturing industry. It’s time to provide a systematic vision of IIoT intelligence. However, existing surveys often focus on specific areas of IIoT intelligence, leading researchers and readers to have biases in their understanding of IIoT intelligence, that is, believing that research in one direction is the most important for the development of IIoT intelligence, while ignoring contributions from other directions. Therefore, this paper provides a comprehensive overview of IIoT intelligence. We first conduct an in-depth analysis of the inevitability of manufacturing transformation and study the successful experiences from the practices of Chinese enterprises. Then we give our definition of IIoT intelligence and demonstrate the value of IIoT intelligence for industries in fucntions, operations, deployments, and application. Afterwards, we propose a hierarchical development architecture for IIoT intelligence, which consists of five layers. The practical values of technical upgrades at each layer are illustrated by a close look on lighthouse factories. Following that, we identify seven kinds of technologies that accelerate the transformation of manufacturing, and clarify their contributions. The ethical implications and environmental impacts of adopting IIoT intelligence in manufacturing are analyzed as well. Finally, we explore the open challenges and development trends from four aspects to inspire future researches.

With the emergence of time-sensitive IIoT applications, such as remote operation and industrial control, a long-distance deterministic forwarding service is highly desirable. However, most of the existing research is limited to local area networks, or requires costly replacement of core network devices. Enabling incremental deterministic networks based on off-the-shelf technologies is a significant challenge. This paper designs a core-agnostic and cost-effective solution named Hirail to achieve the smooth evolution of long-distance deterministic networks. Firstly, we investigate that a time-discrete shaper (TDS) can be deployed at the ingress node to enable millisecond-level bounded delay. TDS functions similarly to the concept of buying time-stamped tickets for each flow prior to getting on a high-speed rail, thus avoiding the expensive modification of core devices. Then, to alleviate the flow aggregation problem under long-distance links, we utilize the inband network telemetry to construct the delay-aware network map and conduct adaptive source routing based on the map. Finally, an adjustable buffer at the last hop is devised for jitter reduction. Evaluation results show that Hirail can meet the bounded delay and jitter demands, and outperforms other solutions in terms of performance and overhead.

In future 6G-enabled Internet of Things (IoT), users and devices will be divided into numerous distributed domains with smaller base station coverage due to the utilization of terahertz high-frequency band communication. Deep reinforcement learning (DRL) agents will be increasingly deployed in the domain to achieve intelligent service provisioning and resource allocation. However, the existing DRL-based method faces the problem of repeated model training and poor generalization ability when service demand fluctuates and environmental changes occur. In addition, limited training samples in each domain also lead to insufficient model training. Inspired by the collaborative learning of human knowledge, we propose a knowledge collaboration-based resource allocation mechanism for future 6G-enabled IoT and address two basic issues: 1) which agent should collaborate with and 2) how to collaborate. Specifically, we first model the distributed network as a graph and use graph attention (GAT) to capture the fluctuant service demands and time-varying resource capacities in temporal and spatial domains, and then calculate the similarity between the agents. We further propose a collective reinforcement learning (CRL) algorithm that facilitates knowledge collaboration between the agents through the policy distribution. Simulation results verify that the proposed GAT-CRL achieves fast convergence as deep deterministic policy gradient (DDPG) in 4K steps, computing the similarity score more accurately with the increasing attention heads, and achieves higher successful flow than the soft actor-critic (about 3.6%–5.4%) and DDPG (about 14.6%–21%) when adapting to unseen traffic patterns/loads and increasing topology scales.

With the advancement of Urban Rail Transits (URTs), the demand for artificial intelligence (AI) based URTs services grows exponentially. Edge intelligence (EI) leverages computing resources on the network edge to provide realtime intelligent services in close proximity. As it enables fast distributed learning, EI is envisioned to be a potential component of URTs, and ideal EI service provision is a critical concern for the intelligent development of URTs. The existing EI-related research concentrates on the computation offloading of general AI-based tasks, whereas both the edge server deployment and AI model training process are not explicitly designed for URTs. The URTs AI service characteristics such as model training demand, priority, and security are largely ignored. In this paper, we propose a novel collaborative EI service provision framework for URTs. Blockchain is used along with the EI server to construct a trusted computing infrastructure. To address the EI service credit crisis, a blockchain-based trust management mechanism including short-term reward incentives and long-term reputation evaluation is designed in the trusted computing infrastructure. An HRL-based collaborative training service optimization model is proposed to improve the learning efficiency and edge resource utilization rate in URTs. Specifically, the proposed two-stage collaborative optimization model jointly considers high-level service scheduling and low-level task offloading. In addition, we present an intelligent train control model based on the state-ofthe-art decision transformer (DT), with the training service as a case study to demonstrate the effectiveness of the proposed collaborative EI service provision. Extensive simulation results show that the proposed EI service provision framework can provide trusted, efficient, and high-quality AI training services, simultaneously improving URTs operational efficiency.

Unmanned aerial vehicles (UAVs)-assisted multi-access edge computing (MEC) platforms are becoming an increasingly popular solution for infrastructure-less Internet of Vehicles (IoVs) due to their mobility and flexibility. To address the challenges of uneven task offloading and vehicle mobility, in this paper, we propose a mobility-aware service offloading and migration scheme for UAV-assisted IoVs. We formulate the service placement, service migration, and UAV deployment as an optimization problem to minimize the serving delay of task addressing for IoVs, under a predefined long-term migration cost budget. To solve the problem, we use the Lyapunov optimization method to transform the long-term optimization into a real-time optimization problem. Additionally, we design a multi-agent deep deterministic policy gradient (MADDPG) algorithm to solve the problem. Compared with traditional central optimization methods, the proposed algorithm can achieve a near-global optimal policy by leveraging only local observation information. Simulation results show that the proposed MADDPG algorithm can achieve good convergence performance, and the proposed scheme can achieve quasi-optimal performance in terms of serving delay, service offloading rate, and service migration cost.

With the emergence of the fifth generation (5G) era, the development of the Internet of Things (IoT) network has been accelerated with a new impetus, making it imperative to strive for a more reliable and efficient network environment. To accomplish this, we introduce and investigate a novel proposal for the intelligent reflecting surface (IRS) enabled uplink secure receive spatial modulation (SM), named IRS-USRSM, to resolve the security issues arising from the open wireless transmission environment in the 5G IoT network. In the IRS-USRSM scheme, we assume that the passive eavesdropper is directly connected to the uplink user and occasionally connected to the IRS. To achieve enhanced secrecy with finite alphabet inputs, a joint transmitter perturbation and IRS reflection design for physical layer security is proposed to guarantee secure and reliable transmission of IRS-USRSM. Specifically, two categories of IRS-based random phase compensation strategies, namely, random perturbation compensation and random path synthesize, along with maximum likelihood detection and suboptimal detection are proposed to meet the variant design requirements between achieved performance and system cost. Furthermore, in order to evaluate the performance limits of the IRS-USRSM, the closed-form results of average bit error probabilities and discrete-input continuous-output memoryless channel capacities are derived using the method of moment generating function. Simulation results are presented to verify the correctness of our theoretical analyses, as well as to demonstrate the efficiency and superiority of the proposed IRS-USRSM scheme.

With the development of internet of things (IoT) technology and its wide application in urban traffic, the next-generation vehicle-to-everything (V2X) communication network should support high-capacity, ultra-reliable, and low-latency massive information exchange to provide unprecedentedly diverse user experiences. The development of the sixth-generation (6G) mobile communication technology will pave the way for realizing this vision. Reconfigurable intelligent surfaces (RISs), a critical 6G technology, is expected to make a big difference in V2X communications when used in conjunction with unmanned aerial vehicles (UAVs), allowing for extremely increased communication capacity and reduced latency. We propose a UAV-enhanced RIS-assisted V2X communication architecture (UR-V2X) suitable for urban three-dimensional (3D) IoT traffic and design an adapted MAC protocol UR-V2X-MAC to accomplish communication resource allocation and scheduling. The UAVs are used as access points and resource allocation centers, while the RISs are used as passive relays to assist V2X communication in proposed architecture. To improve the performance of UR-V2X-MAC, we use a distributed optimization algorithm in the message report phase of the protocol to maximize the system capacity by allocating the transmit power and alternately optimizing the RIS phase shift matrix. We analyze the delay and system capacity characteristics under different parameter settings through theoretical derivation and protocol performance simulation. Analysis and simulation results are presented to demonstrate that UR-V2X-MAC achieves a reduction in communication delay and a significant increase in system capacity through detailed design and alternate optimization compared to the existing V2X MAC protocol and no-RIS case.

In the large-scale dynamic Low Earth Orbit (LEO) satellite networks, the conventional TCP-based single-path transmission encounters challenges such as prolonged propagation delay, frequent connection failures, and suboptimal resource utilization. In this paper, we propose an Integrated Multi-Path Network Coding (IMPNC) transmission scheme. This scheme leverages multiple paths for end-to-end transmission to achieve bandwidth aggregation and redundant backup. The multi-path transmission is facilitated by Multi-Path Quick UDP Internet Connection (MPQUIC) protocol to adapt to the limited satellite bandwidth and caching resources. The proposed approach involves encoding packets at nodes along the paths, addressing the significant out-of-order problem arising from variable delays on different paths. Additionally, we present a Software Defined Networking (SDN)-based domain clustering architecture, which offers a more streamlined control approach, reducing overall complexity. Furthermore, we formulate the domain clustering problems as mixed-integer nonlinear programming and the coding-based routing problem as a Steiner tree problem. Evaluation results demonstrate that the proposed scheme effectively reduces the latency over 25.1%, enhances bandwidth utilization by 19.6%, and ensures reliable data transmission by reducing retransmission probability by 4.1%.

The power inspection is an important application of UAV-assisted power internet of video things (IoVT) for maintaining the safety of the power system. Due to the limitations of distance and angle, the resolution of the images captured by UAV is low, which seriously impacts the effects of small insulator defects detection. To address this problem, we propose a small-size defects detection method based on multi-scale feature interaction transformer for UAV-assisted Power IoVT. For the algorithm, we design a super-resolution reconstruction-assisted small object detection algorithm, the super-resolution module generates high-resolution images with the requirements of object detection function, which greatly improves the small object detection performance. Moreover, we design multi-scale feature interaction transformer network (MFITN), compared with the traditional non-local attention mechanism, the network structure can capture dependencies in multi-scales features, furthermore, the advantage assist the super-resolution module to generate more realistic image information to further improve small object detection. In addition, we propose a distributed model deployment strategy to deploy our high computational complexity algorithm in the edge side of the IoVT system, which can drive the overall algorithm to perform low-latency edge computation by relying only on the limited computing power devices. Experiments demonstrate that our method has better small object detection performance (mAP=81.3%, FPS=49.7), the super-resolution reconstruction is able to recover more realistic detail information, the distributed computing method can reduce the response latency by 33.4%-87.2%, which all contribute UAV-assisted Power IoVT system to realize accurate and fast power insulator defects detection.

Video surveillance as an important function of internet of video things (IoVT) system has been widely used in coal mine monitoring for coal mine safety with excellent results, however, there are still many shortcomings: 1) Existing coal mine IoVT systems have limited detection accuracy for small-sized objects; 2) Coal mine video surveillance systems generally adopt centralized cloud computing, transmission of massive data causes high latency, which seriously affects the response speed of object detection function; 3) The concept drift caused by the data stream seriously affect the detection effect of the offline algorithm. To address the above issues, we propose a small object detection method based federated intelligence to assist coal mine IoVT for object detection. First, we design a lightweight neural network Rep-ShuffleNet to improve YOLOv8, the state-of-the-art YOLO algorithm, to maintain high detection accuracy while dramatically increasing the inference speed, and with the advantage of lightweight, it can be deployed to embedded devices for low-latency edge computing; Moreover, we design a federated learning-based MLC-FL algorithm for local algorithms’ automatic and efficient optimization by asynchronous communication and data interaction reduction strategy. The experimental results show that with the assistance of federated intelligence model optimization strategies, the lightweight YOLOv8 has excellent detection performance (mAP: 94.6%, APsmall: 86.7%, FPS: 21.6), thus to assist coal mine IoVT to realize accurate and real-time underground small object detection.

The growing utilization of unmanned aerial vehicles (UAVs) across diverse industries has led to increased interest in UAV-assisted data acquisition for the Internet of Things (IoT). The security of image data collected by UAVs during transmission within the IoT has become a critical concern. This article focuses on the security challenges associated with UAV-assisted image data collection in the IoT and presents a dedicated framework designed to enhance the security of this process. Given the high-resolution nature of UAV-captured images, traditional encryption methods face difficulties in directly and effectively encrypting such data. To address this issue, this article introduces an efficient chaotic image encryption algorithm integrated into the proposed protection framework. The algorithm features a novel 1-D chaotic system for generating effective chaotic sequences. For the scrambling phase, a chaotic four-spiral transformation method is employed, and the diffusion process utilizes the Fibonacci matrix. This strategic approach aims to minimize pixel correlation within the image, thereby bolstering the overall security of the encryption process. Experimental validation conducted on authentic UAV image data sets demonstrates the superior, practical, secure, and efficient characteristics of the proposed algorithm.

Federated learning, leveraging distributed data from multiple nodes to train a common model, allows for the use of more data to improve the model while also protecting the privacy of original data. However, challenges still exist in ensuring privacy and security within the interactions. To address these issues, this paper proposes a federated learning approach that incorporates blockchain, homomorphic encryption, and reputation. Using homomorphic encryption, edge nodes possessing local data can complete the training of ciphertext models, with their contributions to the aggregation being evaluated by a reputation mechanism. Both models and reputations are documented and verified on the blockchain through consensus process, which then determines the rewards based on the incentive mechanism. This approach not only incentivizes participation in training, but also ensures the privacy of data and models through encryption. Additionally, it addresses security risks associated with both data and network attacks, ultimately leading to a highly accurate trained model. To enhance the efficiency of learning and the performance of the model, a joint adaptive aggregation and resource optimization algorithm is introduced. Finally, simulations and analyses demonstrate that the proposed scheme enhances learning accuracy while maintaining privacy and security.

The integration of Software-Defined Networking (SDN) and Artificial Intelligence (AI) presents promising opportunities for managing and optimizing LEO satellite network routing. However, as the scale and coverage of satellite networks continue to expand, challenges are posed to both centralized and distributed architectures in terms of managing network information and coping with routing complexity. To overcome these challenges, leveraging distributed SDN technology, a stigmergy multi-agent hierarchical deep reinforcement learning routing algorithm is proposed in multi-domain collaborative satellite networks. A pheromone-based mechanism is incorporated to facilitate collaboration during independent training, and hierarchical control is employed to decouple the complexity of cross-domain routing decisions. Simulation results demonstrate that our proposed algorithm exhibits good scalability and performance in large-scale satellite networks.

In this paper, we establish a two-timescale framework for the joint service deployment and task scheduling problem in satellite edge computing networks.We aim to optimize the computing performance of networks with diverse quality-of-service (QoS) guarantees for computing tasks. Specifically, to capture the small-timescale network dynamics and task randomness, we formulate the task scheduling problem as a constrained Markov decision process (CMDP) to minimize the energy consumption, load imbalance and packet loss of networks while ensuring the long-term delay. The Lyapunov technique is employed to deal with the delay constraints. A soft actor-critic (SAC)-based deep reinforcement learning (DRL) framework is designed to learn the stationary scheduling policy. We further explore the significant impact of deploying diverse services on the performance of task scheduling in satellite edge computing. Considering that frequent deployment of services will incur huge deployment overhead, we optimize the service deployment on a larger timescale. The optimization problem is modeled as an integer programming problem to improve the service capability of networks and reduce service deployment costs. A heuristic-based atomic orbital search (AOS) approach is proposed to obtain the superior policy with low complexity. Due to the correlation between the problems of two timescales, a hierarchical solution is constructed to iteratively find the excellent solution. Finally, extensive simulations are conducted to validate the effectiveness and superiority of the proposed scheme.

Integrating the unmanned aerial vehicles (UAVs) assisted mobile edge computing (MEC) network with the blockchain technology emerges its superiority in the network utilization, differentiated service, and security, which has been regarded as a promising technique for time-critical applications. In this paper, we propose a UAV-assisted MEC network architecture and a comprehensive data processing flow, where the UAVs cooperate with the base station in computation as edge servers and act as blockchain nodes. We formulate an optimization problem that jointly considers UAVs’ position, data offloading, and resource allocation for minimizing the total time consumption of data processing. To address this problem, we decouple it as three tractable subproblems and propose a Block Coordinate Descent (BCD)-based iterative algorithm. In addition, we analyze the task migration and resource allocation problem in computation, and obtain analytical solutions by the Karush-Kuhn-Tucker (KKT) conditions. The simulated results indicate that the proposed algorithm leads to substantial performance gains.

How to improve the coverage capability of network including connectivity and throughput is vital for enabling the Internet of Everything (IoE) in B5G/6G. However, the traditional terrestrial networks are confronted with the high-cost and inflexible challenges especially in the remote area and emergency applications. In order to solve these challenges, we consider a heterogeneous aerial network (HetAN), where some low-altitude base stations (LBSs) are deployed as access points for wireless coverage and a high-altitude base station (HBS) hovers as the hub for backhaul of LBSs. Furthermore, we apply the non-orthogonal multiple access (NOMA) to uplink transmission for the terrestrial users, which enable massive connectivity in the IoE. To maximize connectivity and throughput, we jointly optimize the LBSs’ deployment, power control, channel allocation, and rate control by fully exploiting the potential of the HetAN in wide-area coverage. For solving the formulated problem efficiently, we propose an iterative algorithm based on the methods of graph theory, bionic algorithm, and theoretical analysis. Simulation results are provided to reveal the influence of the control variables on network performance and indicate that our algorithm can greatly improve the connectivity and throughput with the other schemes.

With the significant advantages of system complexity and operating costs, train autonomous circumambulate system (TACS) is gradually replacing the traditional communication-based train control system as the next-generation train operation control system development direction. As train operation and control become more decentralized and autonomous, real-time and accurate obstacle detection, apart from route-level protection, is quite desirable in TACS. Most of the existing researches about obstacle detection focus on detection algorithm optimization based on the once-deployed lifelong use principle, whereas model reoptimization based on the actual operating environment under unexpected situations and model sharing among multiusers are largely ignored. In this article, we design a novel obstacle detection system in TACS based on blockchain-empowered edge intelligence (EI). To make full use of the massive raw unannotated data collected online, we first propose an semisupervised learning-based TACS obstacle detection model. Considering the resource-hungry model training, we introduce EI into TACS and propose a multiagent reinforcement learning-based task offloading algorithm for secure and efficient computation offloading coordination. Furthermore, we propose a blockchain-based model sharing scheme to facilitate the multimodel parameter exchange and improve the obstacle detection accuracy. Extensive simulation results show that the designed obstacle detection system can effectively improve the TACS obstacle detection performance.

Federated learning (FL) is a promising technology for autonomous driving, enabling connected and autonomous vehicles (CAVs) to collaborate in decision-making and environmental perception while preserving privacy. However, traditional FL algorithms face challenges related to imbalanced data distribution, fluctuating channel conditions, and potential security risks associated with malicious attacks on local models. This paper proposes a fair and secure FL algorithm that not only addresses the challenges arising from imbalanced data distribution and fluctuating channel conditions, but defends against malicious attacks. Specifically, we first propose a personalized local training round allocation algorithm to balance energy costs and accelerate model convergence. Next, in order to further guarantee security, we embed an attack module based on Gini impurity. Extensive simulations demonstrate that the proposed algorithm achieves energy fairness, reduces global iteration time, and exhibits resistance against malicious attacks.

Vehicular Edge Computing and Networks (VECoNs) have gained popularity for its enhanced Internet of Vehicles (IoV) capabilities. To satisfy the needs of delay-sensitive and computation-intensive in-vehicle applications, VECoNs need to provide low-latency task offloading services. However, existing offloading frameworks generally overlook the spatially and temporally heterogeneous computation task arrival patterns. The former causes overloading and underloading of RSU computational resources and thus hinders further reduction of offloading latency on the macro-scale, while the latter emphasizes the importance of long-term system performance, especially energy constraints, posing challenges to the design of offloading framework and optimization strategies. This paper introduces a novel distributed two-stage task offloading architecture based on Lyapunov and multi-agent deep deterministic policy gradient (MADDPG). On one hand, it jointly optimizes the initial offloading stage within VEC subsystems and the RSU peer offloading stage to minimize offloading delays for each VEC subsystem. On the other hand, it incorporates RSU energy consumption within long-term constraints to formulate the offloading optimization problem. After decoupling the energy coupling between RSU time slots using the Lyapunov algorithm, a Lyapunov and MADDPG-based distributed task offloading (LAMETO) algorithm is presented to solve the optimal problem in a distributed manner. Simulation results show that the proposed framework and algorithm can reduce the system delay, energy consumption, and energy deficit while stabilizing convergence.

“Read-write-own” based Web3 has been proposed as a promising user-centric Internet to open the new generation of the World Wide Web, where Web3 users can independently manage data and derive value from creating content without relying on intermediaries. Connected and autonomous vehicles (CAVs) in Web3 can trade models in a self-controlled and decentralized credible way, which is a fundamentally and principally innovation based on novel architecture. Effectively implementing such paradigms involves proper model trading strategies. However, reinforcement learning (RL)-based strategies face challenges of poor generalization ability, low feasibility, and the exploration-exploitation dilemma. It is also difficult to define an explicit and appropriate reward function. Therefore, in this paper, we propose an intelligence-based reinforcement learning (IRL) approach for CAVs in Web3. We present a framework to enable model transactions between CAVs. Also, we provide a decentralized identifier (DID)-based identity management system for resource description and data verification to access Web3, followed by the mechanism and supporting smart contracts. Furthermore, we formulate the model trading issue as an active inference to form higher-level cognition about the environment without rewards. Then we use IRL to solve it. And we use “intelligence”, a high-level indicator, to quantify the efficiency of such cognition. It can evaluate the difference between the predicted state and the real state in policy exploration. The proposed scheme shows good generalization and can auto-balance exploration and exploitation, simultaneously achieving outperforming performance on the model trading issue with no rewards. In simulations, the performance of the proposed scheme is compared with existing methods.

Real-time detection of driver drowsiness is critical to reduce the risk of road accidents and fatalities. Current facial landmark-based methods usually use a two-stage paradigm, where faces and facial landmarks are localized separately. Additionally, most methods can be hindered by challenging conditions, such as night driving or eyes closed. To address these challenges, we present a refined YOLO network named YOLOFaceMark that can simultaneously detect faces and their facial landmarks. Furthermore, we introduce a drowsiness detection model based on facial landmarks. This model utilizes extracted eye and mouth information to identify drowsy states. We optimize the original YOLO components through structural re-parameterization, channel shuffling, and the design of a dual-branch detection head with an implicit module. These enhancements are designed to improve the accuracy while maintaining computational efficiency. We validate the real-time performance and accuracy of YOLOFaceMark on public datasets, including 300W and COFW. Additionally, we conduct further validation to demonstrate our ability to achieve effective and robust drowsiness detection solely based on the facial landmarks detected by YOLOFaceMark.

Urban Rail Transit Systems (URTS) have increasingly become the backbone of modern public transportation, attributed to their unparalleled convenience, high efficiency, and commitment to sustainable green energy. As we witness a global resurgence of urban rail transit, it becomes evident that most existing URTS still operate on a level of suboptimal intelligence, with their operation and maintenance methods lagging behind other advanced urban transit systems. URTS generate considerable data, offering substantial opportunities for service quality enhancements. Machine Learning (ML), with its demonstrated proficiency in extracting valuable insights from vast data, hold significant promise in the quest to empower URTS. This survey presents a comprehensive exploration of the potential application of ML in URTS. Initially, we delve into the existing challenges of URTS, thereby elucidating the compelling motivation behind the integration of ML into these systems. We then propose a taxonomy of ML paradigms and techniques, discussing in-depth their potential applications in URTS, encompassing perception, prediction, and optimization tasks. Subsequently, we scrutinize a plethora of ML-empowered URTS application scenarios, including but not limited to obstacle perception, infrastructure perception, communication and cybersecurity perception, passenger flow prediction, train delay prediction, fault prediction, remaining useful life (RUL) prediction, train operation and control optimization, train dispatch optimization, and train ground communication optimization. Finally, we present an insightful discussion on the challenges and future directions for URTS, aiming to harness the full potential of ML techniques to deliver superior service and performance.

Train autonomous circumambulate systems (TACS) are a new-generation train control systems. They are characterized by autonomous travel path planning, autonomous protection, and autonomous train operation adjustment. One crucial problem in TACS is real-time communication and computation of autonomous train control systems. Trains need to obtain the real-time state of all the other trains and derive real-time intelligent control commands in TACS. With high capacity and reliable 5G technologies, edge intelligence (EI) can perform complex computing tasks offloaded from trains with little delay. In this paper, we develop a collaborative train and edge computing framework for TACS to provide real-time communication and computation service for train control. To reduce the tracking deviations and ensure the train operation punctuality, ride comfort, and energy-saving ability, we adopt the model predictive control (MPC) algorithm to optimize the autonomous train control process. To cope with the limited onboard computing power, we propose a meta reinforcement learning (MRL) based collaborative computing method to solve the computation offloading problem. Compared with the existing RL-based offloading policy that requires sufficient data samples for training, MRL can rapidly adapt to different computation offloading environments, which is exceptionally suited for the urban rail transit system where different rail lines have different operating environments, and we do not have enough data to finish a regular reinforcement learning and training task. Experimental results illustrate that the proposed framework can provide TACS with reliable and real-time computing services. The train operational efficiency can be significantly improved with our proposed collaborative computing train control algorithm.

The incompleteness of knowledge graphs (KGs) negatively impacts the performance of KGs in downstream applications (e.g., recommendation systems and information retrieval). This phenomenon has brought an increasing rise in research related to knowledge graph reasoning. Recently, emerged reinforcement learning (RL)-based multi-hop reasoning methods can infer missing information through multi-hop reasoning according to the existing information in KGs, which has better reasoning performance and interpretability. However, these methods always use relation-entity pairs that have been pre-cropped as the action space of agents for path reasoning, which leads to two problems: 1) insufficient learning and reasoning ability of reasoning models and 2) the hard convergence of the training process of agents. To address these problems, we propose aHierarchicalSequenceMultiHop (HSMH) reasoning framework, which consists of the interactive search reasoning model, local-global knowledge fusion mechanism, and action optimization mechanism. We use interactive search reasoning models to select relations and entities independently, thus fully mining the semantic information of relations and entities and improving the learning and reasoning ability of reasoning models. In the HSMH framework, we design the local-global knowledge fusion and action optimization mechanisms for path reasoning, which can enhance agents' state information and action space. Specifically, the local-global knowledge fusion mechanism is designed to acquire the local knowledge of entities and neighboring relations and the global knowledge about KG structure. This local-global knowledge can improve the learning ability of reasoning models. In addition, the action optimization mechanism can combine the filtered action space and the additional action space for efficient path reasoning for agents. Experimental results on five benchmark datasets show that our proposed HSMH framework comprehensively outperforms the state-of-the-art multi-hop reasoning model.

With the increasing popularity and demands for large language model applications on mobile devices, it is difficult for resource-limited mobile terminals to run large-model inference tasks efficiently. Traditional deep reinforcement learning (DRL) based approaches have been used to offload large language models (LLMs) inference tasks to servers. However, existing DRL solutions suffer from data inefficiency, insensitivity to latency requirements, and non-adaptability to task load variations, which will degrade the performance of LLMs. In this paper, we propose a novel approach based on active inference for LLMs inference task offloading and resource allocation in cloud-edge computing. Extensive simulation results show that our proposed method has superior performance over mainstream DRLs, improves in data utilization efficiency, and is more adaptable to changing task load scenarios.

In this paper, we optimize network computational performance and ensure diverse quality of service (QoS) for tasks by developing a dual-timescale joint optimization framework for satellite-terrestrial integrated edge computing networks (STECN). In our architecture, STECN can handle intelligent tasks for the Internet of remote things (IoRT) devices based on multiple configured applications deployed. Specifically, we formulate task scheduling as a Markov decision process (MDP) to minimize network energy consumption and task processing delay at small timescales. A deep reinforcement learning (DRL) framework is designed for policy learning. Recognizing the impact of resource slicing on task scheduling in STECN and the deployment overhead from frequent changes, we further optimize resource slicing at larger timescales. To enhance network service capability under dynamic demand, we establish a resource slice gap index, characterizing the difference between actual resources and service demand. By a heuristic-based artificial electric field (AEF) approach, we obtain an optimal strategy with low complexity. Considering the correlation between two timescales, the optimal solution is found by iteratively constructing a hierarchical solution. In addition, to guarantee the global load balancing of the network, we introduce a self-attention mechanism, which allows the knowledge of other satellites to be taken into account when slicing the satellite resources. Finally, extensive simulations confirm the effectiveness and superiority of the proposed scheme.

Distributed clouds decentralize cloud resources, moving from a single high-level point in the network to multiple low-level points, allowing for the dynamic distribution of services across the “cloud-edge-end”. Nevertheless, the “best-effort” traditional networks suffer from unpredictable service quality and limited collaboration between services and networks. To address these shortcomings, we present a novel solution named “Network-as-a-Service (NaaS) Tickets,” inspired by traffic tickets in transportation systems to empower distributed clouds with customized service capabilities. Specifically, we first propose NaaS Tickets-enabled service-customized distributed clouds (NT-SCDC) to realize on-demand and service-oriented interconnection in a wide area. To establish a solid connection between services and networks, we introduce an auction-driven matching mechanism for NaaS Tickets. Then, the matching problem is formulated via an online framework MatOnline, which translates the long-term market problem into a series of one-shot auctions for NaaS Tickets. Based on the Vickrey-Clarke-Groves (VCG) mechanism, we develop MatVCG algorithm to handle one-shot matching problems, guaranteeing truthfulness, individual rationality, and social welfare. Moreover, we improve the performance of MatOnline to find the minimum feasible scale-down ratio with reduced budget expenditure. Experimental results demonstrate our algorithm achieves a stable competitive ratio on social welfare, effectively meeting customized demands

In this paper, a clustered parallel training algorithm is designed for personalized federated learning (Per-FL), called CPPer-FL. CPPer-FL improves the communication and training efficiency of Per-FL from two perspectives, namely, less burden for the central server and lower interaction idling delay. CPPer-FL adopts a client-edge-center learning architecture, which offloads the central server's model aggregation and communication burden to distributed edge servers. Also, CPPer-FL redesigns the cascading model synchronization and updating procedure in conventional Per-FL and changes it to a parallel manner, thus improving the interaction efficiency in the training process. Further, for the proposed hierarchical architecture, two approaches are proposed to cater to Per-FL: similarity-based clustering for client-edge association and personalized model aggregation for parallel model updating, such that clients' personal features can be preserved in the training process. The convergence of CPPer-FL has been formally analyzed and proved. Evaluation results validate the communication efficiency, model convergence, and model accuracy improvement.

Multimodal recommendation is an emerging task with the goal of improving the effectiveness of the recommendation system by utilizing multimodal data (images, texts, etc.). Most previous methods have struggled with the ability to mine item semantic relationships while guaranteeing accurate modeling of user modality preferences, resulting in low recommendation accuracy. To address this issue, this paper proposes a novel and effective Self-suPervised duAl preference enhanCing nEtwork for multimodal recommendation, named SPACE, which further mines user preferences towards historical interactions and multimodal features of items to obtain more precise user and item representation. Specifically, we design an interaction preference enhancing module to learn both interactive and latent semantic relationships between users and items. Then, a modality preference enhancing module is established by introducing self-supervised learning (SSL), which aims to strengthen the role of dominant modality-specific representation of items. Finally, the enhanced interaction and modality representations are fused, and the recommendation performance is largely improved by utilizing dual joint prediction. Extensive experiments are conducted on three real-world datasets, and the simulation results demonstrate that the proposed SPACE model outperforms the state-of-the-art multimodal recommendation methods.

As an important branch of machine learning, reinforcement learning (RL) has proved its efficiency in many emerging applications in science and engineering. A remarkable advantage of RL is that it enables agents to maximize their cumulative rewards through online exploration and interactions with unknown (or partially unknown) and uncertain environments, which is regarded as a variant of data-driven adaptive optimal control methods. However, the successful implementation of RL-based control systems usually relies on a good quantity of online data due to its data-driven nature. Therefore, it is imperative to develop data-efficient RL methods for control systems to reduce the required number of interactions with the external environment. Moreover, network-aware issues, such as cyberattacks, dropout packet and communication latency, and actuator and sensor faults, are challenging conundrums that threaten the safety, security, stability, and reliability of network control systems. Consequently, it is significant to develop safe and resilient RL mechanisms.

Graph learning can effectively characterize the similarity structure of sample pairs, hence multiple kernel clustering based on graph learning (MKC-GL) achieves promising results on nonlinear clustering tasks. However, previous methods confine to a “three-stage” scheme, that is, affinity graph learning, Laplacian construction, and clustering indicator extracting, which results in the information distortion in the step alternating. Meanwhile, the energy of Laplacian reconstruction and the necessary cluster information cannot be preserved simultaneously. To address these problems, we propose a one-stage shifted Laplacian refining (OSLR) method for multiple kernel clustering (MKC), where using the “one-stage” scheme focuses on Laplacian learning rather than traditional graph learning. Concretely, our method treats each kernel matrix as an affinity graph rather than ordinary data and constructs its corresponding Laplacian matrix in advance. Compared to the traditional Laplacian methods, we transform each Laplacian to an approximately shifted Laplacian (ASL) for refining a consensus Laplacian. Then, we project the consensus Laplacian onto a Fantope space to ensure that reconstruction information and clustering information concentrate on larger eigenvalues. Theoretically, our OSLR reduces the memory complexity and computation complexity to$O(n)$and$O(n^2)$, respectively. Moreover, experimental results have shown that it outperforms state-of-the-art MKC methods on multiple benchmark datasets.

The advancement of Internet of Things (IoT) technology has significantly transformed the dynamic between humans and devices, as well as device-to-device interactions. This paradigm shift has led to profound changes in human lifestyles and production processes. Through the interconnectedness of numerous sensors and controllers via networks, the IoT facilitates the seamless integration of humans with diverse devices, leading to substantial economic advantages. Nevertheless, the burgeoning IoT industry and the rapid proliferation of various IoT devices have also introduced a multitude of security vulnerabilities. Cyber attackers frequently exploit cyber attacks to compromise IoT devices, jeopardizing user privacy and property security, thereby posing a grave menace to the overall security of the IoT ecosystem. In this paper, we propose a novel IoT Web attack detection system based on a joint embedded prediction architecture (JEPA), which effectively alleviates the security issues faced by IoT. It can obtain high-level semantic features in IoT traffic data through non-generative self-supervised learning. These features can more effectively distinguish normal data from attack data and help improve the overall detection performance of the system. Moreover, we propose a feature interaction module based on a dual-branch network, which effectively fuses low-level features and high-level features, and comprehensively aggregates global features and local features. Simulation results on multiple datasets show that our proposed system has better detection performance and robustness.

Network function virtualization (NFV) has attracted attention because of its flexible configuration and management of network functions. Based on NFV, the service function chain (SFC) defines a group of virtual network functions (VNFs) connected sequentially, enabling flexible customization and provisioning of network services. In the large-scale and heterogeneous Internet of Things (IoT) environment, e.g., industrial IoT, servers provided by a single infrastructure provider (InP) cannot support the deployment of all VNFs, and SFCs must be deployed across multiple domains. However, SFCs deployed across multiple domains will inevitably bring privacy leakage and resource coordination difficulties, thereby reducing the efficiency of network services. To address these issues, this paper proposes a privacy-preserving deployment mechanism (PPDM) for SFCs that achieves near-optimal SFC deployment across multiple domains while protecting resource and topology privacy. PPDM first performs virtual resource prediction and forms the service intention response matrix (SIRM) based on SFC requests (SFCRs). Second, the multi-domain controller (MDC) discovers a near-optimal SFCs deployment strategy by deep Q-network (DQN) using SIRM as input to protect domains’ privacy. Finally, the learned strategies are distributed to intra-domain controllers (IDCs) to implement specific services. Simulation results demonstrate that the proposed method outperforms privacy-preserving and non-privacy-preserving methods.

Unmanned aerial vehicles (UAVs) are widely recognized as crucial supplementary component of 6G networks. Owing to the key attributes of UAVs (mobility, flexibility, and adjustable altitude), UAVs can serve as flying base stations (BSs), flying relays and mobile terminals in order to expand the service coverage and derive more applications. Softwarization is regarded as dominant attribute of network architecture of 6G, mainly realized by network function virtualization (NFV) and software defined networking (SDN). In this paper, we concentrate on researching the resource orchestration and allocation of end-to-end (E2E) slice services in softwarized UAVs-assisted 6G terrestrial networks. Problem models of UAVs-assisted 6G terrestrial networks and E2E slice are firstly introduced. Then, the problem formulation of resource orchestration and allocation of E2E slice is presented. Afterwards, one novel framework design, abbreviated as ReOrcAll-UAVs-6G, is detailed. When receiving one E2E slice, our ReOrcAll-UAVs-6G checks the available softwarized resources. If having available softwarized resources, our ReOrcAll-UAVs-6G turns to serving the slice and fulfilling slice’s tailored resource demands. During the orchestration and allocation phase, wireless and wired resource requests of this slice are considered and executed. Evaluation work and gained results of ReOrcAll-UAVs-6G and selected approaches are illustrated and analyzed. Gained results reveal that our ReOrcAll-UAVs-6G achieves apparent performance advantage, comparing with all selected approaches.

The emerging computing paradigms offer effective resolutions for the escalating conflict arising from the heightened computational demands of portable terminals and their constrained capacity. Concurrently, the architecture has transitioned from a single-tier structure to a multi-tier collaborative framework, enhancing flexibility and enabling fine-grained computation offloading. Nevertheless, existing research on multi-tier computation offloading faces challenges, including inefficient resource perception and task decomposition; there is a notable absence of an effective hierarchical task scheduling strategy within the multi-tier collaborative architecture. To bridge these gaps, our paper investigates the multi-granularity task decomposition and hierarchical task scheduling in a cloud-edge-end collaborative computing network. We first introduce a large-small resource tree (LST) model to facilitate efficient resource perception across three-tier network nodes. Then we propose a multi-granularity task decomposition algorithm (MTDA) based on long short-term memory (LSTM) network resource prediction to fully utilize the distributed node resources. Finally, we propose a parallelized LST-DDQN task offloading algorithm to maximize the delay and energy consumption weighted utility function. Simulation results demonstrate the efficacy of our proposed task decomposition and parallel scheduling methods, showcasing a reduction in utility by approximately 6.31% to 13.01% compared to baseline algorithms.

A wireless federated learning system is investigated by allowing a server and multiple workers to exchange uncoded information via orthogonal wireless channels. Since the workers frequently upload local gradients to the server via band-limited channels, the uplink transmission from the workers to the server becomes a communication bottleneck. Therefore, a one-shot distributed principle component analysis (PCA) is leveraged to reduce the dimension of uploaded gradients to relieve the communication bottleneck. A PCA-based wireless federated learning (PCA-WFL) algorithm and its accelerated version (i.e., PCA-AWFL) are proposed based on the low-dimensional gradients and the Nesterov’s momentum. For the non-convex empirical risk, a finite-time analysis is performed to quantify the impacts of system hyper-parameters on the convergence of the PCA-WFL and PCA-AWFL algorithms. The PCA-AWFL algorithm is theoretically certified to converge faster than the PCA-WFL algorithm. Besides, the convergence rates of PCA-WFL and PCA-AWFL algorithms quantitatively reveal the linear speedup with respect to the number of workers over the vanilla gradient descent algorithm. Numerical results are used to demonstrate the improved convergence rates of the proposed PCA-WFL and PCA-AWFL algorithms over the benchmarks.

Current mobile networks are facing dramatic growth in wireless traffics due to the prosperity of streaming media services. Cooperative edge caching, enabling multiple edge nodes to cache and share contents by exploiting the spatial/temporal user request differentiation, is regarded as a promising method to enhance Quality of Experience (QoE). However, frequent content sharing between BSs consumes operation cost such as the usage of cross-edge bandwidth and energy consumption. Therefore, new challenges incurred by performance-cost trade-off arise. In this paper, we propose a user preference-aware content caching and migration (PACM) scheme for video content delivery in a cost-efficient edge network. In this scheme, the dynamic user request preference and the long-term content migration cost budget are considered for content placement and delivery. To navigate a good performance-cost trade-off, we formulate the content caching and migration to be a long-term optimization problem. Then, the Lyapunov optimization method is used to decompose the problem into a series of real-time optimizations. As the decomposed problem is NP-hard, we design a novel collective reinforcement learning (CRL) algorithm that can realize online efficient decision-making by interacting with training experience. Simulation results show that the CRL algorithm has a high convergence rate and the proposed scheme can achieve quasi-optimal performance in terms of user-perceived latency, cache hit rate, and video stalling rate.

Intelligent reflecting surface (IRS)-assisted unmanned aerial vehicle (UAV) communications are expected to alleviate the load of ground base stations in a cost-effective way. Existing studies mainly focus on the deployment and resource allocation of a single IRS instead of multiple IRSs, whereas it is extremely challenging for joint multi-IRS multi-user association in UAV communications with constrained reflecting resources and dynamic scenarios. To address the aforementioned challenges, we propose a new optimization algorithm for joint IRS-user association, trajectory optimization of UAVs, successive interference cancellation (SIC) decoding order scheduling and power allocation to maximize system energy efficiency. We first propose an inverse soft-Q learning-based algorithm to optimize multi-IRS multi-user association. Then, successive convex approximation (SCA) and Dinkelbach-based algorithm are leveraged to optimize UAV trajectory followed by the optimization of SIC decoding order scheduling and power allocation. Finally, theoretical analysis and performance results show significant advantages of the designed algorithm in convergence rate and energy efficiency.

This work investigates the secure resource collaboration among selfish edge servers for multi-access edge computing (MEC)-enabled VR systems in a dynamic scenario. Due to the time-varying and stochastic nature of VR user requests, the edge servers usually have significant differences in workload. To this end, we first propose a type judgment method to perceive their service capability and divide them into two types, i.e., the requesting node (RN) with a poor service capability and the cooperative node (CN) with a powerful service capability. To promote collaboration among self-interest nodes, we then model the competitive interactions among RNs and CNs as a multi-leader and multi-follower Stackelberg game. For the RN (as the leader), we design a novel pricing strategy based on deep reinforcement learning (DRL) to motivate CNs to provide resource assistance. Meanwhile, an optimal selling strategy for the CN (as the follower) is presented to maximize its payoffs from the network. To overcome the security problem during the resource collaboration, we finally introduce the blockchain as a secure and trusted platform for resource publishing and trading, where an efficient consensus mechanism called Proof-of-Trust (PoT) is developed to improve the performance of blockchain. The simulation results show that the proposed approach achieves superior performance.

Cooperative multi-agent reinforcement learning (Co-MARL) commonly employs different parameter sharing mechanisms, such as full and partial sharing. However, imprudent application of these mechanisms can potentially constrain policy diversity and limit cooperation flexibility. Recent methods that group agents into distinct sharing categories often exhibit poor performance due to challenges in precisely differentiating agents and neglecting the issue of promoting cooperation among these categories. To address these issues, we introduce a dynamic selective parameter sharing mechanism embedded with multi-level reasoning abstractions (DSPS-MA). Our approach uses self-comparison sequences to infer agents’ abstract concepts, defining the differences between agents and allowing them to dynamically select partners to share parameters based on these abstract concepts. We also design an intrinsic reward to offer comprehensive collaboration guidance for agents, and introduce a policy cosine similarity regularization term to ensure sufficient policy diversity. Empirical evaluations demonstrate that our approach yields higher returns and faster convergence than state-of-the-art methods.

The super data transmission capability and connectivity of wireless technologies have promoted the arrival of the Internet of Things (IoT) era. However, the distinct characteristics of IoT devices make them vulnerable to malicious attacks such as hackers and viruses. This paper designs a novel IoT intrusion detection architecture that combines knowledge extraction and sharing, which can extract human understandable knowledge from the trained deep learning model and apply it to the training process of the detection model. The obtained knowledge can also be shared with other detection systems based on the blockchain, which will effectively improve the intrusion detection capabilities of the IoT and realize collective learning. In addition, we propose a CNN-based semi-supervised learning method under the constraints of rules, which can effectively alleviate the catas-trophic interference generated during the self-training process and improve detection accuracy. Simulation results confirm the effectiveness of the proposed architecture and method.

Federated Learning (FL) has become a common method for edge devices. Due to the limited energy capacity of edge devices, and the vulnerability of FL to malicious attacks from edge devices, vanilla FL still faces several challenges in edge computing, including energy consumption, model heterogeneity, and malicious behavior. To address these challenges, we propose a multi-granularity grouping-based federated learning (MG2FL), which groups and aggregates edge devices with low communication energy consumption and latency to reduce communication costs. Additionally, we introduce a multi-granularity guidance mechanism and a credit model to enhance model accuracy while ensuring security. Experimental results show that compared to the traditional FL algorithms, MG2FL achieves a 5.6% increase in accuracy, with the highest accuracy improvement reaching 11.1% in the presence of malicious edge devices.

With the increasingly sophisticated algorithms of simultaneous localization and mapping (SLAM), it is difficult for mobile terminals with limited resources to exploit the performance of SLAM algorithms fully. Traditional deep reinforcement learning (DRL)-based approaches have offloaded SLAM tasks to servers. However, existing solutions suffer from low data efficiency and poor generalization problems. This paper proposes a novel approach based on recent advances in rewardless active inference for SLAM back-end optimization. Specifically, the reward function is replaced with simple rewardless guidance in active inference. In addition, instead of simply considering the SLAM task as a whole, we delve into the sub-tasks of back-end optimization of SLAM for offloading and resource allocation. Simulation results show the superior performance of the proposed scheme.

The emergence of communication-based train control (CBTC) system within urban rail transport has improved the efficiency of safe train operations. At the same time, the CBTC system enhances the reliability of the train system and lowers latency. Nevertheless, there are still certain critical issues that need to be considered in the CBTC: 1) limited coverage and high maintenance costs of wayside equipment; 2) multiple ground devices configuration and complex system architecture; and 3) insufficient computing capacity of the train leads to heavy latency and energy consumption. The multi-hop ad hoc network coexisting with train-to-train communication and train-to-wayside communication is applied to simplify the networking architecture, together with the employment of mobile edge computing (MEC) servers to provide massive computing and communication resources for trains. Therefore, in this paper, a new multi-hop ad hoc network and MEC-assisted CBTC framework are developed for computing offloading and resource allocation. Offloading decisions, offloading ratio, computing and communication resource allocation are integrated to minimize latency and energy consumption. Furthermore, the proposed problem is a mixed-integer non-convex problem that is transformed into a solvable convex problem, and the consensus alternating direction method of multipliers-based (ADMM) algorithm is employed to solve the problem. The simulation results show that our proposed method has remarkable advantages over other schemes.

With the growth of the communication industry, the demand for spectrum resources has been increasing steadily. However, the spectrum resources are limited and the utilization rate is low. In this case, Cognitive Radio (CR) technologies and Non-Orthogonal Multiple Access (NOMA) technology are proposed to improve the number of user access and spectrum resource utilization. In this paper, we consider a CR-inspired NOMA network and propose a power allocation problem based on a time-varying system. Due to the dynamic nature of CR, traditional methods are often faced with problems such as low training efficiency and high network volatility. To address the problem, we propose a novel Quantum Reinforcement Learning (QRL) algorithm with quantum world model, interacting with the environment in the quantum way. Specifically, we construct a quantum soft actor-critic algorithm using variable quantum circuit (VQC), and the experience data is encoded into a quantum sum tree and transformed into quantum world model. Simulation results show the superior performance of the proposed framework.

The emerging artificial intelligence (AI) puts forward high requirement for big data acquisition, which is difficult to be met with the existing communication technologies in real time. In this paper, we investigate new graph learning based access management scheme for supporting the real-time big data acquisition in the sixth-generation mobile communication system (6G). We model the network scene with a mass of communication links as a fully connected graph which takes into account the accumulative interference of all links. Then, the joint admission and power control problem is formulated as a combinatorial optimization problem. We propose a graph attention network (GAT) based algorithm which can learn the system features by weighted aggregation of neighbor nodes. In addition, we construct a differentiable loss function that can accurately express the optimization objective and train the network by the change of loss. Based on the output of the GAT, we iteratively optimize the link admission and power to active more links. Simulation results demonstrate that the proposed algorithm is superior to the traditional convex optimization based algorithms and the nonmodified GAT based algorithms in the number of activated links. Moreover, the training of the constructed network is unsupervised with high computational efficiency, which makes them suitable for the big data access management.

Quantum reinforcement learning (QRL) can outperform classical reinforcement learning (RL) by utilizing quantum parallel theory and quantum phenomena such as superposition and entanglement. Although some excellent work has been done on QRL, most existing works either fail to show the exponential advantage of quantum computation over classical computation in terms of performance or are too demanding on quantum devices. In this paper, we provide a novel perspective on combining quantum computing and RL with faster convergence speed and relatively relaxed demands on quantum devices. Specifically, we propose a method to construct a world model with quantum circuit that allows it to interact in a quantum way. In addition, we use Grover's algorithm to efficiently extract high-value information from the quantum world model. Extensive simulation results show that the proposed method can have superior performance compared to classical RL algorithms.

Generic object detection methods have achieved preferable results, but it is still challenging to detect objects from complicated traffic scenes like extreme illumination and adverse weather. The existing methods are not robust enough to be extended to new complex traffic scenes. To address this issue, we leverage the idea of ensemble learning for strong robustness and propose a novel Bagging R-CNN framework. Specially, we design a bagging classification branch that uses adaptive sampling to train base learners and make them different from each other. The final predictions are the ensemble of the base learners, achieving strong robustness to the challenging objects. For localizing more accurately, a progressive regression branch is proposed in which bounding boxes are continuously optimized for high quality. Extensive experiment results on TJU-DHD-traffic and Pascal VOC datasets show that our Bagging R-CNN achieves superior detection accuracy over state-of-the-art methods. The source code can be found at https://github.com/PungTeng/BaggingRCNN.

Federated Learning (FL) is a novel distributed machine learning approach to leverage data from Internet of Things (IoT) devices while maintaining data privacy. However, the current FL algorithms face the challenges of non-independent and identically distributed (non-IID) data, which causes high communication costs and model accuracy declines. To address the statistical imbalances in FL, we propose a clustered data sharing framework which spares the partial data from cluster heads to credible associates through device-to-device (D2D) communication. Moreover, aiming at diluting the data skew on nodes, we formulate the joint clustering and data sharing problem based on the privacy-preserving constrained graph. To tackle the serious coupling of decisions on the graph, we devise a distribution-based adaptive clustering algorithm (DACA) basing on three deductive cluster-forming conditions, which ensures the maximum yield of data sharing. The experiments show that the proposed framework facilitates FL on non-IID datasets with better convergence and model accuracy under a limited communication environment.

Online learning has been widely applied in real-time artificial intelligence (AI) applications to learn new classes from the dynamic environment. Although the deployment of AI model training over the edge can facilitate faster processing of real-time data, the learning efficiency is plagued by the limited capacity of distributed data acquisition. In fact, not all data samples are equally important, and the random data selection strategy is not beneficial to accelerate training due to redundant data processing. In this paper, we present an importance-driven data collection framework, which leverages the usefulness of important data to improve the learning efficiency over the wireless edge. Specifically, the novel model convergence metric (MCM) is constructed to evaluate the data importance dynamically for model learning. Moreover, considering the constraint of limited network resources on learning efficiency, we establish an MCM maximization problem of joint data collecting, scheduling, and feeding in an edge computing system. A two-timescale hierarchical reinforcement learning (TTHRL) algorithm is designed to decouple the original problem into two-timescale two-level subproblems, where the top-level agent is responsible for data feeding strategy in the long term and the low-level agent learns data scheduling and collecting strategy in the short term. Simulation results show that our proposed scheme can achieve better performance improvements over the baseline schemes.

Audio-driven portrait animation aims to synthesize portrait videos that are conditioned by given audio. Animating high-fidelity and multimodal video portraits has a variety of applications. Previous methods have attempted to capture different motion modes and generate high-fidelity portrait videos by training different models or sampling signals from given videos. However, lacking correlation learning between lip-sync and other movements (e.g., head pose/eye blinking) usually leads to unnatural results. In this paper, we propose a unified system for multi-person, diverse, and high-fidelity talking portrait generation. Our method contains three stages, i.e., 1) Mapping-Once network with Dual Attentions (MODA) generates talking representation from given audio. In MODA, we design a dual-attention module to encode accurate mouth movements and diverse modalities. 2) Facial composer network generates dense and detailed face landmarks, and 3) temporal-guided renderer syntheses stable videos. Extensive evaluations demonstrate that the proposed system produces more natural and realistic video portraits compared to previous methods.

Accurately reconstructing 3D faces from monocular images and videos is crucial for various applications, such as digital avatar creation. However, the current deep learning-based methods face significant challenges in achieving accurate reconstruction with disentangled facial parameters and ensuring temporal stability in single-frame methods for 3D face tracking on video data. In this paper, we propose TokenFace, a transformer-based monocular 3D face reconstruction model. TokenFace uses separate tokens for different facial components to capture information about different facial parameters and employs temporal transformers to capture temporal information from video data. This design can naturally disentangle different facial components and is flexible to both 2D and 3D training data. Trained on hybrid 2D and 3D data, our model shows its power in accurately reconstructing faces from images and producing stable results for video data. Experimental results on popular benchmarks NoWand Stirling demonstrate that TokenFace achieves state-of-the-art performance, outperforming existing methods on all metrics by a large margin.

The emergence of Neural Radiance Fields (NeRF) has promoted the development of synthesized high-fidelity views of the intricate real world. However, it is still a very demanding task to repaint the content in NeRF. In this paper, we propose a novel framework that can take RGB images as input and alter the 3D content in neural scenes. Our work leverages existing diffusion models to guide changes in the designated 3D content. Specifically, we semantically select the target object and a pre-trained diffusion model will guide the NeRF model to generate new 3D objects, which can improve the editability, diversity, and application range of NeRF. Experiment results show that our algorithm is effective for editing 3D objects in NeRF under different text prompts, including editing appearance, shape, and more. We validate our method on both real-world datasets and synthetic-world datasets for these editing tasks. Please visit https://repaintnerf.github.io for a better view of our results.

Domain Adaptive Object Detection (DAOD) transfers an object detector from a labeled source domain to a novel unlabeled target domain. Recent works bridge the domain gap by aligning cross-domain pixel-pairs in the non-euclidean graphical space and minimizing the domain discrepancy for adapting semantic distribution. Though great successes, these methods model graphs roughly with coarse semantic sampling due to ignoring the non-informative noises and failing to concentrate on precise semantics alignment. Besides, the coarse graph generation inevitably contains abnormal nodes. These challenges result in biased domain adaptation. Therefore, we propose an Improved Graph Generation (IGG) framework which conducts high-quality graph generation for DAOD. Specifically, we design an Intensive Node Refinement (INR) module that reconstructs the noisy sampled nodes with a memory bank, and contrastively regularizes the noisy features. For better semantics alignment, we decouple the domain-specific style and category-invariant content encoded in graph covariance and selectively eliminate only the domain-specific style. Then, a Precision Graph Optimization (PGO) adaptor is proposed which utilizes the variational inference to down-weight abnormal nodes. Comprehensive experiments on three adaptation benchmarks demonstrate that IGG achieves state-of-the-art results in unsupervised domain adaptation.

The main idea of multimodal recommendation is the rational utilization of the item's multimodal information to improve the recommendation performance. Previous works directly integrate item multimodal features with item ID embeddings, ignoring the inherent semantic relations contained in the multimodal features. In this paper, we propose a novel and effective aTtention-guided Multi-step FUsion Network for multimodal recommendation, named TMFUN. Specifically, our model first constructs modality feature graph and item feature graph to model the latent item-item semantic structures. Then, we use the attention module to identify inherent connections between user-item interaction data and multimodal data, evaluate the impact of multimodal data on different interactions, and achieve early-step fusion of item features. Furthermore, our model optimizes item representation through the attention-guided multi-step fusion strategy and contrastive learning to improve recommendation performance. The extensive experiments on three real-world datasets show that our model has superior performance compared to the state-of-the-art models.

As the research and applications of large language model (LLM) become increasingly sophisticated, it is difficult for resource-limited mobile terminals to run large-model inference tasks efficiently. Traditional deep reinforcement learning (DRL) based approaches have been used to offload LLM inference tasks to servers. However, existing solutions suffer from data inefficiency, insensitivity to latency requirements, and non-adaptability to task load variations. In this paper, we propose an active inference with rewardless guidance algorithm using expected future free energy for offloading decisions and allocating resources for the LLM inference task offloading and resource allocation problem of cloud-edge networks systems. Experimental results show that our proposed method has superior performance over mainstream DRLs, improves in data utilization efficiency, and is more adaptable to changing task load scenarios.

To address the security issues in wireless transmission communication systems with finite-alphabet inputs, a secure uplink reception scheme based on receiver spatial modulation is proposed to ensure high security of the wireless transmission system while achieving high spectrum efficiency. The proposed scheme introduces disturbance phase updates at the transmitter end and disturbance compensation at the intelligent reflecting surface (IRS) to ensure both the spectrum efficiency and physical layer security of the wireless transmission system. This enables the legitimate receiver to correctly receive the signal at maximum power while preventing eavesdroppers from accurately intercepting the signal. In this paper, we propose two phase compensation schemes, namely, element-wise random disturbance compensation (ERDC) and group-wise random disturbance compensation. Furthermore, two eavesdropping scenarios, referred to as ideal eavesdropping and jamming eavesdropping, are thoroughly investigated and considered. We introduce the maximum likelihood detection to reliably detecte the indices of the designed receive antenna and the based-band modulated signal, and the closed-form expressions of error performance with ERDC scheme are deduced. Finally, in the simulated and numercial results, the performance analysis verifies error performance discrepancy based security performance evaluation of the proposed scheme, demonstrating its effectiveness and superiority.

With the development of autonomous vehicles and intelligent robots, visual simultaneous localization and mapping (SLAM) has attracted great attentions. Most existing visual SLAM systems assume that the objects are stationary in static environments. However, in the real world, there are many objects that are non-stationary in dynamic environments, which will cause performance degradation of visual SLAM systems. In this paper, to address this issue, we propose a novel visual SLAM system based on multi-task deep neural networks. Specifically, we apply multi-task deep neural networks to extract oriented keypoints and perceive dynamic semantic regions, which are used to perform outlier rejection in the SLAM system. We evaluate our method on public datasets, and the results show that our method outperforms existing visual SLAM systems. The presentation video url is: https://youtu.be/qGE1OvaJvV0.

The virtual world created by the development of the Internet, computers, artificial intelligence (AI), and hardware technologies have brought various degrees of digital transformation to people’s lives. With multiple demands for virtual reality increasing, the metaverse, a new type of social ecology that can connect the physical and virtual worlds, is booming. However, with the rapid growth of data volume and value, the continuous evolution of the metaverse faces the demands and challenges of privacy, security, high synchronization, and low latency. Fortunately, the ever-evolving blockchain and intelligent networking technologies can be used to satisfy the trusted construction, continuous data interaction, and computing demands of the metaverse. Therefore, it is necessary to conduct an in-depth review of the role and gains of blockchain, intelligent networking, and the combination of both in providing the immersive experiences of the metaverse. In this survey, we first discuss the development trend, characteristics, and architecture of the metaverse. Then, the existing work on blockchain, networking, and the combination of the two technologies are reviewed, including overviews, applications, and challenges. Next, applications of the metaverse are summarized, emphasizing the importance of the metaverse and the fields of development. Finally, we discuss some open issues, challenges, and future research directions.

As a promising approach to extending cloud resources and services, blockchain-enabled Internet of Things (IoT)-based smart grid edge computing has attracted much attention. However, the edge node’s resource-constraint nature makes it difficult to store the entire chain as the sensing IoT data volume increases. To address this issue, we propose an FS scheme, a fast and secure multithreshold trapdoor Chameleon hash scheme which serves as the basis for block substitution at the edge nodes to solve the storage limitation problem. The FS scheme is used to achieve a consensus-based block substitution, which allows$t$-out-of-$n$edge nodes to compute a hash collision collaboratively to reliably substitute a historical block without leaking the randomness$R$. Also, inspired by the rationale of fast polynomial interpolation, we optimize the FS scheme to FS-I to reduce the time complexity from$\mathcal {O}(nt)$to$\mathcal {O}(t{\mathrm{ log}}^{2}t)$. In addition, we further optimize FS-I to FS-II by using a fast Fourier transform (FFT) to dramatically improve the computational efficiency of Lagrange interpolation, which leads to a significant improvement in terms of block substitution performance. Finally, We provide security analysis and evaluate the performance through comprehensive experiments and the results show that FS can achieve up to several magnitudes better than DTTCH. The results also demonstrate that the FS scheme can provide high service quality for large-scale IoT-based smart grid systems.

Federated learning (FL) is an emerging technique to support privacy-aware and resource-constrained machine learning, where a base station (BS) will coordinate a set of distributed Internet of Things (IoT) devices to train a shared machine learning model with their local data sets. Nevertheless, due to the frequent interactions between BS and distributed IoT devices for the aggregating/distributing learning model parameters, the performance of FL is fundamentally restricted by the randomness of channel condition. To address this issue, we utilize the intelligent reflecting surface (IRS) to improve the efficiency of learning model aggregation/distribution. In addition, we consider two transmission protocols to enable the model aggregation from IoT devices to BS, i.e., frequency division multiple access (FDMA) and nonorthogonal multiple access (NOMA). For both protocols, we formulate the total training latency minimization problem under the available energy constraints of IoT devices, to jointly optimize the phase shifts of IRS, communication resource scheduling, and transmit power and local computing frequencies of IoT devices. Moreover, we further develop the efficient multidimensional resource management algorithms to solve the formulated training latency minimization problems. Numerical results demonstrate that the proposed IRS-assisted FL systems can achieve significant latency reduction as compared with other benchmark methods, and the NOMA-based model aggregation method exhibits a lower total training latency than the FDMA-based counterpart.

With the growing demand for latency-sensitive and compute-intensive services in the Internet of Things (IoT), multiaccess edge computing (MEC)-enabled IoT is envisioned as a promising technique that allows network nodes to have computing and caching capabilities. In this article, we propose a cache-aided MEC (CA-MEC) offloading framework for joint optimization of communication, computing, and caching (3C) resources in the MEC-enabled IoT. Our goal is to optimize the offloading decision and resource allocation strategy to minimize the system latency subject to dynamic cache capacities and computing resource constraints. We first formulate this optimization problem as a multiagent decision problem, a partially observable Markov decision process (POMDP). Then, the deep graph convolution reinforcement learning (DGRL) method is applied to motivate the agents to learn optimal strategies cooperatively in a highly dynamic environment. Simulations show that our method is highly effective for computation offloading and resource allocation and performs superior results in a large-scale network.

Assessment of the left ventricle segmentation in cardiac magnetic resonance imaging (MRI) is of crucial importance for cardiac disease diagnosis. However, conventional manual segmentation is a tedious task that requires excessive human effort, which makes automated segmentation highly desirable in practice to facilitate the process of clinical diagnosis. The Internet of Medical Things (IoMT) and artificial intelligence (AI) for efficient medical data collection and analysis have been deemed effective approaches to remote and automatic diagnosis. In this article, we propose a novel reinforcement-learning-based framework for left ventricle contouring, which mimics how a cardiologist outlines the left ventricle in a cardiac image. Since such a contour drawing process is simply moving a paintbrush along a specific trajectory, it is thus analogized to a path finding problem. Following the algorithm of proximal policy optimization (PPO), we train a policy network, which makes a stochastic decision on the agent’s movement according to its local observation such that the generated trajectory matches the true contour of the left ventricle as much as possible. Moreover, we design a deep learning model with a customized loss function to generate the agent’s landing spot (or coordinate of its initial position on a cardiac image). We further propose an alternative approach for generating the landing spot based on interventricular septum detection, which is more efficient since no extra effort in data preprocessing and model training is involved. The experimental results show that the coordinates of the generated landing spots with both of the two approaches are sufficiently close to the true contour and the proposed reinforcement-learning-based approach outperforms the existing U-net model and its improved version, even with a limited training set.

With the development of autonomous driving technology, it is foreseeable that connected and automated vehicles (CAVs) will be fully popularized in people’s lives. During this process, transportation systems are expected to evolve into the era of mixed autonomy, where CAVs and human-driven vehicles (HDVs) coexist in road networks and share available road resources. To materialize the much-anticipated potential of CAVs, a thorough understanding of CAVs’ effects on transportation systems is indispensable. On the one hand, attributing to advanced sensing, communication, and computation capabilities, CAVs provide opportunities to enhance mixed traffic safety, improve energy savings and suppress shockwave spread. On the other hand, due to advantages in large-scale information and cloud-computing resources, CAVs have the ability to occupy more road resources compared with HDVs, resulting in a reduction in the travel efficiency of HDVs, and even of the entire transportation systems. In this article, by clarifying the key differences between HDVs and CAVs, we comprehensively review the potential impacts of CAVs when they are appearing on road networks coexisting with HDVs. It can be regarded as the first-of-its-kind paper that systematically overviews the impacts of CAVs in the era of mixed autonomy on both positive and negative emotions. Specifically, the main focuses of this article are: 1) what are the key differences between CAVs and HDVs? 2) what are the positive impacts of CAVs’ appearance on mixed traffic systems? 3) will the introduction of CAVs cause some negative effects simultaneously? and 4) what kinds of strategies should be employed to relieve these negative effects? Hopefully, this article can not only call for an objective attitude toward the introduction of CAVs, but also provide foresighted advice to address possible challenges during the popularization of CAVs, so as to create a cooperative, safe, and efficient mixed traffic ecosystem.

This paper investigates the joint resource allocation and group formation for platoon in vehicle-to-everything (V2X) networks under vehicular channel uncertainty. To achieve the high spectrum efficiency and overcome the platoon head communication range limitation, an adaptive multicast-based group cooperation communication model is developed for the platoon with dynamic topology. Considering the heterogeneous characteristics of different types of links, i.e., high capacity for vehicle-to-infrastructure (V2I) links and ultra-reliability for vehicle-to-vehicle (V2V) links, we attempt to maximize the V2I capacity whilst satisfying a probability constraint for ultra-reliable V2V-supported intra-platoon communication. To handle the intractable probability constraint, a support vector clustering (SVC) based method is developed to capture the distributional geometry of massive uncertain channel samples as a sphere in high-dimensional feature space with asymmetric structure. Based on it, the probability constraint is transformed into a tractable linear convex set. After that, an exploration-selection-alternating-iterative algorithm is developed to solve the formulated problem with coupled optimization variables. Specifically, in the exploration process, a two-stage algorithm is proposed for the resource allocation problem under fixed group formation decision, which includes power control and spectrum allocation. During the selection process, a performance difference-based decision transition rate is designed to optimize group formation solution. Simulation results demonstrate the proposed data-driven approach can overcome the over-conservatism of the traditional symmetric-geometry-based uncertainty sets, and the multicast-based group cooperation communication model corresponds to a higher performance on V2I capacity than other traditional schemes.

Graph-based clustering has become an active topic due to the efficiency in characterizing the relationships between the samples via graph. To improve the quality of graph, recent works propose to utilize global and local information. However, existing methods may lead to a degenerated graph when facing noisy and uneven distributed data. Since 1) they preserve the local information by referring the similarity between each sample-pair, whose confidence is easily disturbed by the poor quality samples; and 2) although the global information is relatively robust to the noisy, existing methods have island effect that lies between local and global structures learning, such that the information of both can not be utilized mutually. To alleviate these issues, this paper presents explicit local coupling global structure clustering (ELGSC) to explicitly learn the local structure and global structure information via a coupling scheme. To be specific, we learn$l(\ll n)$pseudo samples as the anchors to reflect local hot spots distribution, where$n$is the number of samples. By referring the relationship between each anchor-sample pair, ELGSC is capable of obtaining an effective local bipartite graph to capture the local structure. Meanwhile, the self-expressiveness learning is adopted to pursue a lower-rank global affinity graph. Finally, a higher-order coupling learning framework is proposed to couple the learning of global affinity graph and local bipartite graph. Thus, local and global structure information could be propagated each other on both graphs. The experimental results on real datasets demonstrate the efficacy of the proposed method over state-of-the-arts.

Serverless edge computing is seen as a promising enabler to execute differentiated Industrial Internet of Things (IIoT) applications without managing the underlying servers and clusters. In IIoT serverless edge computing, IIoT workflow scheduling for cloud-edge collaborative processing is closely related to the service quality of users. However, serverless functions decomposed by IIoT applications are limited in their deployment at the edge due to the resource-constrained nature of edge infrastructures. In addition, the scheduling of complex IIoT applications supported by serverless computing is more challenging. Therefore, considering the limited function deployment and the complex dependencies of serverless workflows, we model the workflow application as directed acyclic graph and formulate the scheduling problem as a multiobjective optimization problem. A dueling double deep Q-network-based solution is proposed to make scheduling decisions under dynamically changing systems. Extensive simulation experiments are conducted to validate the superiority of the proposed scheme.

There has been a notable increase in the research and development of Vehicular Ad-hoc Networks (VANETs) to efficiently and safely manage large amounts of traffic. Such networks are, however, also prone to various cyber threats to data integrity, privacy, authentication, and network availability, and given the potential risk to life under the event of a malfunction and misinformation, it is important to provide security measures against such threats. This paper presents the Multi-branch Reconstruction Error (MbRE) Intrusion Detection System (IDS) for edge-based anomaly detection in VANETs for data integrity, network availability and user authentication-based misbehaviors without the need to train on them. Vehicular data is first sequenced and separated into three data branches - frequency (F) derived from the message timestamps, pseudo-identities (I), and the motion data (M) i.e. position and velocity. The proposed model comprises of three Convolutional Neural Networks (CNN)-based reconstruction models trained to reconstruct normal F-I-M vehicular behavior. The IDS classifies each branch of a sequence as 0/1 based on the reconstruction error threshold for the respective branch and, therefore, has the ability to detect 8 possible binary encoded behaviors for each sequence of vehicular data. These results are then used to find the overall behavior of each vehicle using carefully selected detection thresholds. MbRE is able to classify frequency, identity and motion-based behavior samples with an accuracy of 100%, 98.5-100%, and 95.4-100%, respectively, without the need to train on such behaviors. The study also emulates the IDS on Google Colaboratory and Jetson Nano to show its practicality in cloud and edge environments.

Currently, the complex traffic environment challenges the fast and accurate response of a connected autonomous vehicle (CAV). More importantly, it is difficult for different CAVs to collaborate and share knowledge. To remedy that, this paper proposes a selective federated reinforcement learning (SFRL) strategy to achieve online knowledge aggregation strategy to improve the accuracy and environmental adaptability of the autonomous driving model. First, we propose a federated reinforcement learning framework that allows participants to use the knowledge of other CAVs to make corresponding actions, thereby realizing online knowledge transfer and aggregation. Second, we use reinforcement learning to train local driving models of CAVs to cope with collision avoidance tasks. Third, considering the efficiency of federated learning (FL) and the additional communication overhead it brings, we propose a CAVs selection strategy before uploading local models. When selecting CAVs, we consider the reputation of CAVs, the quality of local models, and time overhead, so as to select as many high-quality users as possible while considering resources and time constraints. With above strategic processes, our framework can aggregate and reuse the knowledge learned by CAVs traveling in different environments to assist in driving decisions. Extensive simulation results validate that our proposal can improve model accuracy and learning efficiency while reducing communication overhead.

Federated learning (FL), as a distributed machine learning technology, allows large-scale nodes to utilize local datasets for model training and sharing without revealing privacy, which has significant efficiency and advantages in artificial intelligence (AI)-based knowledge sharing of connected and autonomous vehicles (CAVs). However, for FL, there are challenges to ensure the security of knowledge, deal with the lazy behavior of participants, and enforce effective incentives. To bridge the gaps, in this paper, we first propose a hierarchical blockchain-supported FL architecture that utilizes the immutable and transparent properties of blockchain to enable secure storage and sharing of knowledge and transaction information with scalability. Then, considering the cost and laziness of the participants in the FL process, we propose an incentive mechanism combined with the supervision game to attract high-quality participants based on a comprehensive evaluation of model quality and participants’ reputation. Extensive simulation results validate that our proposal can improve learning accuracy and efficiency while ensuring security.

Edge intelligence (EI) is becoming one of the research hotspots among researchers, which is believed to help empower intelligent transportation systems (ITS). ITS generates a large amount of data at the network edge by millions of devices and sensors. Data-driven artificial intelligence (AI) is at the core of ITS development. By pushing the AI frontier to the network edge, EI enables ITS AI applications to have lower latency, higher security, less pressure on the backbone network and better use edge big data. This paper surveys Edge Intelligence in Intelligent Transportation Systems. We first introduce the challenges ITS faces and explain the motivation of using EI in ITS. We then explore the framework of using EI in ITS, including the EI-based ITS architecture, the data gathering and communication methods, the data processing and service delivery, and the performance indexes. The enabling technologies, such as AI models, the Internet of Things, and Edge Computing technologies used in EI-based ITS, are reviewed intensively. We discuss the edge intelligence applications and research fields in ITS in depth. Typical application scenarios, such as autonomous driving, vehicular edge computing, intelligent vehicular transportation system, unmanned aerial vehicle (UAV) in ITS environment, and rail transportation control and management, are explored. The general platforms of EI, the EI training and inference in ITS, as well as the benchmark datasets, are introduced. Finally, we discuss some of the challenges and future directions of using EI in ITS.

The increasingly intense cyber-attacks have always been a crucial issue to the communication-based train control (CBTC) system due to exposed wireless channels. Both cyber-attack intrusion detection and defense policy calculation demand substantial computing resources. Combined with high capacity and reliability 5G technologies, edge intelligence (EI) is believed to help empower CBTC systems in terms of security and efficiency. This paper proposes an EI-enabled structure for CBTCs to defend against cyber-attacks, where the EI server provides real-time intelligent computing services for trains to derive real-time defense policies. We formulate the cyber-attack and defense process in EI-enabled CBTCs as a two-level game model, where system security and edge computing resource allocation are jointly optimized. In the lower-level game, we model interactions between the cyber attacker and system defender as a discrete repeated security game (DRSG), which is also a non-zero sum and incomplete information game. The fictitious play (FP) is introduced to derive a Nash equilibrium (NE) based optimal defense scheme. In the upper-level game, considering that the EI server cannot simultaneously update the optimal defense scheme for all trains due to the limited computation resources, we construct a multi-stage computation resource allocation game (MCRAG). We derive the optimal computation resource allocation scheme by the neural fictitious self-play (NFSP), where a deep Q-learning network (DQN) and a supervised learning network are jointly built to learn the strategy. Extensive simulation results show that our proposed EI-enabled CBTC system and the two-level game model can effectively defend against various attacks.

Due to the high integration of wireless communication and networking technologies, the communication-based train control (CBTC) systems are exposed to additional cyber-attack surfaces, allowing sophisticated attackers to combine cyber attack vectors with physical attack means to achieve malicious goals. Notably, the decentralized authentication features are missing in existing communication protocols which make the CBTC be easily compromised by data tampering attacks, and lead to serious operational accidents. With outstanding advantages in decentralized authentication, blockchain provides new effective solutions for decentralized identity authentication in CBTC. Consequently, it is critical to study the complex physical consequences of cyber breaches from a cross-layer defense perspective. In this paper, we propose a novel cross-layer defense method for cyber security in blockchain empowered CBTC against data tampering attacks. In the physical layer, the joint Kalman filter and$\chi ^{2} $detector is proposed for the train state estimation and detection. In the cyber layer, an asymmetric encryption-based secure communication protocol with identity authentication and the blockchain-based distributed key management system with the adaptive consensus mechanism are designed for data communication security. Considering the unavailable direct observation of the CBTC cyber security states, a partially observable Markov (POMDP) decision model is constructed to derive the optimal adaptive consensus strategies for balancing cyber security and efficiency. Extensive simulation results show that the proposed blockchain empowered CBTC cross-layer defense method can effectively improve the cyber security protection capability and minimize the impact of data tampering attacks on the train operation.

With the large-scale deployment of the Internet of Vehicles (IoV) and 5G technologies, it is inevitable to share data frequently for superior in-vehicle services. However, due to the dynamically changing and widely distributed Vehicular Ad-hoc Networks (VANETs), data sharing still faces challenges in security, efficiency, and reliability. In this paper, we propose a secure and reliable data-sharing scheme (SDSS) for edge-enabled IoV networks. It assigns multiple attribute authorities to alleviate the management burden and support a large attribute universe catering to the various services in IoV. To enhance efficiency and flexibility, edge computing is introduced for quickly responding to vehicles’ requests and assisting resource-constrained vehicle computation. And an online/offline mechanism is designed to further alleviate the computational pressure of sharing data online. In addition, we put forward a cooperative key generation approach to guarantee the security of users’ private keys. The security analysis proves that SDSS ensures resistance to collusion attacks and indistinguishability under chosen-ciphertext attacks (IND-CCA). Moreover, it can avoid the single point of failure and resist denial of service (DoS) attacks with the help of multiple distributed edge nodes. The experiment demonstrates SDSS is practicable for IoV data sharing.

Connectionist artificial intelligence (AI) can power many critical tasks for connected and autonomous vehicles (CAVs). However, connectionist AI lacks interpretability and usually needs large amount of data for learning. A Markov logic network (MLN), which combines first-order logic (FOL) with statistical learning, learns weighted FOL formulas for inference. MLNs can incorporate domain expert knowledge in the form of FOL formulas to achieve data-efficient learning and transparent decision process. In this paper, we propose a hybrid driving decision-making system, which integrates a MLN module and a deep Q-network (DQN) for enhanced driving safety. The MLN module evaluates the safety of ranked actions from DQN to reduce potential collisions. A collective MLN (Co-MLN) learning algorithm is proposed and it enables CAVs collectively learn a global MLN model for safe state transitions, given distributed small amount of noisy data. A hybrid DQN-MLN learning algorithm is also developed for CAVs to collectively learn to drive in new driving environments. Simulations performed using a highway driving simulator show that the proposed Co-MLN algorithm is highly data-efficient and the learned hybrid driving system can effectively reduce collisions. In addition, the learned MLN module provides transparency for safety-critical driving decisions.

Air-to-ground (A2G) communication based on Unmanned aerial vehicle (UAV) is an important part of the future communication system. In this paper, an A2G channel model with UAV three-dimensional (3D) wobbles (pitch, roll, and yaw) based on the geometry-based stochastic model (GBSM) is proposed. On this basis, the UAV’s internal vibration is modeled as a sinusoidal random process, and the UAV wobble caused by the atmospheric flow is modeled as the uniform distribution random process. We derive the channel temporal correlation function (CF) with UAV 3D wobbles, analyze the variation of the temporal CF with different carrier frequencies, and amplitudes of the wobble angles. It is found that, even if the UAV wobbles slightly, the channel temporal correlation will be significantly affected. Numerical results show that the channel CF will decrease rapidly with the increase of the amplitudes of wobble angles and the carrier frequency. Therefore, the coherence time of millimeter wave (mmWave) band is significantly less than that of sub-6 GHz band. The consistency of simulation results and measurement results in published papers ensures the availability of the proposed model. For the MUAVs scenario, when the distance between different UAVs is much greater than the wavelength, the A2G channels between different UAVs and user equipment (UE) on the ground are not correlated to each other, and the temporal auto-correlation function (ACF) of each UAV is the same as that of the SUAV scenario. This work contributes to the theoretical exploration and system design of A2G communication based on UAV.

Data sparsity and cold start problems are common in recommender systems. Adding some side information, such as knowledge graph and users' trust relationship, is an effective method to alleviate these problems. However, few work jointly explore the fine-grained implicit relationships between the external heterogeneous graphs to enhance the recommendation accuracy. To address this issue, in this paper, we propose a new method named Trust-aware Multi-task Knowledge Graph (TMKG), which uses multi-task learning to integrate two kinds of side information of trust graph and knowledge graph in an end-to-end manner. Firstly, we mine the intra-graph and inter-graph high-order connections through the node propagation and aggregation, and optimize the embedding of nodes through the implicit relationships obtained. Furthermore, through the shared cross unit, the connection relationships between each layer is mined, and the high-order interaction of nodes of different layers is obtained. We conduct extensive experiments on real-world datasets and prove that our model has the superior performance compared with the state-of-the-art models.

Edge intelligence is emerging as a new interdiscipline to push learning intelligence from remote centers to the edge of the network. However, with its widespread deployment, new challenges arise in terms of training efficiency and service of quality (QoS). Massive repetitive model training is ubiquitous due to the inevitable needs of users for the same types of data and training results. Additionally, a smaller volume of data samples will cause the over-fitting of models. To address these issues, driven by the Internet of intelligence, this paper proposes a distributed edge intelligence sharing scheme, which allows distributed edge nodes to quickly and economically improve learning performance by sharing their learned intelligence. Considering the time-varying edge network states including data collection states, computing and communication states, and node reputation states, the distributed intelligence sharing is formulated as a multi-agent Markov decision process (MDP). Then, a novel collective deep reinforcement learning (CDRL) algorithm is designed to obtain the optimal intelligence sharing policy, which consists of local soft actor-critic (SAC) learning at each edge node and collective learning between different edge nodes. Simulation results indicate our proposal outperforms the benchmark schemes in terms of learning efficiency and intelligence sharing efficiency.

Ambient backscatter communication is an energy efficient technique for massive Internet of Things (IoT). Combining with flexibly deployed unmanned aerial vehicles (UAVs), the UAV-aided ambient backscatter communication can establish wireless links for isolated IoT nodes efficiently. In this paper, we investigate the outage performance of the UAV-aided air-ground network with underlaid ambient backscatter communications, where the emitted signals from the air-ground link are leveraged as radio frequency (RF) carrier for ambient backscattering. The air-ground channel is modeled as a probabilistic line-of-sight (LoS) channel with Nakagami-$m$fading. Then, the ground communication is modeled as a non-line-of-sight (NLoS) channel with Rayleigh fading. For the downlink, we derive the expressions of the outage probabilities of the backscatter link and the air-ground link. In addition, the asymptotic cases of infinite transmit power and infinite fading parameter are analyzed. For the uplink, the outage probabilities of the backscatter link and air-ground are analyzed, with the cases of infinite transmit power and fading parameter discussed. Simulation results show that the analytical results match well with the Monte Carlo results, which verifies the effectiveness of the proposed scheme.

With the rapid development of intelligent transportation systems, there is an increasingly strong demand for low-latency and high-bandwidth vehicular services. Unmanned aerial vehicles (UAVs) can be used as a supplement to the ground networks, to relieve the communication pressure on ground facilities, such as base stations. In this paper, we use multiple UAVs to provide services for vehicles and model the multi-UAV scenario as a collaborative multi-agent system. In addition, we take vehicle safety as the top priority and the delay requirement as the constraints. Then we exploit the Lagrange multiplier to combine the constraint function and cost function, so as to reduce the resource consumption as much as possible on the premise of ensuring the safety of the vehicles. The influence of spectrum efficiency and computing power should also be taken into account when allocating resources. We adopt the multi-agent reinforcement learning to train the UAVs, and meanwhile introduce the attention mechanism so that each UAV can optimize itself better with the information of other UAVs. Through extensive simulations, the effectiveness of our proposed method is verified. Particularly, the limited resources can allocated efficiently according to the vehicle’s needs under the premise of ensuring vehicle safety.

Multi-access edge computing (MEC) enables the computation-intensive and latency-critical application to be processed at the network edge, which reduces the transmission latency and energy consumption. The quality of the wireless channel seriously affects the performance of the edge network. Consequently, the performance of the edge network can be significantly improved from the perspective of communication. The recently advocated intelligent reflecting surface (IRS) intelligently controls the radio propagation environment to improve the quality of wireless communication links. This paper proposes an edge heterogeneous network with the assistance of intelligent reflecting surface. Specifically, the macro base station and small base stations are equipped with MEC servers, and IRS is adopted to provide an additional computation offloading link. The user association, computation offloading and resource allocation, as well as IRS phase shift design are optimized with the aim of minimizing the long-term energy consumption subject to the constraints imposed on quality of service (QoS) and available resources. The challenge of the optimization problem is rooted from the fact that update timescale of user association is different from others. Hence, a two-timescale mechanism is invoked by marrying tools from matching theory and deep reinforcement learning. More specifically, the user association decision takes place in the long timescale. In the short timescale, the computation offloading, resource allocation and IRS phase shift design strategy is performed. The effectiveness of the proposed two-timescale mechanism is verified by the simulation results.

Combining with flexibly deployed unmanned aerial vehicles (UAVs) and energy-efficient ambient backscatter communication, the UAV-aided ambient backscatter communication can establish wireless links for isolated IoT nodes efficiently. In this paper, we investigate a UAV air-ground networks with underlaid ambient backscatter communications, where the emitted signal from the UAV is leveraged as radio frequency (RF) carrier for ambient backscattering. First, we establish a system model of the UAV air-ground networks with underlaid ambient backscatter communications. Then, the expressions of the outage probabilities for both the backscatter link and the air-ground link are derived. In addition, the asymptotic outage probabilities of infinite transmit power and infinite fading parameter are analyzed. Simulations show that the analytical results match well with the Monte Carlo results, which verifies the effectiveness of the proposed scheme.

The space-air-ground (SAG) integrated networks will play a major role in the sixth generation (6G) mobile networks, which will provide global coverage, full connection and pervasive intelligence services for multiple ground Internet of Things (IoT) devices. Moreover, massive computing tasks can be either performed by local devices, or offloaded to edge servers, such as low orbit satellites, high altitude platforms (HAPs) and remote base stations. Nevertheless, the joint computation and communication resource allocation solutions are becoming challenging due to the large-scale state space, time-varying network scenarios, and limited battery capacity. In this paper, we propose a SAG-integrated three-layer heterogenous network model to maximize the sum-rate of ground IoT devices, which further enhances the deep integration of communication and computation resources. Additionally, we develop a Lyapunov-assisted multi-agent proximal policy optimization algorithm to process the task scheduling, HAP selection, battery harvesting, and CPU cycle frequency optimization. Extensive simulation results corroborate that the proposed method has superior performance gains in terms of the remaining battery capacity, energy consumption, and maximum average sum-rate compared with the state-of-the-art baselines.

Driven by numerous emerging mobile devices and various quality of service requirements, mobile edge computing (MEC) has been recognized as a prospective paradigm to promote the computation capability of mobile devices, as well as reduce energy overhead and service latency of applications for the Internet of Things (IoT). However, there are still some open issues in the existing research works: 1) limited network and computing resource, 2) simple or non-intelligent resource management, 3) ignored security and reliability. In order to cope with these issues, in this article, 6G and blockchain technology are considered to improve network performance and ensure the authenticity of data sharing for the MEC-enabled IoT. Meanwhile, a novel intelligent optimization method named as collective reinforcement learning (CRL) is proposed and introduced, to realize intelligent resource allocation, meet distributed training results sharing and avoid excessive consumption of system resources. Based on the designed network model, a cloud-edge collaborative resource allocation framework is formulated. By joint optimizing the offloading decision, block interval and transmission power, it aims to minimize the consumption overheads of system energy and service latency. Then the formulated problem is designed as a Markov decision process, and the optimal strategy can be obtained by the CRL. Some evaluation results reveal that the system performance based on the proposed scheme outperforms other existing schemes obviously.

Vehicular edge computing is an emerging enabler to support vehicular-based computation-intensive tasks. By reason of the time-varying vehicular wireless environments and the stochastic task generation, the dynamically unbalanced task load distribution among resource-constrained edge infrastructures leads to the performance bottleneck and low efficiency of computation resource utilization. We employ an aerial relay station that can establish relay connections between vehicles and nearby heterogeneous edge infrastructures to relieve this situation. The computation offloading strategy design in the multivehicle multi-edge infrastructure scenario that is closely linked to system latency performance will be particularly complicated. To address this issue, a model-free multi-agent reinforcement learning is adopted, and we propose a practical constraint in the problem formulation. Simulation experiments show that the proposed strategy can guarantee load balancing among edge infrastructures.

Wireless systems are upgraded to use green energy (e.g., solar, wind, and tide energy) such that the greenhouse gas emission can be neutralized. This work incorporates the on-grid energy into a green coordinated multi-point (CoMP) system to handle the volatile arrival of green energy. In the green CoMP, the long-term weighted throughput maximization problem is investigated by expecting a non-positive consumption of the long-term on-grid energy. Motivated by the capacity-achieving property and simple implementation, an online zero-forcing dirty paper precoder is proposed to update the precoding matrices by combining statistical learning with the Lyapunov learning. A tradeoff relation is theoretically established to show that the long-term weighted throughput approaches the $\mathcal{O}(V)$ -neighbor of optimal value while the long-term consumed on-grid energy increases at a rate of $\mathcal{O}\left( {{{\log }^2}(V)/\sqrt V } \right)$, where V is an introduced control parameter. Numerical results are used to verify the performance of the online zero-forcing dirty paper precoder.

In this paper, an integrated intelligent reflecting surface (IRS)-unmanned aerial vehicle (UAV) communication scheme is proposed where the IRS is mounted on the UAV as a mobile relay between the base station (BS) and the ground user. We present two schemes to maximize the spectrum efficiency (SE) and the energy efficiency (EE) of the system by jointly optimizing the active beamforming, passive beamforming and UAV trajectory. First, to tackle the SE maximization problem, we divide it into three sub-problems to optimize the variables iteratively. For the active and passive beamforming, the closed-form solutions can be directly derived. The suboptimal trajectory design can be obtained by utilizing the successive convex approximation (SCA). Furthermore, considering the limited on-broad energy of UAV, a scheme to maximize the EE is proposed. The optimal active beamforming and the passive beamforming can be similarly obtained. For the non-convex fractional programing of trajectory optimization, it can be solved via the Dinkelbach’s method. Numerical results demonstrate that the effectiveness of the proposed algorithms.

The Unmanned Aerial Vehicle (UAV) swarm has attracted widespread attention from both academia and industry. It has been widely adopted in disaster recovery, military communication, agricultural production, and industrial automation. In critical situations or places where communication infrastructure is lacking, deploying a UAV swarm network is a cost-effective solution. However, considering the high speed of UAV devices, designing an effective routing mechanism has been a challenging problem. In this paper, enlightened by the recent success of multi-agent reinforcement learning, we propose a multi-agent policy gradients-based UAV routing algorithm. We adopt a centralized training and decentralized executing framework, where a centralized training platform is implemented to guide the policy updating of each UAV node. Moreover, we introduce a counterfactual baseline scheme in our algorithm to improve the convergence speed. Extensive simulation results validate the effectiveness of the proposed algorithms compared to the state-of-the-art schemes.

Immutability, decentralization, and linear promoted scalability make sharded blockchain a promising solution, which can effectively address the trust issue in the large-scale Internet of Things (IoT). However, currently, the throughput of sharded blockchains is still limited when it comes to high proportions of cross-shard transactions (CST). On the other hand, assemblage characteristics of collaborative computing in IoT have not been received attention. Therefore, in this paper, we present a clustering-based sharded blockchain strategy for collaborative computing in the IoT, where the sharding of the blockchain system is implemented in two steps: k-means clustering-based user grouping and the assignment of consensus nodes. In this framework, how to reasonably group the IoT users while simultaneously guaranteeing the system performance is the key point. Specifically, we describe the data transactions among IoT devices by data transaction flow graph (DTFG) based on a dynamic stochastic block model. Then, formed as a Markov decision process (MDP), the optimization of the cluster number (shard number) and the adjustment of consensus parameters are jointly trained by deep reinforcement learning (DRL). Simulation results show that the proposed scheme improves the scalability of the sharded blockchain in the IoT application.

Deep reinforcement learning (DRL) has been studied in a variety of challenging decision-making tasks, e.g., autonomous driving. \textcolor{black}{However, DRL typically suffers from the action shaking problem, which means that agents can select actions with big difference even though states only slightly differ.} One of the crucial reasons for this issue is the inappropriate design of the reward in DRL. In this paper, to address this issue, we propose a novel way to incorporate the smoothness of actions in the reward. Specifically, we introduce sub-rewards and add multiple constraints related to these sub-rewards. In addition, we propose a multi-constraint proximal policy optimization (MCPPO) method to solve the multi-constraint DRL problem. Extensive simulation results show that the proposed MCPPO method has better action smoothness compared with the traditional proportional-integral-differential (PID) and mainstream DRL algorithms. The video is available at https://youtu.be/F2jpaSm7YOg.

In recent years, multi-access edge computing (MEC) is emerging to provide computation and storage resources to the Internet of things (IoT) devices to assist them in high-performance demanding tasks. Real-time task requests from the IoT devices often have strict delay constraints. However, in practice, the delay requirements of task requests often fail to be satisfied because of the inappropriate computation processing method and inefficient resource allocation in MEC networks. In this article, we present a novel framework for MEC networks with unmanned aerial vehicles (UAVs) and intelligent reflecting surfaces (IRSs) to facilitate computation offloading with delay constraints. In addition, we propose a novel constrained reinforcement learning method with a dynamic balance mechanism named$Q_{c}-DQN$. Finally, we conduct extensive simulations to verify the effectiveness of our proposed method. Compared to the benchmark schemes, our scheme not only improves the overall network performance and reduces the task completion time, but also meets the delay constraints.

In recent years, multi-access edge computing (MEC) is emerging to provide computation and storage capabilities to the Internet of things (IoT) devices to improve the quality of service (QoS) of IoT applications. In addition, intelligent reflecting surface (IRS) techniques have attracted great interests from both academia and industry to improve the communication efficiency. Although existing works leverage the IRS technique in MEC networks, they mainly focus on the single-IRS single-area scenario. However, in practice, multi-IRS will be deployed in multi-area scenarios in future networks. Consequently, considering the single-IRS single-area scenario will have inferior performance. In this paper, to address the aforementioned issue, we propose an efficient resource provisioning scheme for multi-IRS multi-area scenarios in MEC networks. We first model the problem as a cooperative multi-agent reinforcement learning process, where each agent manages one area and all agents share the network bandwidth and computation resources. Then, we propose a multi-agent actor-critic method with an attention mechanism for resource management with latency guarantee. Finally, we conduct extensive simulations to verify the effectiveness of the proposed scheme. Our scheme can reduce the required computation resources by up to 11.84% when compared with the benchmark works. It is also shown that our proposed scheme can improve the efficiency of resource allocation and scale well with the increasing demand from IoT devices.

In this paper, we design a task scheduling scheme for Internet of Things (IoT) workflow applications in serverless edge computing. Notice the fact that complex applications in traditional serverless computing are decomposed into several stateless, dependent functions, whose execution environments are pre-deployed at the resource-finite edge domain, we model the workflow application as Directed Acyclic Graph (DAG) by considering the distribution of edge resources and the deployment of serverless functions. We further formulate the scheduling problem as a multi-objective optimization problem to reduce the time consumption, energy consumption, and cost simultaneously. Then, considering the diversity of solution space and the fast convergence to optimal solutions, an improved hybrid algorithm that combines Particle Swarm Optimization and Genetic Algorithm (PSO–GA) is introduced and utilized to make the scheduling decision. Finally, extensive simulation experiments are conducted to validate the superiority of the proposed scheme.

Machine learning (ML) algorithms are essential components in autonomous driving. In most existing connected and autonomous vehicles (CAVs), a large amount of driving data collected from multiple vehicles are sent to a central server for unified training. However, data privacy and security have become crucial during the data-sharing process. Federated learning (FL) for data security has arisen nowadays, and it can improve the data privacy of distribute machine learning. However, the malicious attackers can still be able to attack the training process. Due to the complete reliance on the central server, FL is very fragile. To address the above problem, we propose Bift: 1) a fully decentralized ML system combined with FL and 2) blockchain to provide a privacy-preserving ML process for CAVs. Bift enables distributed CAVs to train ML models locally using their own driving data and then to upload the local models to get a better global model. More importantly, Bift provides a consensus algorithm named Proof of Federated Learning to resist possible adversaries. We evaluate the performance of Bift and demonstrate that Bift is scalable and robust, and can defend against malicious attacks.

Compared with 5G networks, 6G networks are guaranteed to provide various tailored end-to-end network services and emerging cloud-edge applications. Network slicing (NS) is regarded as the key enabler of 6G networks. Softwarization and virtualization technologies, such as software-defined networking and network function virtualization, are accelerating the way toward NS of 6G networks. The resource allocation issue in 6G NS is very crucial, worthy more research attention. In this article, we propose one efficient resource allocation algorithm, labeled asTailoredSlice-6G, so as to realize the tailored slices in 6G. When receiving one slice request, ourTailoredSlice-6Gwill identify the slice resource type in the first place. Then, ourTailoredSlice-6Gwill select its most suitable subalgorithm to do the resource allocation and slicing deployment. Each type of slice corresponds to its specific resource allocation subalgorithm, inserted in theTailoredSlice-6Galgorithm. In addition, each subalgorithm inTailoredSlice-6Gis guaranteed to run within polynomial time. Thus,TailoredSlice-6Ghaving the potential to be promoted to real networking application. To highlight the merits ofTailoredSlice-6G, we do the comprehensive simulation. Simulation results vividly reveal that ourTailoredSlice-6Goutperforms the selected heuristics that are representative in the literature.

Colonoscopy screening is one of the most effective diagnostic tools for detecting intestinal diseases, such as bleeding, polyp, Meckel’s diverticulum, and ulcer. However, the missed rate of manual detection is high due to the lack of experience or fatigue among clinicians. To address this issue, this work proposed a novel autodiagnosis framework built on Internet of Medical Things (IoMT) systems, which can be deployed among multiple hospitals in a distributed fusion. This work presents a two-stage knowledge distillation (TSKD) method coupled with Bayesian optimization (BO) that can exploit distributed colonoscopy data to learn a compact diagnostic model achieving a good tradeoff between predictive performance and resource consumption (e.g., memory and computation). The proposed framework is extensively evaluated in real-world data sets in comparison with its counterparts. A prototype of on-site diagnostic device is implemented to demonstrate the potential for real-world deployment.

Industrial Internet of Things (IIoT) systems connect a plethora of smart devices, such as sensors, actuators, and controllers, to enable efficient industrial productions in manners observable and controllable by human beings. Plain model-based and data-driven diagnosis approaches can be used for fault detection and isolation of specific IIoT components. However, the physical models, signal patterns, and machine learning algorithms need to be carefully designed to describe system faults. Besides, the ever-increasing level of connectivity among devices can induce exponential complexity. Knowledge-based fault diagnosis approaches improve interoperability via ontologies so that high-level reasoning and inquiry response can be provided to nonexpert users. Therefore, knowledge-based fault diagnosis approaches are preferred over plain model-based and data-driven diagnosis approaches in recent IIoT systems. In the context of IIoT systems, this work reviews the recent progress on the construction of knowledge bases via ontologies and deductive/inductive reasoning for knowledge-based fault diagnosis. Besides, general inductive reasoning methods are discussed to shed light on their successful applications in knowledge-based fault diagnosis for IIoT systems. Following the trend of large-system decentralization, future fault diagnosis also requires decentralized implementations. Therefore, we conclude this survey by discussing several interesting open problems for decentralized knowledge-based fault diagnosis for IIoT systems.

With the rapid development of wireless communication technologies, the emerging multimedia applications make mobile Internet traffic grow explosively while putting forward higher service requirements for the next-generation wireless networks. Therefore, how to achieve low-latency content transmission by effectively allocating heterogeneous network resources to improve the network quality of service and end-user quality of experience is a key issue to be solved urgently in the current Internet. In this article, we propose a deep reinforcement learning (DRL)-based resource allocation scheme to improve content distribution in a layered fog radio access network (FRAN). We formulate the optimal resource allocation problem as a minimal delay model, where in-network caching is deployed and the same content requests from mobile users can be aggregated in the queue of each base station. To cope with the increasing user requests and overcome capacity constraints of the FRAN, moreover, a cloud–edge cooperation offloading scheme is utilized in our model, where the integrated allocation of caching, computing, and communication resources and joint optimization between in-network caching and routing are considered to promote resource utilization and content delivery. In our solution, a new DRL policy is designed to make cross-layer cooperative caching and routing decisions for the arriving content requests according to request history information and available network resources in the system. Simulation results demonstrate that our proposed model can performs much better than the existing cloud–edge cooperation schemes in the FRAN.

In the last few decades, ciphertext-policy attribute-based encryption (CP-ABE) technology has attracted great interest, since it can provide fine-grained, flexible, and access control for sensitive data to implement a high secure and efficient data-sharing mechanism. In this article, based on the linear secret sharing scheme (LSSS), an efficient scheme is proposed to realize a collaborative decryption function. For any user group, when the user’s attribute set cannot access the ciphertext alone, the private key of other users in the same group can be used for collaborative decryption with the permission of the data owner. Our scheme uses the LSSS matrix that can significantly reduce the computation and storage overhead when comparing with the existing schemes. Then, a multiauthorization model is created based on the Bohen–Lynn–Shacham technology in order to solve the key-management issue. Finally, we implemented the specific functions of the framework through JAVA, and built a private chain to verify the feasibility of data transfer between users.

Video surveillance in intelligent transportation systems (ITSs) is in the rapid growth stage, where video analytics is a potential technology to improve the safety of the Internet of Autonomous Vehicles (IoAV). However, massive video data transmission and computation-intensive video analytics bring an overwhelming burden for vehicular networks. Moreover, owing to the unstable network connection, the video data are not always reliable, which makes data sharing a lack of security and scalability in IoAV. In this work, we first propose a video analytics framework, where the multiaccess edge computing (MEC) and blockchain technologies are integrated into IoAV to optimize the transaction throughput of the blockchain system as well as reducing the latency of the MEC system. Furthermore, based on deep reinforcement learning, the joint optimization problem is modeled as a Markov decision process (MDP), and the asynchronous advantage actor–critic (A3C) algorithm is adopted to solve this problem. Simulation results demonstrate that our approach can fast converge and significantly improve the performance of blockchain-enabled IoAV with MEC.

Driven by numerous emerging mobile devices and various Quality-of-Service (QoS) requirements, mobile-edge computing (MEC) has been recognized as a prospective paradigm to promote the computation capability of mobile devices, as well as reduce energy overhead and service latency of applications for the Internet of Things (IoT). However, there are still some open issues in the existing research works: 1) limited network and computing resource; 2) simple or nonintelligent resource management; and 3) ignored security and reliability. In order to cope with these issues, in this article, 6G and blockchain technology are considered to improve network performance and ensure the authenticity of data sharing for the MEC-enabled IoT. Meanwhile, a novel intelligent optimization method named as collective reinforcement learning (CRL) is proposed and introduced, to realize intelligent resource allocation, meet distributed training results sharing, and avoid excessive consumption of system resources. Based on the designed network model, a cloud–edge collaborative resource allocation framework is formulated. By joint optimizing the offloading decision, block interval, and transmission power, it aims to minimize the consumption overheads of system energy and latency. Then, the formulated problem is designed as a Markov decision process, and the optimal strategy can be obtained by the CRL. Some evaluation results reveal that the system performance based on the proposed scheme outperforms other existing schemes obviously.

Low earth orbit (LEO) small satellites have attracted great interests in civilian and military applications due to their low cost and high service performance. However, the enormous scale and high dynamism of small satellites pose challenges to network flexibility and scalability. Therefore, the hierarchical satellite network structure is introduced as an effective approach to enhance the satellite network capabilities further. In this regard, small satellites’ clustering is of fundamental importance for designing such a hierarchical structure. Satellite clusters are always prone to instability due to unpredictable link failures and frequent topology changes. In this article, we study the small satellite clustering problem of jointly optimizing the cluster reliability and the network management overhead. A coalition game-theoretic framework is introduced to obtain low computational complexity by adopting the clustering-decision-making process in an automated and fully distributed fashion. A distributed coalition formation algorithm based on the optimization of reliability and management overhead is developed for the clustering problem. Finally, extensive simulations have been conducted, and the results show that our proposed clustering scheme is able to produce better results than the baseline schemes.

By delegating the infrastructure management, such as provisioning or scaling to third-party providers, serverless edge computing has recently been widely adopted in several applications, especially Internet of Things (IoT) applications. Task scheduling is a critical issue in serverless edge computing as it significantly impacts the quality of user experience. In contrast to the centralized scheduling in the cloud center, serverless edge task scheduling is more challenging due to the heterogeneous and resource-constrained nature of edge resources. This article aims to study the distributed task scheduling for the IoT in serverless edge computing networks, in which heterogeneous serverless edge computing nodes are rational individuals with interests to optimize their own scheduling utility while the nodes only have access to local observations. The task scheduling competition process is formulated as a partially observable stochastic game (POSG) to enable serverless edge computing nodes to noncooperatively schedule tasks and allocate computing resources depending on their locally observed system state, which takes into account the associated task generation state, data queue state, communication channel state, and previous computing resource allocation state. To solve the proposed POSG and deal with the partial observability, a multiagent task scheduling algorithm based on the dueling double deep recurrent$Q$-network (D3RQN) method is developed to approximate the optimal task scheduling and resource allocation solution. Finally, extensive simulation experiments are conducted to validate the effectiveness and superiority of the proposed scheme.

Based on the trustless feature of blockchain, this article designs a general configurable blockchain-enabled smart manufacturing system to achieve flexible manufacturing in response to large-scale manufacturing services. With a transaction pool containing all the pending manufacturing tasks but aligning with the logic flow, the complex manufacturing structure can be uniformly tackled. Furthermore, in virtue of the contradiction between large-scale manufacturing and limited blockchain throughput, we formulate a joint optimization of the block size, task scheduling, and the supply-demand configuration to maximize the customers’ net profit with the probabilistic delay requirements, which addresses the critical issue of efficiency and latency in the blockchain-based live manufacturing process. Meanwhile, the production quality and price preference are involved. For solution, a mixed online bipartite matching-based DQN algorithm is proposed, which circumvents the high dimensionality by separating the task-manufacturer matching from the time-correlated problem. Simulation results show that the proposed flexible framework can well adopt to dynamic customer population, and achieves better convergence.

Recent developments in the Internet of Vehicles (IoV) enable interconnected vehicles to support ubiquitous services. Various emerging service applications are promising to increase the Quality of Experience (QoE) of users. On-board computation tasks generated by these applications have heavily overloaded the resource-constrained vehicles, forcing it to offload on-board tasks to other edge intelligence (EI)-assisted servers. However, excessive task offloading can lead to severe competition for communication and computation resources among vehicles, thereby increasing the processing latency, energy consumption, and system cost. To address these problems, we investigate the transmission-awareness and computing-sense uplink resource management problem and formulate it as a time-varying Markov decision process. Considering the total delay, energy consumption, and cost, quantum-inspired reinforcement learning (QRL) is proposed to develop an intelligence-oriented edge offloading strategy. Specifically, the vehicle can flexibly choose the network access mode and offloading strategy through two different radio interfaces to offload tasks to multiaccess edge computing (MEC) servers through WiFi and cloud servers through 5G. The objective of this joint optimization is to maintain a self-adaptive balance between these two aspects. Simulation results show that the proposed algorithm can significantly reduce the transmission latency and computation delay.

The growing number of Internet of Things (IoT) devices brings enormous time-sensitive applications, which require real-time transmission to effectuate communication services. The ultrareliable and low-latency communication (URLLC) scenario in the fifth generation (5G) has played a critical role in supporting services with delay-sensitive properties. Time-sensitive networking (TSN) has been widely considered as a promising paradigm for enabling the deterministic transmission guarantees for 5G. However, TSN is a hybrid traffic system with time-sensitive traffic and best effort traffic, which require effective routing and scheduling to provide a deterministic and bounded delay. While joint optimization of time-sensitive and non-time-sensitive traffic greatly increases the solution space and brings a significant challenge to obtain solutions. Therefore, this article proposes a graph convolutional network-based deep reinforcement learning (GCN-based DRL) solution for the joint optimization problem in practical communication scenarios. The GCN is integrated into deep reinforcement learning (DRL) to obtain the network’s spatial dependence and elevate the generalization performance of the proposed method. Specifically, the GCN adopts the first-order Chebyshev polynomial to approximate the graph convolution kernel, which reduces the complexity of the algorithm and improves the feasibility for the joint optimization task. Furthermore, priority experience replay is employed to accelerate the convergence speed of the model training process. Numerical simulations demonstrate that the proposed GCN-based DRL algorithm has good convergence and outperforms the benchmark methods in terms of the average end-to-end delay.

Immutability, decentralization, and linear promoted scalability make the sharded blockchain a promising solution, which can effectively address the trust issue in the large-scale Internet of Things (IoT). However, currently, the throughput of sharded blockchains is still limited when it comes to high proportion of cross-shard transactions (CSTs). On the other hand, the assemblage characteristic of the collaborative computing in IoT has not been received attention. Therefore, in this article, we present a clustering-based sharded blockchain strategy for collaborative computing in the IoT, where the sharding of the blockchain system is implemented in two steps:K-means-clustering-based user grouping and the assignment of consensus nodes. In this framework, how to reasonably group the IoT users while simultaneously guaranteeing the system performance is the key point. Specifically, we describe the data transactions among IoT devices by data transaction flow graph (DTFG) based on a dynamic stochastic block model. Then, formed as a Markov decision process (MDP), the optimization of the cluster number (shard number) and the adjustment of consensus parameters are jointly trained by deep reinforcement learning (DRL). Simulation results show that the proposed scheme improves the scalability of the sharded blockchain in the IoT application.

Internet of Vehicles (IoV) is one of the most promising applications of Internet of Things (IoT) in the automotive industry, which can empower moving vehicles to exchange information with neighboring cars, roadside infrastructure, remote servers, traffic control centers, and so on. IoV expects to support a wide range of vehicular services, such as road safety, path planning, infotainment, and smart parking, which will play a vital role in intelligent transportation systems (ITSs)[1]–[3]. The main enabling platforms for IoV consist of dedicated short-range communications (DSRCs)-based networks and cellular networks (C-V2X). However, these terrestrial networks alone might not be able to support the vehicular applications well in all the cases and scenarios, due to the issues of limited coverage and capacity, as well as costly deployment.

In this article, we focus on a downlink cellular network, where multiple unmanned aerial vehicles (UAVs) serve as aerial base stations for ground users through frequency-division multiple access (FDMA). With user locations and channel parameters inaccessible, the UAVs coordinate to make a decision on resource allocation and trajectory design in a decentralized way. Aiming at optimizing both overall and fairness throughput, we model resource allocation and trajectory design as a decentralized partially observable Markov decision process (Dec-POMDP) and propose multiagent reinforcement learning (RL) as a solution. Specifically, we use parameterized deep$Q$-network (P-DQN) for the action space comprising both discrete and continuous actions and the QMIX framework is leveraged to aggregate each UAV’s local critics. For fairness throughput optimization, we introduce an entropy-like fairness indicator to the reward to make the total return decomposable. In addition, we further propose a novel distributed learning framework for overall throughput optimization such that each UAV can contribute its local gradient, and model training can be implemented in parallel without need of observation data sharing among the UAVs. Simulation results show that the proposed multiagent RL approach as well as the distributed learning framework are efficient in model training and present acceptable performance close to that achieved by deterministic optimization, which relies on convention optimization techniques with user locations and channel parameters explicitly known beforehand. For fairness throughput optimization, we also show that ground users achieve individual throughputs close to each other, which verifies the effectiveness of the proposed fairness indicator as the reward definition in the RL framework.

A novel paradigm that combines federated learning with blockchain to empower edge intelligence over vehicular networks (FBVN) can enable latency-sensitive deep neural network-based applications to be executed in a distributed pattern. However, the complex environments in FBVN make the system latency much harder to minimize by traditional methods. In this article, we model the training and transmission latency of each autonomous vehicle (AV) and consensus latency of the blockchain in-edge side in FBVN. Considering the dynamic and time-varying wireless channel conditions, unpredictable packet error rate, and unstable data sets quality, we adopt duel deep$Q$-learning (DDQL) as the solving approach. We propose a federated DDQL algorithm, in which the learning agent is deployed on each AV side, and the sensing states on each AV do not need to be shared so that it increases scalability and flexibility for practical implementation. Simulation results show that the proposed algorithm has better performance in reducing system latency compared with the other schemes.

Orbital angular momentum (OAM) has the potential to dramatically enhance the amount of information in the Internet of Underwater Things (IoUT) system. Nevertheless, underwater-turbulence-induced scintillation will destroy the orthogonality of OAM modes, hence degrading the performance of the system. In this article, a random-amplitude-mask-based adaptive optics (AOs) technique is proposed for the sake of mitigating the turbulence effects in the OAM-based underwater wireless optical communication (UWOC) system. Combined with phase retrieval algorithms, the magnitudes of linear measurements obtained from the distorted OAM beams modulated with a series of random amplitude masks and focused by a lens are employed for the phase estimation. Furthermore, we present a comprehensive performance comparison against state-of-the-art phaseless wave-front sensing techniques. Moreover, the mixture exponential-generalized gamma (EGG) distribution is applied for characterizing the probability density function (PDF) of reference-channel irradiance of OAM beams coupled into a single-mode fiber (SMF). In the end, the performance metrics, such as the outage probability, the average bit-error-rate (BER), and the ergodic capacity are analyzed with the aid of PDF for both single-input-single-output (SISO) and multiinput-multioutput (MIMO) systems. In a nutshell, this article provides new insights for the applications of AO in the OAM-based UWOC system, which can serve as a candidate for supporting IoUT devices.

Spatial beamforming is an efficient way to realize dual-functional radar-communication (DFRC). In this paper, we study the DFRC design for a general scenario, where the dual-functional base station (BS) simultaneously detects the target as a multiple-input-multiple-output (MIMO) radar while communicating with multiple multi-antenna communication users (CUs). This necessitates a joint transceiver beamforming design for both MIMO radar and multi-user MIMO (MU-MIMO) communication. In order to characterize the performance tradeoff between MIMO radar and MU-MIMO communication, we first define the achievable performance region of the DFRC system. Then, both radar-centric and communication-centric optimizations are formulated to achieve the boundary of the performance region. For the radar-centric optimization, successive convex approximation (SCA) method is adopted to solve the non-convex constraint. For the communication-centric optimization, a solution based on weighted mean square error (MSE) criterion is obtained to solve the non-convex objective function. Furthermore, two low-complexity beamforming designs based on CU-selection and zero-forcing are proposed to avoid iteration, and the closed-form expressions of the low-complexity beamforming designs are derived. Simulation results are provided to verify the effectiveness of all proposed designs.

Viewport prediction and prefetch have an important influence on VR video streaming performance. This work proposes a novel federated learning-based viewport prediction model training algorithm, ComPer-FedAvg. The proposed algorithm leverages a VR video’s common viewing pattern and users’ personal viewing patterns to train the prediction model in a distributed and privacy-preserving manner. Further, considering the VR video viewport prediction accuracy, a stochastic game is formulated to solve the VR streaming network’s communication resource allocation problem, where limited communication resource blocks are auctioned to users to achieve the optimal overall VR viewing experience. For each user, the auction is decomposed into two disjoint subproblems, namely, the optimal number of data rate requesting and true value claiming (bidding). The optimal true value claiming has been analytically proved to be equal to the VR viewing reward with given data rate. Due to the lack of global information when users request data rate, we reformulate users’ data rate requesting problem as a POMDP problem. A novel deep reinforcement learning algorithm is adopted to solve the problem. Evaluation and simulation results show the proposed viewport prediction and VR streaming schemes outperform conventional solutions in terms of prediction accuracy and VR viewing experience.

Object Detection in low-light aerial images is a challenging problem due to considerable variation in brightness and varying contrast. Deep Learning-based approaches have recently demonstrated great promise in image enhancement. Many existing neural networks used for image quality enhancement first encode the input into low-resolution representations and then decode these representations back to a higher resolution for the contextual information. However, this method leads to the loss of semantic content. Recent research has demonstrated the advantage of maintaining high-resolution information along with lower resolution representations, which maintains image features throughout the network. In this paper, we propose a novel architecture named RNet for low-light image enhancement of aerial images. The proposed network contains multi-resolution branches for better understanding of different levels of local and global context through different streams. The performance of RNet is evaluated on a recent synthetic dataset. We also present a comprehensive evaluation with a representative set of state-of-the-art enhancement techniques and neural net architectures.

In this paper, we study the resource allocation in high mobility vehicle-to-everything (V2X) networks with only slowly varying large-scale channel parameters. For satisfying the diversity requirements of different types of links, i.e., low delay for vehicle-to-infrastructure (V2I) connections and ultra-reliability for vehicle-to-vehicle (V2V) connections, we formulate a joint power, spectrum and vehicle local computing ratio allocation problem to minimize the delay of V2I links whilst satisfying the V2V reliability constraint. For solving the formulated problem, a Feasible Region Transformation Method is firstly developed to convert the probabilistic V2V reliability requirement into a computable constraint. In addition, a Robust Signal to Interference Plus Noise Ratio (SINR) Modified Method is proposed to give the computable expression for the V2I throughput. Then, a parallel Deep Neural Network (DNN) framework is designed for the resource allocation in V2X networks, where one is the transmit power control unit and the other is the local computing ratio allocation unit. After that, a Feedback-oriented Learning Method is proposed to train the parallel DNN-based resource allocation framework, in which the output of DNN is used as feedback to dynamically revise the training loss function along with the training process. Afterwards, the Hungarian method is employed to obtain the optimal spectrum matching. Finally, we conduct the simulations to show that the proposed learning-based algorithm has better performance compared with other general algorithms.

The vigorous developments of the Internet of Things make it possible to extend its computing and storage capabilities to computing tasks in the aerial system with the collaboration of cloud and edge, especially for artificial intelligence (AI) tasks based on deep learning (DL). Collecting a large amount of image/video data, unmanned aerial vehicles (UAVs) can only hand over intelligent analysis tasks to the back-end mobile edge computing (MEC) server due to their limited storage and computing capabilities. How to efficiently transmit the most correlated information for the AI model is a challenging topic. Inspired by task-oriented communication in recent years, we propose a new aerial image transmission paradigm for the scene classification task. A lightweight model is developed on the front-end UAV for semantic block transmission with the perception of images and channel states. To achieve the tradeoff between transmission latency and classification accuracy, deep reinforcement learning (DRL) is applied to explore the semantic blocks which have the greatest contribution to the back-end classifier under various channel states. Experimental results show that the proposed method can significantly improve classification accuracy by more than 4% under the same conditions, compared to other semantic saliency learning methods.

Unmanned aerial vehicles (UAVs) have been widely employed in wireless communications, and the performance can be enhanced with the assistance of intelligent reflecting surface (IRS). However, the finite energy of UAVs greatly limits the endurance and becomes a bottleneck for IRS-UAV communications. In this paper, an integrated IRS-UAV communication scheme is proposed where the IRS is mounted on the UAV to connect the base station and the ground user. We present two schemes to maximize the spectrum efficiency (SE) and the energy efficiency (EE) of the system by jointly optimizing the active beamforming, passive beamforming and UAV trajectory. First, to tackle the SE maximization problem, we divide it into three subproblems to optimize the variables iteratively. For the active and passive beamforming, the closed-form solutions can be directly derived. The suboptimal trajectory design can be obtained by utilizing the successive convex approximation. Furthermore, considering the limited on- broad energy of UAV, a scheme to maximize the EE is proposed. The optimal active beamforming and the passive beamforming can be similarly obtained. For the non-convex fractional programming of trajectory optimization, it can be solved via the Dinkelbach’s method. Numerical results demonstrate that the proposed algorithms are effective for the IRS-UAV networks to maximize the SE and EE, respectively.

Computation over multiple access channels (CoMAC) has been proposed to solve the problem of spectrum scarcity in wireless networks, which combines communication and computation efficiently using the superposition property of wireless channels. In this paper, we consider a multi-cluster CoMAC network, whose performance is affected by the inter-cluster interference and the non-uniform fading. To minimize the sum mean squared error of signals aggregated at different fusion centers (FCs), we propose a transceiver design for multi-cluster CoMAC. Specifically, we adopt a uniform-forcing transmitter design to formulate the receiver design as a quadratic sum-of-ratios problem with nonconvex quadratic constraints. Then, we propose a branch-and-bound algorithm to find its optimal solution with a given error tolerance. To solve the problem in a decentralized way, we develop a distributed algorithm based on the primal decomposition theory. Each subproblem is solved by using the successive convex approximation method. Further combining Lagrange duality, we derive the optimal solution structure of each subproblem, based on which we can find the solution with lower complexity. Simulation results demonstrate the effectiveness of the proposed distributed transceiver design.

The performance of distributed computing is restricted by the slowest worker nodes, known as stragglers, in the system. Coded computation has emerged as an efficient technique to mitigate the straggler effects in distributed computing. Most existing works only considered the computation straggler for single-hop networks. However, in multi-hop networks, the straggler effects will occur not only on worker nodes but also on relay nodes. In this paper, we consider a heterogeneous multi-hop network. The nodes in the network are heterogeneous, i.e., their computation capacities and transmission capacities are different. We propose a hierarchical coding scheme for such a network. Firstly, we reorganize it into a hierarchical network containing multiple layers. Each layer in the network consists of several groups. Then, a new hierarchical coding scheme is proposed, where coding is applied to each group to mitigate the stragglers. By taking both the computation time and transmission time into consideration, the overall task completion time is derived. To improve the performance of the network, heterogeneous hierarchical coded computation (HHCC) algorithm is proposed to provide an asymptotically optimal task allocation strategy. Compared with existing uniform uncoded, load balanced uncoded, and heterogeneous coded matrix multiplication schemes, HHCC has significant improvement.

Currently, increasing attention is devoted to artificial intelligence (AI) enabled software defined industrial networks (AI-SDINs). Toward energy-efficient and secure data transmission in AI-SDINs, a metric called energy-cost-per-useful-bit (ECPUB), which means energy cost of transmitting per useful bit is presented to evaluate energy efficiency and a novel energy-efficiency based secure multipath routing scheme is then put forward. Specifically, the ECPUB incorporates the utility and the law of diminishing marginal utility, for revealing the relationship among energy consumption, residual energy, and useful bits required. Moreover, in this article, an energy-efficiency based secure multipath routing scheme (E2SMR) is proposed by adopting the ECPUB and (t,n) threshold secret sharing scheme, for enhancing the security under the premise of guaranteeing energy efficiency. Extensive simulation results show that ECPUB can evaluate the energy efficiency and facilitate the balance of network load, while E2SMR can prolong the lifetime of the network and simultaneously ensure the network functionality securely.

Accurately discovering hazards and issuing appropriate warnings to drivers in advance or performing autonomous control is the core of the Collision Avoidance (CA) system used to solve traffic safety problems. More comprehensive environmental awareness, diversified communication technologies, and autonomous control can make the CA system more accurate and effective, thereby improving driving safety. In addition, the assistance of Artificial Intelligence (AI) technology can make the CA system adapt to the environment and facilitate fast and accurate decisions. Considering the current lack of a thorough survey of driving safety with sensing, vehicular communications, and AI-based collision avoidance, in this paper, we survey existing researches for state-of-the-art data-driven CA techniques. Firstly, we discuss the major steps of CA and key research issues. For each step, we review the existing enabling techniques and research methods for CA in detail, including sensing and vehicular communication for safe driving, as well as CA algorithm design. Particularly, we present a comparison between the most common AI algorithms for different functions in the CA system. Testbeds and projects for CA are summarized next. Finally, several open challenges and future research directions are also outlined.

The complex traffic and road environment pose considerable challenges to the accuracy, timeliness, and adaptive ability of connected and autonomous vehicles (CAVs) in making driving decisions. This paper uses vehicle collaboration and integrates the adaptive learning capabilities of machine learning and the interpretation capabilities of expert systems (ESs) in a unified architecture to form a hybrid autonomous driving guidance system, which not only solves the “bottleneck” of knowledge acquisition during the construction of expert systems but also solves the “black box” phenomenon of machine learning in the decision-making process. First, an autonomous driving strategy based on deep reinforcement learning (DRL) is proposed for CAVs to make decisions and extract corresponding rules. Next, we design an ES knowledge base expansion method including rule extraction, rule sharing, and rule test. Particularly, vehicular blockchain is adopted to ensure user privacy and data security during the rule-sharing process. Third, hybrid autonomous driving guidance combining ES and machine learning is proposed for CAVs to make accurate and efficient decisions in different driving environments. Once the strategy is well trained, it can effectively guide CAVs to cope with the complex traffic environment. Extensive simulations validate the performance of our proposal in terms of decision-making accuracy, effectiveness, and safety.

With the rapid development of intelligent transportation systems, there is an increasing demand for a variety of vehicular services, such as automated driving assistance, emergency alert, infotainment, etc. However, in some situations (e.g., remote areas or maritime scenarios), the terrestrial networks alone cannot serve the vehicular applications very well due to the infrastructure deployment and maintenance issues. Satellite networks have become an effective supplement to terrestrial networks, which complement well in terms of coverage, flexibility, reliability, and availability. In this paper, we consider the low orbit multi-beam satellite-terrestrial networks to serve for vehicles. We model this problem as a cooperative multi-agent reinforcement learning process, where each beam acts as an agent, and the global bandwidth is cooperatively shared among all the agents. A multi-agent actor-critic method with attention mechanism is proposed to allocate resources for vehicles with strict delay requirements and minimum bandwidth consumption. When allocating bandwidth, the channel efficiency, the angle of the beams and the priorities of requests in different regions are also considered. Centralized training and distributed execution is performed in the training of the agents. Extensive simulation results verify the effectiveness of our proposed method, where all the agents can well cooperative to achieve efficient resource allocation on-demand for the vehicles under strictly limited bandwidth resources.

In intelligent transportation systems (ITS), the vehicular ad-hoc network (VANET) is an enabling technology that can provide information exchange services among connected and autonomous vehicles (CAVs). Video streaming over VANETs is a potential application to ensure the safety of drivers and passengers and improve infotainment services. However, owing to the dynamic network topology, video transmission in VANETs is very challenging in terms of latency, reliability, and security. Therefore, a comprehensive summary of the state-of-art video streaming over VANETs is surveyed in this work. Firstly, related works and background knowledge are introduced. Then, a systematic survey on resource allocation (RA) scheme for video streaming in VANETs is provided, and some prevailing and feasible optimization tools are elaborated. Furthermore, enabling technologies of video streaming over VANETs are summarized with a special focus on the integration of video communication, caching, and computing. Finally, we give some challenges and future research directions.

As the general development trend of the automotive industry, connected and autonomous vehicles (CAVs) can be used to increase transportation safety, promote mobility choices, reduce user costs, and create new job opportunities. However, with the increasing level of connectivity and automation, malicious users are able to easily implement different kinds of attacks, which threaten the security of CAVs. Hence, this paper provides a comprehensive survey on the cyber-security in the environment of CAVs with the aim of highlighting security problems and challenges. Firstly, based on the types of communication networks and attack objects, it classifies various cyber-security risks and vulnerabilities in the environment of CAVs into in-vehicle network attacks, vehicle to everything network attacks, and other attacks. Next, it regards cyber-risk as another type of attacks in the environment of CAVs. Then, it describes and analyzes up-to-date corresponding defense strategies for securing CAVs. In addition, it concludes several available cyber-security and safety standards of CAVs, which is helpful for the practical application of CAVs. Finally, several challenges and open problems are discussed for the future research.

Digital mymargin Twin (DT) and Intelligent Reflective Surface (IRS), the most two promising technologies of 6G make the Internet of Vehicles (IoV) more adaptive. However, future autonomous driving needs powerful networking resources and high-quality wireless communications to guarantee the Quality of Service (QoS). Especially considering the time-varying physical operating environments of IoV, it is extremely urgent to improve resource utilization and wireless channel quality. In this work, we propose a Digital Twin-Driven Vehicular Task Offloading and IRS Configuration Framework (DTVIF) to efficiently monitor, learn, and manage the IoV. Specifically, we adopt Mobile Edge Computing (MEC) and IRS to provide augmented computing capacities for vehicles and improve transmission performance when vehicles communicate to MEC servers. DT is employed to achieve real-time data collection and digital representation of physical operating environments of IoV to better support decisions making. In order to reduce the overall delay and energy consumption of DTVIF, we propose a Two-Stage Optimization for Jointly Optimizing Task Offloading and IRS Configuration (TSJTI) algorithm based on Deep Reinforcement Learning (DRL) and Transfer Learning (TFL). In the first stage, we introduce Double Deep$Q$-learning Networks (DDQN) to find the optimal offloading decision. In the second stage, based on the parameters learned from the first stage, we migrate the parameters from the first stage to find the optimal IRS configuration based on the Deep Deterministic Policy Gradient (DDPG) method. The simulations demonstrate that the proposed algorithm can effectively reduce the processing latency of task offloading and reduce the average energy consumption in DTVIF.

Vehicular ad hoc networks (VANETs) have become an important branch of future 6G smart wireless communications. As an emerging key technology, multi-access edge computing (MEC) provides low-latency, high-speed, and high-capacity network services for the VANETs. In this paper, we propose a novel framework for blockchain-based, hierarchical multi-access edge computing for the future VANET ecosystem (BMEC-FV). In the underlying VANET environment, we propose a trust model to ensure the security of the communication link between vehicles. Multiple MEC servers calculate the trust between vehicles through computing offloading. Meanwhile, the blockchain system plays an important role to manage the entire BMEC-FV architecture. We aim to optimize the throughput and the quality of services (QoS) for MEC users in the lower layer of the system architecture. In this framework, the main challenge is how to effectively reach consensus among blockchain nodes while ensuring the performance of MEC systems and blockchains. The blocksize of blockchain nodes, the number of consensus nodes, reliable features of each vehicle, and the number of producing blocks for each block producer are considered in a joint optimization problem, which is modeled as a Markov decision process with state space, action space, and reward function. Since it is difficult for this to be solved by traditional methods, we propose a novel deep compressed neural network scheme. Simulation results illustrate the superiority of the BMEC-FV ecosystem.

The appropriate design of a vehicular ad hoc network (VANET) has become a pivotal way to build an efficient smart transportation system, which enables various applications associated with traffic safety and highly-efficient transportation. VANETs are vulnerable to the threat of malicious nodes stemming from its dynamicity and infrastructure-less nature and causing performance degradation. Recently, software-defined networking (SDN) has provided a feasible way to manage VANETs dynamically. In this article, we propose a novel software-defined trust based VANET architecture (SD-TDQL) in which the centralized SDN controller is served as a learning agent to get the optimal communication link policy using a deep$Q$-learning approach. The trust of each vehicle and the reverse delivery ratio are considered in a joint optimization problem, which is modeled as a Markov decision process with state space, action space, and reward function. Specifically, we use the expected transmission count ($ETX$) as a metric to evaluate the quality of the communication link for the connected vehicles’ communication. Moreover, we design a trust model to avoid the bad influence of malicious vehicles. Simulation results prove that the proposed SD-TDQL framework enhances the link quality.

Following the roadmap of carbon neutrality, wireless communication systems are upgrading to use green energy that comes from renewable sources, e.g., sun, tide, and wind. Due to the volatile arrival of green energy, the on-grid energy is used as a backup for a green coordinated multiple point system. In this work, a weighted sum-rate maximization problem in thegreencoordinated multiple point system is investigated by expecting non-positive consumption of the on-grid energy in the long term. Motivated by the capacity-achieving property and simple implementation, an online zero-forcing dirty paper precoder is proposed to update the precoding matrices by combining statistical learning with the Lyapunov learning technique. A tradeoff relation is theoretically established to show that the long-term weighted sum rate approaches the${\mathcal{ O}}(V)$-neighbor of optimal value while the long-term on-grid energy increases at a rate of${\mathcal{ O}}({\scriptstyle {}^{\scriptstyle \log ^{2}(V)}}\hspace {-0.224em}/\hspace {-0.112em}{\scriptstyle \sqrt {V}})$, where$V$is an introduced control parameter. Numerical results are used to verify the performance of the proposed online adaptive precoder.

To achieve the service-oriented features of the 5G, network slicing aims to create logical virtual networks where multiple services are provided on a common physical infrastructure. The performance of network slicing depends on the intelligent management of multi-dimensional resources, which are exactly what multi-access edge computing (MEC) provides. This paper proposes joint optimization of communication, computing and caching (3C) resources in multi-access edge network slicing. The optimization objective of the two-level resource allocation problem is to maximize the utility obtained by mobile virtual network operators while ensuring the quality of service (QoS). The deep reinforcement learning (DRL) approach is employed which enables the resource allocation scheme to intelligently adapt to the dynamic environment. Specifically, we propose a novel DRL approach named twin-actor deep deterministic policy gradient (twin-actor DDPG). Since the action space is continuous, the DDPG is adopted where the actor generates the deterministic policy while the critic evaluates the policy and guides the actor to obtain the optimal policy. A novel twin-actor structure is put forward to replace the actor of the DDPG, thus the slice-level action and user-level action can be generated respectively. The convergence and effectiveness of the proposed DRL based algorithm is are verified by numerical simulation.

Heterogeneous networks (HetNets) with end-to-end (E2E) network slicing are regarded as effective approaches to meet diverse service requirements from vertical industries. Due to the dense deployment of base stations (BSs) and the complicated associations between BSs and E2E network slices (NSs) in the scenario, the handoff problem faces challenges of the huge system state space and handoff action space and the considerable communication overhead. In this paper, we take these issues into account and consider a distributed E2E NS handoff decision framework in the HetNet. A decentralized Markov decision process (DEC-MDP)-based model is formulated for the distributed E2E NS handoff problem, and the jointly observable and random characteristics of the DEC-MDP are analyzed. To obtain a theoretical performance reference, the original distributed E2E NS handoff problem is simplified, and a Nash equilibrium-based performance bound is given. More practically, the multi-agent double deep Q-network-based distributed handoff (MA-DDQN-DH) algorithm with the centralized training and decentralized executing framework is proposed. Simulation results show that the Nash equilibrium-based performance bound is reasonable, and the proposed MA-DDQN-DH algorithm performs well in the comparison.

This paper investigates a joint resource allocation for ultra-reliable and low-latency radio access networks (URLLRANs) with edge computing. Compared with conventional networks, URLLRANs have more restrictive latency and reliability requirements, and always feature short packet communications. It is a challenging work to provide edge computing services in URLLRANs, since the processing and transmission delay as well as packet loss during computation and communications should all be taken into considerations. Along these lines, to specify the trade-off between latency and reliability, this paper defines computation rates and transmission rates for short packets. Different from the existing work, the proposal takes effective information as well as energy consumption as performance metrics based on the definition. The packet request rates, computation latency, service rates, communication power, blocklength, and transmission information amounts are jointly optimized to reduce energy consumption and meanwhile generate more effective information for both the computation system and the communication system. To solve the NP-hard problem, the locally optimal solution and global optimal solution are both derived. Simulation results validate the performance advantage of the proposal and also indicate that the locally optimal solution can greatly reduce the computation complexity with only a small performance loss when compared with the global optimal solution.

In microgrid, peer-to-peer (P2P) electricity trading has quickly ascended to the spotlight and gained enormous popularity. However, there are inevitable credit problems and system security problems. Besides, the current model in the electricity trading system cannot balance the utilities of multiple trading entities. In this paper, we propose a blockchain-based distributed P2P electricity trading system. We define elecoins as currency in circulation within our trading system. In order to jointly optimize the utilities of both parties in the elecoins trading, we formulate the elecoins purchasing problem as a hierarchical Stackelberg game. Then, we design a distributed multi-agent utility-balanced reinforcement learning (DMA-UBRL) algorithm to search the Nash equilibrium. Finally, we factually build a blockchain system with a blockchain explorer and deploy an electricity trading smart contract (ETSC) on Ethereum, with a website interface for operating. The numerical results and the implemented realistic system show the advantages of our work.

Nowadays, the rise of the Internet of Vehicles (IoV) has led to the rapid development of smart transportation. To increase the computing capacity of mobile vehicles and decrease the content delivery latency of suppliers, mobile edge computing (MEC) is considered as an indispensable solution. However, there are some essential issues to be considered: 1) security and privacy of data transmission, and 2) reasonable resource allocation for collaborative computing and caching. In this paper, to solve above issues, blockchain technology is adopted to ensure reliable transmission and interaction of data. Meanwhile, we develop an intelligent resource framework about computing and caching for blockchain-enabled MEC systems in IoV. Through jointly considering and optimizing offloading decision of computation task carried by vehicle, caching decision, the number of offloaded consensus nodes, block interval and block size, the energy consumption and computation overheads can be decreased, and the data throughput of the blockchain can be increased significantly. Moreover, the proposed optimization problem is modeled and formulated as a Markov decision process. Facing the complexity and dynamic of resource allocation, the asynchronous advantage actor-critic approach is considered and applied to solve the optimization problem. Experiment results demonstrate that the advantages of the proposed optimization scheme are obvious compared with other existing schemes.

As a compelling collaborative machine learning framework in the big data era, federated learning allows multiple participants to jointly train a model without revealing their private data. To further leverage the ubiquitous resources in end-edge-cloud systems, hierarchical federated learning (HFL) focuses on the layered feature to relieve the excessive communication overhead and the risk of data leakage. For end devices are often considered as self-interested and reluctant to join in model training, encouraging them to participate becomes an emerging and challenging issue, which deeply impacts training performance and has not been well considered yet. This paper proposes an incentive mechanism for HFL in end-edge-cloud systems, which motivates end devices to contribute data for model training. The hierarchical training process in end-edge-cloud systems is modeled as a multi-layer Stackelberg game where sub-games are interconnected through the utility functions. We derive the Nash equilibrium strategies and closed-form solutions to guide players. Due to fully grasping the inner interest relationship among players, the proposed mechanism could exchange the low costs for the high model performance. Simulations demonstrate the effectiveness of the proposed mechanism and reveal stakeholder's dependencies on the allocation of data resources.

Unmanned aerial vehicles (UAVs) have been used in millimeter-wave (mmWave) networks as relays to assist remote or blocked communication nodes. In this paper, we perform secrecy analysis for UAV-based mmWave relaying networks, where a cooperative jamming scheme is proposed via utilizing the destination and an external UAV to cooperatively disrupt the eavesdroppers at the two stages of relaying, respectively. Considering the probability of line-of-sight (LoS) between the UAV and ground nodes, the three-dimensional (3D) antenna gain, and the Nakagami-m small-scale fading model, closed-form SOP of the network is obtained by employing the Gauss-Chebyshev quadrature. Simulation results are presented to validate the theoretical expressions of SOP and to show the effectiveness of the proposed scheme.

Recently, electric vehicles (EVs) have been widely used under the call of green travel and environmental protection, and diverse requirements for charging are also increasing gradually. In order to ensure the authenticity and privacy of charging information interaction, blockchain technology is proposed and applied in charging station billing systems. However, there are some issues in blockchain itself, including lower computing efficiency of the nodes and higher energy consumption in the consensus process. To handle the above issues, in this paper, combining blockchain and mobile edge computing, we develop a reliable billing data transmission scheme to improve the computing capacity of nodes and reduce the energy consumption of the consensus process. By jointly optimizing the primary and replica nodes offloading decisions, block size and block interval, the transaction throughput of the blockchain system is maximized, as well as the consumption costs of latency and energy consumption is minimized. Moreover, we formulate the joint optimization problem as Markov decision process (MDP). To tackle this dynamic and continuity of the system state, the actor–critic reinforcement learning is introduced to solve the MDP problem. Finally, simulation results demonstrate that the performance improvement of the proposed scheme through comparison with other existing schemes.

Covert communication can hide the information transmission process from the warden to prevent adversarial eavesdropping. However, it becomes challenging when the warden can move. In this paper, we propose a covert communication scheme against a mobile warden, which maximizes the connectivity throughput between a multi-antenna transmitter and a full-duplex jamming receiver with the covert outage probability (COP) limit. First, we analyze the monotonicity of the COP to obtain the optimal location the warden can move. Then, under this worst situation, we optimize the transmission rate, the transmit power and the jamming power of covert communication to maximize the connection throughput. This problem is solved in two stages. Under this worst situation, we first maximize the connection probability over the transmit-to-jamming power ratio within the maximum allowed COP for a fixed transmission rate. Then, the Newton's method is applied to maximize the connection throughput via optimizing the transmission rate iteratively. Simulation results are presented to evaluate the effectiveness of the proposed scheme.

Nowadays, blockchain has become a promising tamper-evident and tamper-resistant distributed ledger technology that achieves the security and privacy through the cryptography, consensus mechanism and chained data structure. In this paper, we propose a general blockchain-based smart manufacturing system (BSM) that utilizes the decentralization, immutability, auditability of blockchain to achieve flexible manufacturing that responds to on-demand services in time. For management, manufacturing services are divided into tasks and for unified scheduling, these tasks are queued along the logical flow in the transaction pool. Considering the contradiction between large scale manufacturing and limited transactional throughput, a joint task scheduling over blockchains and supply-demand configuration design is proposed to obtain the maximum customers' net profit while balancing the timeliness as well as the production and blockchain payoff. Moreover, a maximum weight matching based Alternating Optimization framework (MWMAO) is proposed as the solution. Simulation results show that the proposed framework has superiority on the profitability.

In recent years, with the rapid development of the Internet of Things (IoT), various applications based on IoT have become more and more popular in industrial and living sectors. However, the hypertext transfer protocol (HTTP) as a popular application protocol used in various IoT applications faces a variety of security vulnerabilities. This article proposes a novel HTTP anomaly detection framework based on edge intelligence (EI) for IoT. In this framework, both clustering and classification methods are used to quickly and accurately detect anomalies in the HTTP traffic for IoT. Unlike the existing works relying on a centralized server to perform anomaly detection, with the recent advances in EI, the proposed framework distributes the entire detection process to different nodes. Moreover, a data processing method is proposed to divide the detection fields of HTTP data, which can eliminate redundant data and extract features from the fields of an HTTP header. Simulation results show that the proposed framework can significantly improve the speed and accuracy of HTTP anomaly detection, especially for unknown anomalies.

Mobile-edge computing (MEC) is considered as a promising technology to enable low latency applications while consuming less energy, and nonorthogonal multiple access (NOMA) is regarded as a hopeful method of increasing spectrum efficiency and the wireless network capacity. In this article, we consider a NOMA-MEC-based Internet-of-Things (IoT) network, and propose a joint optimization framework to maximize the effective system capacity, i.e., the number of IoT devices whose tasks are processed successfully, and meanwhile to maximize the total energy saving. First, we concentrate on improving the effective system capacity from the wireless side by introducing NOMA, and from the IoT device side by task offloading decision optimization, where distributed optimization is conducted and closed-form solution is obtained. Then, we maximize the total energy saving also from two aspects, i.e., the device-side computation resource allocation, and the wireless side joint admission control, user clustering, orthogonal subcarrier assignment, and transmit power control, where we resort to graph theory and propose a low-complexity heuristic algorithm to solve it. Abundant simulation results demonstrate our proposed joint optimization algorithm performs well in both effective system capacity optimization and energy saving maximization.

With the rapid development of automotive industry and telecommunication technologies, live streaming services in the Internet of Vehicles (IoV) play an even more crucial role in vehicular infotainment systems. However, it is a big challenge to provide a high quality, low latency, and low bitrate variance live streaming service for vehicles due to the dynamic properties of wireless resources and channels of IoV. To solve this challenge, we propose a novel live video transcoding and streaming scheme that maximizes the video bitrate and decreases time-delays and bitrate variations in vehicular fog-computing (VFC)-enabled IoV, by jointly optimizing vehicle scheduling, bitrate selection, and computational/spectrum resource allocation. This joint optimization problem is modeled as a Markov decision process (MDP), considering time-varying characteristics of the available resources and wireless channels of IoV. A soft actor–critic deep reinforcement learning (DRL) algorithm that is based on the maximum entropy framework, is subsequently utilized to solve the above MDP. Extensive simulation results based on the data set of the real world show that compared to other baseline algorithms, the proposed scheme can effectively improve video quality while decreasing latency and bitrate variations, and access excellent performance in terms of learning speed and stability.

Nowadays, many disruptive Internet-of-Things (IoT) applications emerge, such as augmented/virtual reality online games, autonomous driving, and smart everything, which are massive in number, data intensive, computation intensive, and delay sensitive. Due to the mismatch between the fifth generation (5G) and the requirements of such massive IoT-enabled applications, there is a need for technological advancements and evolutions for wireless communications and networking toward the sixth-generation (6G) networks. 6G is expected to deliver extended 5G capabilities at a very high level, such as Tbps data rate, sub-ms latency, cm-level localization, and so on, which will play a significant role in supporting massive IoT devices to operate seamlessly with highly diverse service requirements. Motivated by the aforementioned facts, in this article, we present a comprehensive survey on 6G-enabled massive IoT. First, we present the drivers and requirements by summarizing the emerging IoT-enabled applications and the corresponding requirements, along with the limitations of 5G. Second, visions of 6G are provided in terms of core technical requirements, use cases, and trends. Third, a new network architecture provided by 6G to enable massive IoT is introduced, i.e., space-air-ground-underwater/sea networks enhanced by edge computing. Fourth, some breakthrough technologies, such as machine learning and blockchain, in 6G are introduced, where the motivations, applications, and open issues of these technologies for massive IoT are summarized. Finally, a use case of fully autonomous driving is presented to show 6G supports massive IoT.

The past few years have witnessed the compelling applications of the Internet of Things (IoT) in our daily life. The explosive growth of the number of IoT devices also presents a great challenge in network security, especially the DDoS attack. Current DDoS defense mechanisms adopted out-of-band architecture, which is accomplished by a process that receives monitoring data from routers and switches, then analyzes that flow data to detect attacks. However, facing IoT devices growing rapidly, this out-of-band architecture confronted with limited processing capacity, bandwidth resources, and service assurance problems. Recently, with the development of the programming switch, it opens up new possibilities for in-network DDoS detection, where the detection algorithms could be directly implemented inside the routers and switches. Benefit from switch processing performance, the in-network mechanism could achieve high scalability and line speed performance. Therefore, in this article, we design a machine learning-based in-network DDoS detection framework. We implement the lightweight variational Bayes algorithm in each switch to detect the anomaly traffic. Besides, considering the shortage of training data in each switch, a centralized platform is introduced to synchronize parameters among distributed switches to realize collaborative learning. Extensive simulations are conducted to evaluate our proposed algorithm in comparison to some state-of-the-art schemes.

With the emergence of 5G, wireless sensor networks, and related technologies, Internet of Things (IoT) has gained prominence as an emerging paradigm to meet the demands of flexible, agile, and ubiquitous accessibility of cyberspace from physical systems. However, the current centralized IoT architecture is heavily restricted by the problems of single points of failure, data privacy, security, and robustness. Recently, blockchains have been found attractive as potential solutions to some of these problems, due to their ability to maintain immutable open ledgers that are accessible to everyone but are tamper-proof. In addition, rapid development of edge computing has enabled a large range of new IoT applications. Edge computing pushes cloud services from the network core to the network edges in closer proximity to IoT devices. Thus, blockchain and edge computing are attractive technologies to meet new and existing challenges by enabling new IoT applications and services through secure, reliable, flexible, and powerful devices and systems while motivating new business models in the growing digital economies. They can provide attractive solutions, such as schemes for decentralized services, service virtualization, rapid resource optimization, and flexible and reliable management and maintenance.

Wireless virtual reality (VR)-enabled medical treatment (WVMT) system, integrating the VR technology and the platform of the Internet of Medical Things (IoMT), is a promising application in future medical industries. Multiaccess edge computing (MEC) is an effective approach to support the ubiquitous applications of WVMT systems. Due to the high requirements of medical services, the computation efficiency and security are two issues in WVMT systems. In this article, we propose a blockchain-enabled task offloading scheme, where the viewport rendering tasks of VR devices (VDs) can be offloaded to edge access points (EAPs). The blockchain is integrated into the system to reach the consensus of the global information of task offloading and data processing to resist malicious attacks. To reduce VDs’ computation load under the promise of high VR QoE, we formulate the computation offloading and resource allocation to be a Markov decision problem, considering block consensus, content correlation, and fluctuating channel conditions. Then, a novel collective reinforcement learning (CRL) algorithm is proposed to adaptively allocate resources based on the requirements of viewport rendering, block consensus, and content transmission. In the simulations, the convergence rate and the performance in terms of energy consumption and stalling rate are evaluated. simulation results demonstrate the effectiveness of the proposed scheme.

The ever-growing number of Internet-of-Things (IoT) devices makes multiaccess edge computing (MEC)-assisted status update system more and more attractive, which can be deployed to enable remote data acquisition and analysis from urban space. The ambient computing resource at edge automatically extracts valuable status update information from the data collected by IoT devices, which supports the real-time remote monitoring applications. In this article, we employ the concept of Age of Information (AoI) to quantify the freshness of status updates. To combat the limited battery capacity at IoT devices, energy harvesting (EH) is leveraged to capture the green energy from ambient environment. Specifically, we investigate an age minimization problem by considering the randomness in energy arrivals, heterogeneity in harvesting mode, and the stochasticity in transmission and computing process. The formulated problem is a long-term stochastic optimization problem. Then, we transform the original problem into a series of per-time slot deterministic optimization problem. An online scheduling policy is proposed to obtain the energy management decisions at devices, and the transmission and computing scheduling decisions among multiple devices without any prior knowledge on the network dynamics, which is facilitated to be implemented. Simulation results show that the performance of our proposed algorithm is competitive when compared with other existing schemes.

Distributed deep learning deployed in an edge computing environment is a promising approach for extracting accurate information from raw sensor data from Internet of Things (IoT). But the distributed training suffers from heavy communication overheads between a master node and multiple compute nodes due to frequent transmission of gradients, which limits the training efficiency of the distributed deep learning. In this article, we propose a novel algorithm named ProbComp-LPAC (ProbComp: probability compression and LPAC: layer parameters adaptive compression), which can reduce the communication overhead and improve the training efficiency of the distributed deep learning. ProbComp-LPAC adopts a probability equation to select the gradients and uses different compression rates in different layers of deep neural networks. Comparing with other methods, such as adaptive compression (AdaComp) and lazily aggregated quantized compression (LAQ), the performance of ProbComp-LPAC is not only faster in the training speed but also higher in the accuracy of the test.

Recently, the term “Internet of Things” (IoT) has elicited escalating attention. The flexibility, agility, and ubiquitous accessibility have encouraged the integration between machine learning (ML) with IoT. However, there are many challenges that present the key inhibitors in moving ML to the public solution, such as centralized training, poor training efficiency, and heavy computing capabilities requirements. Therefore, bringing learning intelligence to edge IoT nodes has been spotlighted for some researches. Meanwhile, how to govern the use of learning results efficiently, reliably, scalably, and safely is hampered by the heterogeneity and nonconfidence among IoT nodes. In this article, we propose a blockchain-based collective Q-learning (CQL) approach to address the above issues, where lightweight IoT nodes are used to train parts of learning layers, then employing blockchain to share learning results in a verifiable and permanent manner. We further improve the traditional Proof of Work (PoW). Instead of solving a meaningless puzzle, we regard the learning process in the IoT node as a piece of work. Accordingly, the winner is the IoT node with the minimum reduced percentage of the learning loss function, referred to as the Proof-of-Learning (PoL) consensus protocol. Specifically, in order to show how the CQL approach works, we use it to address a networking integrated cloud-edge-end resource allocation in IoT. The experimental results reveal the superior performance of the proposed scheme.

To cope with the unprecedented surge in demand for data computing for the applications, the promising concept of multiaccess edge computing (MEC) has been proposed to enable the network edges to provide closer data processing for mobile devices (MDs). Since enormous workloads need to be migrated, and MDs always remain resource-constrained, data offloading from devices to the MEC server will inevitably require more efficient transmission designs. The integration of nonorthogonal multiple access (NOMA) technique with MEC has been shown to provide applications with lower latency and higher energy efficiency. However, the existing designs of this type have mainly focused on the transmission technique, which is still insufficient. To further advance offloading performance, in this work, we propose an application-driven NOMA-enabled computation offloading scheme by exploring the characteristics of applications, where the common data of the application is offloaded through multidevice cooperation. Under the premise of successfully offloading the common data, we formulate the problem as the maximization of individual offloading throughput, where the time allocation and power control are jointly optimized. By using the successive convex approximation (SCA) method, the formulated problem can be iteratively solved. Simulation results demonstrate the convergence of our method and the effectiveness of the proposed scheme.

Logistics is a driver of countries' and firms' competitiveness and plays a vital role in economic growth. However, the current logistics industry still faces high costs and low efficiency. The development of smart logistics brings opportunities to solve these problems. As one of the important technologies of the modern information and communication technology (ICT), the Internet of Things (IoT) can create oceans of data and explore the complex relationships between the transactions represented by these data with the help of various mathematical analysis technologies. These features are helpful to promote the development of smart logistics. In this article, we provide a comprehensive survey on the literature involving IoT technologies applied to smart logistics. First, the related work and background knowledge of smart logistics are introduced. Then, we highlight the enabling technologies for IoT in smart logistics. Furthermore, we review how IoT technologies are applied in the realm of smart logistics from the perspectives of logistics transportation, warehousing, loading/unloading, carrying, distribution processing, distribution, and information processing. Finally, some challenges and future directions are discussed.

The evolution of the Internet of Things (IoT) makes an increased emphasis on extending their computing and storage capabilities by relying particularly on the cloud/edge computing (EC) for cyber-physical systems (CPSs). Especially, in software-defined CPS (SD-CPS), different software-defined networking (SDN) controllers share information and cooperate to make global decisions. To further enhance system security during the information sharing process, we introduce blockchain technology into SD-CPS. However, because many security-related decisions are sensitive to latency, it is vital to minimize the system latency in blockchain-empowered SD-CPS. In this article, a blockchain-empowered distributed SD-CPS framework is proposed to realize consensus and distributed resource management by offloading data in a hybrid network paradigm that combines cloud computing and EC. Moreover, to adaptively implement offloading and control strategies while guaranteeing data security, we design a resource management scheme for reducing system latency and provide the flexibility of cooperation. To foster intelligence, we formulate the joint communication, computation, and consensus problems as a Markov decision process and use deep reinforcement learning to balance resource allocation, reduce latency, and guarantee data security. Compared with other schemes, simulation results verify the effectiveness of the proposed scheme, which performs better on self-adaptation decision making and system delay reduction.

Driven by the growing popularity of mobile applications, such as the Internet of Things (IoT), fog computing has been envisioned as a promising approach to enhance the computation capability of mobile devices and reduce the energy consumption. In this article, we aim to investigate the dynamic computation offloading problem in the IoT fog system under the fast time-varying wireless channel conditions. Our work differs from the existing work, which is based on the assumption that the channel-state information can be perfectly obtained by the offloading agent (e.g., the IoT device). In reality, due to hardware limitation, short sensing time, and network connectivity issues in IoT fog systems, it is difficult for the IoT device to have the perfect knowledge of a dynamic channel environment. Therefore, in this article, we propose a partially observable offloading scheme to enable the IoT device to make the optimal offloading decision with imperfect channel-state information. The optimization problem is formulated as a partially observable Markov decision process (POMDP) formulation, with the objective of minimizing the IoT device's energy consumption while meeting its requirement on task processing delay. To find the optimal offloading solution, an offline algorithm based on the deep recurrent $Q$ -network (DRQN) is developed. Finally, extensive simulation experiments are performed to evaluate the effectiveness of the proposed offloading scheme.

Secure information transmission and energy efficiency (EE) optimization are very important for simultaneous wireless information and power transfer (SWIPT)-aided heterogeneous networks. However, most of the existing works consider perfect channel state information (CSI) and linear energy harvesting (EH) models, which are too ideal in practical systems. In this article, we focus on the EE-based robust optimization with imperfect CSI and nonlinear EH models in a SWIPT-aided two-tier heterogeneous macro-femto network with multiple eavesdroppers. In particular, we formulate a robust beamforming problem by jointly optimizing the beamforming vectors of the macro base station (BS) and femto BSs, the power splitting (PS) factors of energy receivers, and the artificial noise vectors of BSs, under multiple constraints including the quality of service requirement of each user, the minimum harvested energy, the maximum transmit power, and the PS factor. Although the formulated robust optimization problem is nonconvex, an EE-based iterative algorithm is developed to obtain the solutions. Simulation results demonstrate the proposed algorithm is superior to other algorithms in terms of EE and security.

Currently, developments in wireless sensing technologies have shown that wireless signals can be employed to transmit information between wireless communication devices and are also able to realize passive target wireless sensing. Wireless sensing has diverse Internet-of-Things applications in indoor human detection, such as in device-free localization, activity recognition and fall detection, respiration detection, gait recognition, user identification, and so forth. Deep learning (DL), with the latest breakthroughs in machine learning (ML) and artificial intelligence (AI), seems to be a feasible technique for device-free wireless sensing (DFWS) and human detection in a more intelligent and autonomous manner. Although DL has attracted wide spread attention in computer vision (CV), AI games, speech recognition, automated vehicles, and other fields, its application in wireless sensing systems (WSSs) is relatively new, and little attention has been paid to it. Motivated by these developments, this article clarifies the motivation and mechanism of the DL-aided WSSs for human detection. First, we survey the most advanced architecture of DL that may be powerful for WSSs. We also review conventional ML and DL approaches to human detection based on red green blue (RGB)/depth camera and radar: one reason is to introduce the successful experience in these areas to the field of wireless sensing and another reason is that the possibility of combining and fusing information from the heterogeneous types of sensors is expected to improve the overall performance of practical human detection systems. We provide a comprehensive survey of the state-of-the-art research on wireless sensing for human detection with a focus on WSSs. Furthermore, a general structure of the DL-based WSS is introduced in detail for hitherto unexplored applications and future wireless sensing scenarios. We also discuss some open research issues in wireless sensing for human detection, including data acquisition for DL model training, calibration of signals from commercial devices, multimodal sensing, simultaneous user identification and activity recognition, multiuser human detection, and generalization ability of DL models, to indicate future research directions.

The cache-enabled multicast content pushing, which multicasts the content items to multiple users and caches them until requested, is a promising technique to alleviate the heavy network load by enhancing the traffic offloading. This, in turn, has called for the optimization of content pushing strategy while considering both the transmission and caching resources, which jointly result in the complicated coupling among pushing decisions and lead to high computational complexity. Unlike most existing approaches which simplify the pushing problem via bypassing the complicated coupling, in this paper, we propose a multicast content pushing strategy to maximize the offloaded traffic with the cost on content caching based on structured deep learning. Specifically, we design the convolution stage to extract the spatio-temporal correlations of one content item between different pushing decisions, and construct the fully-connected stage to capture the spatial coupling among the decisions of pushing different content items to different user devices. Moreover, to address the absence of the ground truth on multicast content pushing, we relax the transmission constraint to derive a performance upper bound for guiding the training direction. This relaxed problem is solved based on dynamic programming in a bottom-up manner. Compared to the state-of-the-art baselines including both the traditional model-based and the general neural network-based strategies, the proposed pushing strategy achieves significant performance gain in both the random-generated dataset and the real LastFM dataset. In addition, it is also shown that the proposed strategy is robust to the uncertainty of user request information.

Computation over multiple access channel (CoMAC) scheme provides a promising solution to future large-scale wireless networks by utilizing the superposition property of the wireless channel to compute a class of functions with a summation structure (e.g., mean, norm, etc.). However, its implementation usually requires all nodes' channel state information (CSI) and its performance is limited by the channel condition of the worst node. In order to avoid massive CSI aggregation and improve the limited performance, we propose an automatic repeat request (ARQ)-aided CoMAC scheme in this paper. The transmitters and signaling procedures are designed to achieve the tradeoff between the achievable function rate and the transmission delay. The corresponding performance of the proposed ARQ-aided CoMAC scheme and the traditional ARQ-aided communication scheme are compared for both homogeneous networks and heterogeneous networks. By optimizing the ARQ level, we further maximize the achievable function rate of the proposed scheme. Asymptotic closed-form expressions are derived by resorting to the extreme value theory and point mass approximation. Monte Carlo simulations are given to illustrate and verify the performance of the proposed designs.

Energy-efficient designs play an important role in the Internet of Things (IoT) that monitors multiple phenomena, due to the limited power supply and complicated observation. In this paper, taking into account the power consumptions of observation, coding, and communication, we propose a joint sparse observation and coding scheme for energy-efficient monitoring of multiple phenomena using IoT. Through the analysis of outage performance, we find that the sparse observation and coding scheme can achieve the performance of the full observation scheme in which all nodes observe all phenomena with lower power consumption due to the dynamic and selective observation and coding. With the derived achievable rates and network power consumption, we study the trade-off between achievable rates and network power consumption that is determined by both the observation matrix and the coding matrix. For given rate constraints, we propose an optimization problem to minimize the network power consumption by jointly designing the observation and coding matrices. To solve this NP-hard problem efficiently, we propose a low-complexity algorithm with the convex-concave procedure. Moreover, to improve performance in high noise environment, we adopt collaboration among nodes to suppress observation noises and equalize bad observations by utilizing observation diversity. Finally, simulation results illustrate the superior performance of the proposed schemes.

For future wireless networks, enormous numbers of interconnections are required, creating a multi-hop topology and leading to a great challenge on data aggregation. Instead of collecting data individually, a more efficient technique, computation over multi-access channels (CoMAC), has emerged to compute functions by exploiting the signal-superposition property of wireless channels. However, it is still an open problem on the implementation of CoMAC in multi-hop wireless networks considering fading channel and resource allocation. In this paper, we propose multi-layer CoMAC (ML-CoMAC) by combining CoMAC and orthogonal communication to compute functions in the multi-hop network. Firstly, to make the multi-hop network more tractable, we reorganize it into a hierarchical network with multiple layers that consists of subgroups and groups. Then, in the hierarchical network, the implementation of ML-CoMAC is given by computing and communicating subgroup and group functions over layers, where CoMAC is applied to compute each subgroup function and orthogonal communication is adopted for each group to obtain the group function. The general computation rate is derived and the performance is further improved through time allocation and power control. The closed-form solutions to optimization problems are obtained, which suggests that orthogonal communication and existing CoMAC schemes are generalized.

In this paper, we investigate the application of the non-orthogonal multiple access (NOMA) technique into the unmanned aerial vehicle (UAV) aided relay networks. Specifically, we first incorporate the NOMA protocol with the decode-and-forward (DF) relay protocol to enhance the performance of the cell edge users in a macrocell network. Theoretical analysis indicates that the NOMA-DF-relay protocol outperforms the conventional orthogonal multiple access (OMA) based DF-relay protocol in terms of data rate. To fully exploit the advantages of the proposed protocol, we formulate a joint UAV height optimization, channel allocation, and power allocation problem with the objective to maximize the total data rate of the cell edge users under the coverage of the UAV. For solving the formulated problem effectively, we first analyze its property and employ the golden section method to propose a general framework to obtain the optimal height of the UAV. Then, we design a low-complexity iterative algorithm to solve the joint channel-and-power allocation problem based on the matching theory and the Lagrangian dual decomposition technique. Finally, simulation results demonstrate that the NOMA-DF-relay protocol is superior to the OMA-DF-relay protocol even when the system parameters are not optimized, and the proposed algorithms can further significantly improve the network performance in comparison with the other schemes.

Wireless virtual reality (VR) is increasingly used in industrial Internet of Things (IIoTs). However, ultra-high viewport rendering demands and excessive terminal energy consumption restrict the application of wireless VR. Pervasive edge computing emerges as a promising method for wireless VR. In this article, we propose an energy-aware resource management scheme for wireless-VR-supported IIoTs. To reduce the energy consumption of VR equipments (VEs) while ensuring a smooth immersive VR experience, we formulate the viewport rendering offloading, computing, and spectrum resource allocation to be a joint optimization problem, considering content correlation between VEs, fluctuating channel conditions, and VR quality of experience. By applying dual approximation, the original problem is transformed to be a Markov decision process and an reinforcement learning (RL)-based online learning algorithm is designed to find the optimal policy. To improve the learning efficiency, the quantum parallelism is integrated into the RL to overcome “curse of dimensionality”. In the simulations, the convergence rate and the performance in terms of energy consumption and stalling rate are evaluated. Simulation results demonstrate the effectiveness of the proposed scheme.

Smart agriculture is able to optimize the information resources of agriculture, which can improve the quality and productivity of agricultural products. Wireless sensor networks (WSNs) provide smart agriculture with effective solutions for collecting, transmitting, and processing of information. However, the large number of sensor networks consume too much energy that violates the principle of green communication. Simultaneous wireless information and power transfer (SWIPT) technology utilizes radio-frequency signals to transmit information and provide energy to WSNs, which can extend the lifetime of WSNs effectively. In this article, an architecture design of smart agriculture is first proposed by exploiting the SWIPT. Then, an energy efficiency optimization scheme is studied to achieve green communication, in which the subcarriers' pairing and power allocation are jointly optimized. The process of communication is divided into two phases. Specifically, in the first phase, source sensor sends information to relay sensor and destination sensor. Relay sensor utilizes a part of the subcarriers to receive the information, and utilizes the remaining subcarriers to collect energy. Destination sensor uses all the subcarriers to receive the information. In the second phase, relay sensor utilizes the energy collected in the first phase to forward the information to destination sensor. An effective iterative optimization algorithm is proposed to resolve the proposed optimization problem through Lagrangian dual function. Simulation results validate that the performance of the algorithm can improve energy efficiency of the system effectively.

The positioning accuracy of the existing vehicular Global Positioning System (GPS) is far from sufficient to support autonomous driving and ITS applications. To remedy that, leading methods such as ranging and cooperation have improved the positioning accuracy to varying degrees, but they are still full of challenges in practical applications. Especially for cooperative positioning, in addition to the performance of methods, cooperators may provide false data due to attacks or selfishness, which can seriously affect the positioning accuracy. By fully exploiting the characteristics of blockchain and edge computing, this paper proposes a vehicular blockchain-based secure and efficient GPS positioning error evolution sharing framework, which improves vehicle positioning accuracy from ensuring security and credibility of cooperators and data. First, by analyzing the GPS error, a bridge can be established between the sensor-rich vehicles and the common vehicles to achieve cooperation by sharing the positioning error evolution at a specific time and location. Particularly, the positioning error evolution is obtained by a deep neural network (DNN)-based prediction algorithm running on the edge server. We further propose to use blockchain technology for storage and sharing the evolution of positioning errors, mainly to guarantee the security of cooperative vehicles and mobile edge computing nodes (MECNs). In addition, the corresponding smart contracts are designed to automate and efficiently perform storage and sharing tasks as well as solve inconsistencies in time scales. Extensive simulations based on actual data indicate the accuracy and security of our proposal in terms of positioning error correction and data sharing.

While communication-based train control (CBTC) systems play a crucial role in the efficient and reliable operation of urban rail transits, its high penetration level of communication networks opens doors to Man-in-the-Middle (MitM) attacks. Current researches regarding MitM attacks do not consider the characteristics of CBTC systems. Particularly, the limited computing capability of the on-board computers prevents the direct implementation of most existing intrusion detection and defense algorithms against the MitM attack. In order to tackle this dilemma, in this article, we first introduce edge intelligence (EI) into CBTC systems to enhance the computing capability of the system. A cross-layer defense scheme, which includes the detection and defense stages, are proposed next. For the cross-layer detection stage, we propose a Long Short-Term Memory (LSTM) and Support Vector Machine (SVM) based detection method to combine the detection probability calculated from the train control parameter sequence and operation log files. For the cross-layer defense stage, we construct a Bayesian game based defense model to derive the optimal defense policy against MitM attacks. To further improve the accuracy of the defense scheme as well as optimize the communication resource allocation scheme, we propose an optimal communication resource allocation scheme based on the Asynchronous Advantage Actor-Critic (A3C) algorithm at last. Extensive simulation results show that the proposed scheme achieves excellent performance in defending against MitM attacks.

Communication-based Train Control (CBTC) systems are the burgeoning directions for developing future train control systems. With the adoption of wireless communication and network techniques, train control systems are more vulnerable to cyber-attacks. Notably, the jamming attacks, aiming at the handoff process that is the weakest part of train ground communication systems, will cause long disruption of communication. It will have a severe impact on train control operation efficiency. Current research regarding industry control system security is hard to model the impact of the jamming attacks on the train control system quantitatively, and current countermeasure schemes against jamming attacks are not designed for the operating mechanism of train control systems. This paper first builds the train control security state transition probability model under jamming attacks. A cross-layer defense scheme is then proposed from the aspect of the physical layer, the cyber layer and the management layer. In the physical layer, this paper designs a model prediction control algorithm to track dynamic target signals, in the hopes of eventually tracking the dynamic target quickly and smoothly. In the cyber layer, a multi-stage and zero-sum stochastic game model is built for the channel selection for the attack and the defense, whereby the channel selection randomized policy will be obtained. In the management layer, a dynamic train travel speed profile generation algorithm is proposed to mitigate the jamming attacks’ impact on train control systems. Extensive simulation results are shown that jamming attack impact on CBTC can be mitigated effectively with our proposed cross-layer defense scheme.

Covert communication can hide the information transmission process from the warden to prevent adversarial eavesdropping. However, it becomes challenging when the location of warden is uncertain. In this paper, we propose a covert communication scheme against a warden with uncertain locations, which maximizes the connectivity throughput between a multi-antenna transmitter and a full-duplex jamming receiver with the limit of covert outage probability (the probability of the transmission found by the warden). First, we analyze the monotonicity of the covert outage probability to obtain the optimal location for the warden. Then, under this worst situation, we optimize the transmission rate, the transmit power and the jamming power of covert communication to maximize the connection throughput. This problem is solved in two stages. First, we derive the transmit-to-jamming power ratio limit from the maximum allowed covert outage probability. With this constraint, the connection probability is maximized over the transmit-to-jamming power ratio for a fixed transmission rate. Since the connection probability and the transmission rate are coupled, the bisection method is applied to maximize the connectivity throughput via optimizing the transmission rate iteratively. Simulation results are presented to evaluate the effectiveness of the proposed scheme.

Content popularity prediction plays an important role on proactive content caching. Different to most of the existing works which focus on improving the popularity prediction accuracy, in this article, we consider the content caching oriented popularity prediction through a weighted clustering approach in order to improve the caching performance. We formulate the loss of the cache hit ratio as the system regret to indicate the caching performance, and construct a clustering-based popularity prediction framework for overcoming the user request sparsity with considering the similarity of popularity evolution trends. For depicting the explicit relationship between the caching performance and the popularity prediction accuracy, we derive the popularity prediction error distribution of each content, and design the caching threshold. By extracting the insights in the relationship between the popularity prediction accuracy and the user clustering strategy, we develop a weighted clustering-based popularity prediction algorithm, which takes the caching regret probability of files as the weights. Based on two real-world datasets, the simulation results demonstrate that the proposed popularity prediction scheme achieves better caching performance than the state-of-the-art schemes.

Employing unmanned aerial vehicles (UAVs) in millimeter-wave (mmWave) networks as relays has emerged as an appealing solution to assist remote or blocked communication nodes. In this case, the network security becomes a great challenge due to the presence of malicious eavesdroppers. In this paper, we perform a secrecy analysis for a UAV-based mmWave relaying network. We first investigate the relaying scheme without jamming where the UAV decodes and forwards the information from the source to the destination with malicious eavesdropping. Furthermore, to enhance the secrecy performance, we propose a cooperative jamming scheme via utilizing the destination and an external UAV to cooperatively disrupt the eavesdroppers at the two stages of relaying, respectively. Using the probability of line-of-sight (LoS) between the UAV and ground nodes, the three-dimensional (3D) antenna gain, and the Nakagami-m small-scale fading model, the secrecy outage probability (SOP) of the two schemes with and without jamming is analyzed. Closed-form expressions for the SOP of the two schemes are obtained by employing the Gauss-Chebyshev quadrature. Simulation results are presented to validate the theoretical expressions of SOP and to show the effectiveness of the proposed schemes.

The revolutions of computing and communication have opened up demands for the high quality of service (QoS), such as high data transmission, high reliability, and low latency. These new opportunities have spawned numerous studies on edge computing and artificial intelligence (AI), even the cooperation between them, referred to as edge intelligence. However, there are a number of handicaps that prevent edge intelligence from being used as a generic platform. The most intractable one is the heterogeneity and un-credibility among edges, hindering the way of sharing the learning results reliably, flexibly, and efficiently. In this paper, we propose a blockchain-assisted edge intelligence (B-EI) approach to solve the problem. The edge learning nodes train their local intelligence, followed by the improved blockchain to share the local intelligence, constructing edge intelligence among the heterogeneous and uncredible edges. Specifically, the improved blockchain employs a novel learning-measured consensus protocol, named Proof of Learning. The edges, also acted as the blockchain nodes, compete to have more superior local intelligence, instead of solving a hashed result. The superior local intelligence is then shared and distributed with other edges. It is not only beneficial to achieve edge intelligence, but also efficient to employ the computation resource, by replacing the hashing as the intelligence training. In order to show the potential benefits, we then use the proposed B-EI approach to solve a joint resource assignment problem. Simulation results show that our scheme outperforms the other state-of-art solutions, in terms of training episodes, and resource utility.

Network slicing and Multi-access Edge Computing (MEC) have been envisioned as promising technique in the fifth generation mobile communication (5G). In this work, we study joint optimization of radio and computation resource in network slicing with MEC to maximize utility of Mobile Virtual Network Operator (MVNO), while meeting slice Quality of Service (QoS) requirements. On account of the dynamic change of slice demands and environment information, it is hard to solve resource allocation problems with conventional methods. Inspired by the superiority of deep reinforcement learning (DRL) in decision-making problems with the high state space and continuous action space. We formulate the utility maximization problem as a markov decision process (MDP). With an MVNO controller, the problem can be solved utilizing deep deterministic policy gradient (DDPG) algorithm to execute the dynamic resource allocation scheme. Simulation results show that utility performance of the proposed algorithm outperforms than the benchmark algorithms and enables dynamic resource allocation scheme.

On-path caching is a building block in Named Data Networking that helps eliminate redundant traffic. The performance of redundancy elimination depends on both Content Store (CS) and Pending Interest Table (PIT), i.e., CS caches content for future reuse, and PIT aggregates repetitive requests in a short period. However, contemporary proactive caching strategies only take account of CS while neglecting PIT. In this work, we integrate both PIT and CS into the proactive caching model, derive how to calculate aggregated request rate, and propose an algorithm to calculate the aggregated request rate across the tree topology. Then we formulate caching placement into optimization problems and solve them with a decomposition-based evolutionary algorithm. The simulation results show that the proposed scheme outperforms conventional solutions.

Unmanned aerial vehicles (UAVs) are anticipated to be integrated into the next generation wireless networks as new aerial mobile users, which can provide various live streaming applications such as surveillance, reconnaissance, etc. For such applications, due to the dynamic characteristics of traffic and wireless channels, how to guarantee the quality of service (QoS) is a challenging task. In this paper, with recent advances in scalable video coding (SVC), we study secure video streaming in wireless networks with UAVs. By jointly optimizing video levels selection and power allocation, the research tries to maximize the energy efficiency, which is the ratio of video quality to power consumption, while satisfying the secrecy timeout probability (STP) requirement. The aforementioned problem is modeled as a constrained Markov decision process (CMDP). And then, the study employs a state-of-the-art reinforcement learning algorithm, namely safe deep Q-learning network (safe-DQN), to solve the CMDP problem, in which a safety policies set is induced by constructing a Lyapunov function. Extensive simulation results with different system parameters show the effectiveness of the proposed algorithm compared with other existing reinforcement learning algorithms.

Recently, software-defined Industrial Internet of Things (SDIIoT), the integration of software-defined networking (SDN) and Industrial Internet of Things (IIoT), has emerged. It is perceived as an effective way to manage IIoT dynamically. Aiming to improve scalability and flexibility of SDIIoT, multi-SDN has been applied to form a physically distributed control plane to handle the large amount of data generated by industrial devices. However, as the core of multi-SDN, reaching consensus among multiple SDN controllers is a thorny issue. To meet the required design principle, this paper proposes a blockchain-enabled distributed architecture with SDIIoT to synchronize local views between distinct SDN controllers and finally reach the consensus of global view. On the other hand, both the cryptographic operations of blockchain and the noncryptographic computational tasks have access to the same computational resource pool of mobile edge cloud (MEC). In order to simultaneously optimize the throughput of blockchain and the energy consumption caused by computing, we adaptively allocate computational resources and the block size by jointly considering the trust features of SDN controllers and the resource requirements of non-cryptographic operations. To implement the truly distributed manner of blockchain, we describe our problem as a partially observable Markov decision process (POMDP) and propose a novel deep recurrent Q-network (DRQN) approach to solve it. In the simulation results, we compare two different protocols of blockchain and show the effectiveness of our scheme in either of them.

Federated learning is an emerging concept that trains the machine learning models with the distributed datasets, without sending the raw data to the data center. But in an edge computing enviroment where the wireless network resource is constrained, the key problem of federated learning is the communication overhead for parameters synchronization, which wastes bandwidth, increases training time, and even impacts the model accuracy. Gradient sparsification has received increasing attention, which only updates significant gradients and accumulates insignificant gradients locally. However, how to preserve the accuracy after a high ratio sparsification has always been ignored. In this paper, a General Gradient Sparsification (GGS) framework is proposed for adaptive optimizers, to correct the sparse gradient update process. It consists of two important mechanisms: gradient correction and batch normalization update with local gradients (BN-LG). With gradient correction, the optimizer can properly treat the accumulated insignificant gradients, which makes the model converge better. Furthermore, updating the batch normalization layer with local gradients can relieve the impact of delayed gradients without increasing the communication overhead. We have conducted experiments on LeNet-5, CifarNet, DenseNet-121, and AlexNet with adaptive optimizers. Results show that when 99.9% gradients are sparsified, validation datasets are maintained with top-l accuracy.

Non-orthogonal multiple access (NOMA) aided unmanned aerial vehicle (UAV) is becoming a promising technique for future wireless networks. However, its security remains a great challenge due to the line-of-sight in UAV communications and high transmit power for weak users in NOMA. Thus, in this paper, we propose a power allocation (PA) scheme for NOMA-UAV networks with circular trajectory, to maximize the sum rate of common users while guaranteeing the security for a specific user. To achieve this, we consider three cases based on the distance from the UAV to the secure user. Specifically, the lowest transmit power is assigned to the secure user in each time slot to guarantee its security, with the remaining power allocated to common users to maximize their sum rate. To further improve the transmission rate of the secure user, we also derive the upper bound for its decoding threshold, and analyze the linear relationship between the secure decoding threshold and the sum rate of common users. Simulation results are demonstrated to evaluate the effectiveness of the proposed secure PA scheme in NOMA-UAV networks.

To utilize the potential commutating resources of parked vehicles (PVs) in the large parking lot, we design a large-scale parked vehicular computing system via software defined blockchain. However, the parking time for PVs is uncertain and some computational services have delay requirements. Therefore, in this paper, we propose a delay-sensitive joint blockchain parameters and resource optimization framework including block size and block generation time, as well as the offloading strategy and computing frequency adjustment. Such a design causes the problem to be highly coupled and non-convex, for which we use an alternating optimization (AO) strategy and perform multiple transformations to ensure convexity. Finally, the simulation results show the effectiveness of the proposed scheme.

To prolong the time duration of smart mobile devices (SMDs) or enable low‐latency tasks, mobile edge computing (MEC) has emerged as a promising paradigm by offloading tasks to nearby MEC servers (MECSs). In this study the authors propose an optimisation problem to minimise the weighted sum of the total delay and energy consumption of all SMDs in a multi‐MECS‐multi‐SMD network via multi‐dimensional optimisation on offloading strategy making, load balancing, computation resource allocation and transmit power control. Since the problem is NP‐hard, the authors decompose it into three subproblems to solve. First, they propose a low complexity heuristic algorithm to obtain the offloading strategies while guaranteeing load balancing between the multiple MECSs. Then they solve computation resource allocation subproblem using Lagrange dual decomposition. Finally, employing fractional programming, the authors transform the transmit power control subproblem into a convex programming problem where the closed‐form solution is obtained. The proposed simulation results verify the convergence of the proposed iterative algorithms, and demonstrate that the proposed joint optimisation could achieve good performance in both delay and energy reduction.

Recently, the development of the Internet of Things (IoT) provides plenty of opportunities and challenges in various fields. As an essential part of IoT, machine-to-machine (M2M) communications open a novel way that the machine-type communication devices (MTCDs) are connected and communicated without any human intervention. Meanwhile, delay-tolerant data play an important role in M2M communications-based IoT, and it puts more emphasis on powerful data caching, computing, and processing, as well as the security and stability of data transmission. To meet these requirements in M2M communications networks, in this article, we introduce some promising technologies, such as edge computing and blockchain, and propose a joint optimization framework about caching, computation, and security for delay-tolerant data in M2M communications networks based on dueling deep Q-network (DQN). According to the dynamic decision process by DQN, the optimal selection and decision of caching servers, computing servers, and blockchain systems can be made to achieve maximum system rewards, which includes higher efficiency of data processing, lower network costs, and better security of data interaction. Extensive simulation results with different system parameters show that our proposed framework can effectively improve the system performance for blockchain-enabled M2M communications compared to the existing schemes.

Immersive virtual reality (VR) video is becoming increasingly popular owing to its enhanced immersive experience. To enjoy ultrahigh resolution immersive VR video with wireless user equipments, such as head-mounted displays (HMDs), ultralow-latency viewport rendering, and data transmission are the core prerequisites, which could not be achieved without a huge bandwidth and superior processing capabilities. Besides, potentially very high energy consumption at the HMD may impede the rapid development of wireless panoramic VR video. Multiaccess edge computing (MEC) has emerged as a promising technology to reduce both the task processing latency and the energy consumption for HMD, while bandwidth-rich terahertz (THz) communication is expected to enable ultrahigh-speed wireless data transmission. In this article, we propose to minimize the long-term energy consumption of a THz wireless access-based MEC system for high quality immersive VR video services support by jointly optimizing the viewport rendering offloading and downlink transmit power control. Considering the time-varying nature of wireless channel conditions, we propose a deep reinforcement learning-based approach to learn the optimal viewport rendering offloading and transmit power control policies and an asynchronous advantage actor-critic (A3C)-based joint optimization algorithm is proposed. The simulation results demonstrate that the proposed algorithm converges fast under different learning rates, and outperforms existing algorithms in terms of minimized energy consumption and maximized reward.

Mobile-edge computing (MEC) is a promising paradigm to improve the quality of computation experience of mobile devices because it allows mobile devices to offload computing tasks to MEC servers, benefiting from the powerful computing resources of MEC servers. However, the existing computation-offloading works have also some open issues: 1) security and privacy issues; 2) cooperative computation offloading; and 3) dynamic optimization. To address the security and privacy issues, we employ the blockchain technology that ensures the reliability and irreversibility of data in MEC systems. Meanwhile, we jointly design and optimize the performance of blockchain and MEC. In this article, we develop a cooperative computation offloading and resource allocation framework for blockchain-enabled MEC systems. In the framework, we design a multiobjective function to maximize the computation rate of MEC systems and the transaction throughput of blockchain systems by jointly optimizing offloading decision, power allocation, block size, and block interval. Due to the dynamic characteristics of the wireless fading channel and the processing queues at MEC servers, the joint optimization is formulated as a Markov decision process (MDP). To tackle the dynamics and complexity of the blockchain-enabled MEC system, we develop an asynchronous advantage actor–critic-based cooperation computation offloading and resource allocation algorithm to solve the MDP problem. In the algorithm, deep neural networks are optimized by utilizing asynchronous gradient descent and eliminating the correlation of data. The simulation results show that the proposed algorithm converges fast and achieves significant performance improvements over existing schemes in terms of total reward.

Inappropriate lane following and changing behaviors of connected and autonomous vehicles (CAVs) can result in accidents, such as rear-end collision and side collision. To remedy that, the use of deep reinforcement learning (DRL) for autonomous driving decisions is currently a widely used promising solution. In this case, the accuracy and effectiveness of such a machine learning (ML) model is quite essential for this artificial intelligence (AI)-enabled CAVs. This article proposes a blockchain-based collective learning (BCL) framework for autonomous lane-changing systems. Four key issues, namely, learning efficiency, data security, users' privacy, as well as communication burden, are addressed by applying collective learning, vehicular blockchain, and knowledge transfer. First, we model the lane-changing problem as a DRL process and learn the autonomous lane-changing strategy through the deep deterministic policy gradient (DDPG) algorithm. Second, a single CAV involves a limited number of driving scenarios, and the independent learning method has the problem of inefficiency. Therefore, we propose a collective learning framework to utilize the “collective intelligence” shared by CAVs. Third, a vehicular blockchain is then applied to ensure the security and privacy of the user and data. In addition, the introduction of the blockchain can incentivize more users to participate in collective learning. Finally, in order to accelerate the learning process and achieve higher level performance while further reducing the communication burden, we use the corresponding knowledge extracted from the ML model such as human learning, as privileged information for sharing instead of directly sharing local ML models. Extensive simulation results validate the effectiveness and efficiency of our proposal in terms of learning efficiency, driving safety, as well as system security and robustness.

Due to high mobility and high dynamic environments, object detection for vehicular networks is one of the most challenging tasks. However, the development of integration techniques, such as software-defined networking (SDN) and network function visualization (NFV), in networking, caching, and computing provides us with new approaches. In this article, we propose a novel context-aware object detection method based on edge-cloud cooperation. Specifically, an object detection model based on deep learning is established in the cloud server. Different from other methods, to further explore the underlying inner spatial features of collected images, the visual objects of images are regarded as nodes and the spatial relations between objects as edges, then a type of message-passing method is employed to update the nodes' features. In the mobile edge computing (MEC) servers, the context information and captured images of the vehicular environments are extracted and then are used to adjust the object detection model from the cloud server. In this way, the cloud server cooperates with the MEC servers to realize context-aware object detection, which improves the adaptation and performance of the detection model under different scenarios. The simulation results also demonstrate that the proposed method is more accurate and faster than the previous methods.

Object detection for autonomous driving is a huge challenge in the cross-domain adaptation scenario, especially for the time- and resource-consuming task. Distributed deep learning (DDL) has demonstrated a considerably good balance between efficiency and computation complexity. However, the reliability of DDL is low. Moreover, the cost of training data and model is not priced well. In this article, a novel blockchain-enabled model sharing approach is proposed to improve the performance of object detection with cross-domain adaptation for autonomous driving systems. Based on the blockchain and mobile-edge computing (MEC) technology, a domain-adaptive you-only-look-once (YOLOv2) model is trained across nodes, which can reduce significantly the domain discrepancy for different object categories. Furthermore, smart contracts are developed to perform data storage and model sharing tasks efficiently. The reliability of model sharing is ensured with blockchain consensus. We evaluate the proposed method under public data sets. The simulation results demonstrate that the efficiency and reliability of the proposed approach are better than the reference model.

Recently, software-defined Industrial Internet of Things (SDIIoT), the integration of software-defined networking (SDN) and Industrial Internet of Things (IIoT), has emerged. It is perceived as an effective way to manage IIoT dynamically. Aiming to improve the scalability and flexibility of SDIIoT, multi-SDN has been applied to form a physically distributed control plane to handle a large amount of data generated by industrial devices. However, as the core of multi-SDN, reaching consensus among multiple SDN controllers is a thorny issue. To meet the required design principle, this article proposes a blockchain-enabled distributed SDIIoT to synchronize local views between distinct SDN controllers and finally reach the consensus of the global view. On the other hand, both the cryptographic operations of blockchain and the noncryptographic tasks have access to the same computational resource pool of mobile edge cloud (MEC). In order to optimize the system energy efficiency, we adaptively allocate computational resources and the batch size of the block by jointly considering the trust features of SDN controllers and the resource requirements of noncryptographic operations. To implement the truly distributed manner of blockchain, we describe our problem as a partially observable Markov decision process (POMDP) and propose a novel deep reinforcement learning (DRL) approach to solve it. In the simulation results, we compare three different protocols of blockchain and show the effectiveness of our scheme in each of them.

Although vehicular global positioning system (GPS) has been widely applied in many traffic scenarios, it is far from achieving lane-level positioning due to its low accuracy. Existing cooperative positioning (CP) methods have improved vehicular positioning accuracy to varying degrees, which still have challenges in further improving the system's robustness and security. In this article, we propose a novel framework of blockchain-enabled Internet of Vehicles (IoV) with CP for improving vehicular GPS positioning accuracy, system robustness, and security. First, a self-positioning correction scheme for the intelligent vehicles is proposed to improve their positioning accuracy, which uses the multitraffic signs as benchmarks to correct the vehicular position (given by GPS) by deep neural network (DNN) algorithm. We further design a multi-intelligent vehicle positioning error sharing model to reduce GPS positioning error of common vehicles (CoVs) in the same segment or area. In addition, to realize information sharing between vehicles and ensure system security, the connections among intelligent vehicles, CoVs, and roadside units are built by proposing a blockchain-enabled architecture that includes IoV subsystem and blockchain subsystem, where the corresponding mechanism of the information choosing, information sharing, and penalty is designed. Extensive simulation results show the accuracy, robustness, and security of our proposal in terms of vehicular positioning, information transferring, and sharing.

Federated learning is an emerging concept that trains the machine learning models with the local distributed data sets, without sending the raw data to the data center. But, in the Internet of Things (IoT) where the wireless network resource is constrained, the key problem of federated learning is the communication overhead for parameter synchronization, which wastes bandwidth, increases training time, and even impacts the model accuracy. Gradient sparsification has received increasing attention, which only updates significant gradients and accumulates insignificant gradients locally. However, how to preserve the accuracy after a high ratio sparsification has been ignored in the literature. In this article, a general gradient sparsification (GGS) framework is proposed for adaptive optimizers, to correct the sparse gradient update process. It consists of two important mechanisms: 1) gradient correction and 2) batch normalization (BN) update with local gradients. With gradient correction, the optimizer can properly treat the accumulated insignificant gradients, which makes the model converge better. Furthermore, updating the BN layer with local gradients can relieve the impact of delayed gradients without increasing the communication overhead. We have conducted experiments on LeNet-5, CifarNet, DenseNet-121, and AlexNet with adaptive optimizers. Results show that when 99.9% gradients are sparsified, validation data sets are maintained with top-1 accuracy.

Recently, fog computing has emerged as a prospective technique to provide pervasive and agile computation services for Internet-of-Things (IoT) devices and support advanced applications. Introducing the energy harvesting (EH) technique into the fog computing system can extend the battery lifetime and provide a higher quality of experiences (QoE) for IoT devices. In the EH-enabled IoT fog system, computation offloading is an important issue and has attracted much attention. In most existing works, it is assumed that the IoT device is fully aware of the system state. However, in practical offloading problems, the IoT device may not be able to obtain accurate system state information, and only have a partial observation of the environment. Therefore, in this article, we investigate the decentralized partially observable offloading problem in the EH-enabled IoT fog system, in which multiple IoT devices cooperate to maximize the network performance while meeting their QoE requirements. We formulate the optimization problem as a decentralized partially observable Markov decision process (Dec-POMDP) in which each IoT device makes the task offloading decisions according to its local observation of the environment. The Lagrangian approach and the policy gradient method are adopted to find the optimal solution for the proposed problem. Due to the high complexity of solving the Dec-POMDP, a learning-based decentralized offloading algorithm with low complexity is presented to find the approximate optimal solution. Finally, extensive experimental evaluation and comparison are carried out to show the effectiveness of the proposed scheme.

The increasing demand of data computing and storage for cloud-based services motivates the development and deployment of large-scale data centers. This paper studies the resource allocation problem for the data center networking system when multiple data center operators (DCOs) simultaneously serve multiple service subscribers (SSs). We formulate a hierarchical game to analyze this system where the DCOs and the SSs are regarded as the leaders and followers, respectively. In the proposed game, each SS selects its serving DCO with preferred price and purchases the optimal amount of resources for the SS's computing requirements. Based on the responses of the SSs' and the other DCOs', the DCOs decide their resource prices so as to receive the highest profit. When the coordination among DCOs is weak, we consider all DCOs are noncooperative with each other, and propose a sub-gradient algorithm for the DCOs to approach a sub-optimal solution of the game. When all DCOs are sufficiently coordinated, we formulate a coalition game among all DCOs and apply Kalai-Smorodinsky bargaining as a resource division approach to achieve high utilities. Both solutions constitute the Stackelberg Equilibrium. The simulation results verify the performance improvement provided by our proposed approaches.

Federated learning is a communication-efficient training process that alternate between local training at the edge devices and averaging of the updated local model at the center server. Nevertheless, it is impractical to achieve perfect acquisition of the local models in wireless communication due to the noise, which also brings serious effect on federated learning. To tackle this challenge in this paper, we propose a robust design for federated learning to decline the effect of noise. Considering the noise in two aforementioned steps, we first formulate the training problem as a parallel optimization for each node under the expectation-based model and worst-case model. Due to the non-convexity of the problem, regularizer approximation method is proposed to make it tractable. Regarding the worst-case model, we utilize the sampling-based successive convex approximation algorithm to develop a feasible training scheme to tackle the unavailable maxima or minima noise condition and the non-convex issue of the objective function. Furthermore, the convergence rates of both new designs are analyzed from a theoretical point of view. Finally, the improvement of prediction accuracy and the reduction of loss function value are demonstrated via simulation for the proposed designs.

The next generation wireless network is expected to connect billions of nodes, which brings up the bottleneck on the communication speed for distributed data fusion. To overcome this challenge, computation over multiple access channel (CoMAC) was recently developed to compute the desired functions with a summation structure (e.g., mean, norm, etc.) by using the superposition property of wireless channels. This work aims to maximize the achievable function rate of reliable CoMAC in wireless networks. More specifically, considering channel fading and transceiver design, we derive the achievable function rate adopting the quantization and the nested lattice coding, which is determined by the number of nodes, the maximum value of messages and the quantization error threshold. Based on the derived result, the transceiver design is optimized to maximize the achievable function rate of the network. We first study a single cluster network without inter-cluster interference (ICI). Then, a multi-cluster network is further analyzed in which the clusters work in the same channel with ICI. In order to avoid the global channel state information (CSI) aggregation during the optimization, a low-complexity signaling procedure irrelevant with the number of nodes is proposed utilizing the channel reciprocity and the defined effective CSI.

The boom of video streaming industry has resulted in the increasing demands for transcoding services from heterogeneous users. Recent advances of blockchain technology allow some startups to realize decentralized collaborative transcoding through device-to-device (D2D) networks, where a group of transcoders are selected to perform transcoding cooperatively. For the blockchain-enabled D2D transcoding systems, it's imperative to jointly design transcoder selection, task scheduling and resource allocation schemes in order to provide efficient and trustworthy transcoding services. In this paper, viewing the involved multi-dimensional complex factors and channel fluctuation, we propose a novel deep reinforcement learning (DRL) based transcoder selection framework for blockchain enabled D2D transcoding systems where both the platform dynamics and channel statistics are captured. To reduce the action space size, we adopt a two-stage decision approach to first select the transcoders through a normal DRL based framework and then obtain the optimal task scheduling, power control, and resource allocation scheme by solving a stochastic optimization problem with the constrained stochastic successive convex approximation (CSSCA) approach. Simulation results show that our proposed framework can achieve high transcoding revenue while meeting the quality of service (QoS) requirements, and it can well handle dynamic cases.

Cooperative beamforming can provide significant performance improvement for relaying systems with the help of the channel state information (CSI). In time-division duplexing (TDD) mode, the estimated CSI will deteriorate due to the reciprocity mismatch. In this work, we examine the impact and the calibration of the reciprocity mismatch in relaying systems. To evaluate the impact of the reciprocity mismatch for all devices, the closed-form expression of the achievable rate is first derived. Then, we analyze the performance loss caused by the reciprocity mismatch at sources, relays, and destinations respectively to show that the mismatch at relays dominates the impact. To compensate the performance loss, a two-stage calibration scheme is proposed for relays. Specifically, relays perform the intra-calibration based on circuits independently. Further, the inter-calibration based on the discrete Fourier transform (DFT) codebook is operated to improve the calibration performance by cooperation transmission, which has never been considered in previous work. Finally, we derive the achievable rate after relays perform the proposed reciprocity calibration scheme and investigate the impact of estimation errors on the system performance. Simulation results are presented to verify the analytical results and to show the performance of the proposed calibration approach.

Blind estimation of parameters for M-ary frequency-shift-keying (M-FSK) signals is great of importance in intelligent receivers. Many existing algorithms have assumed white Gaussian noise. However, their performance severely degrades when grossly corrupted data, i.e., outliers, exist. This article solves this issue by developing a novel approach for parameter estimation of M-FSK signals in the presence of alpha-stable noise. Specifically, the proposed method exploits the generalized first- and second-order cyclostationarity of M-FSK signals with alpha-stable noise, which results in closed-form solutions for unknown parameters in both time and frequency domains. As a merit, it is computationally efficient and thus can be used for signal preprocessing, symbol timing estimation, signal and noise power estimation. Furthermore, substantial theoretical analysis on the performance of the proposed approach is provided. Simulations demonstrate that the proposed method is robust to alpha-stable noise and that it outperforms the state-of-the-art algorithms in many challenging scenarios.

INTERNET of Thing (IoT) technology has attracted intensive interest in the automotive industry to meet the new demands in the market while continuing to achieve their conservative goals [item 1) in the Appendix]. As for Industrial Internet of Things (IIoT), randomly moving wireless nodes are often carried by humans and communicate with each other when they are in close proximity. The interaction between nodes shows strong regularity or sociality, i.e., a wireless node always communicates with several social-closed or distance-closed nodes. This special section collects the latest ideas and research on the social and cognitive mobile computing in IIoT. Particularly, 15 original articles are accepted and included in the collection on the following pages. The topics of these articles are mainly concerned with social and cognitive mobility modeling, routing protocol, resource allocation, and so forth. We believe that these articles will play a role in inspiring our readers. Summaries of accepted articles are provided.

The four papers in this special section focus on the use of blockchain in the healthcare field. With the development of society, health has received increasing attentions. The development of science and technology has also promoted the protection of health. In recent years, the rapid development of computing and networking technologies has improved the ability to collect, measure, and analyze health-related data, and thus tremendous opportunities have opened up for healthcare computing. Meanwhile, these technologies have also brought new challenges and issues.

Realizing the ultra-low latency and high-accuracy solutions for rear-end collision is still challenging, especially under the condition in which many uncertainties exist. This paper proposes an artificial intelligence-based warning algorithm for rear-end collision avoidance. Three key issues are addressed by applying the neural network approach, including noises in positioning, inaccurate risk assessment, and enhanced comfort level of passengers. First, to filter the noises in positioning, wireless vehicular communications are leveraged; accurate relative lane positioning can be achieved to justify when two vehicles are in the same lane. Second, an online neural network model is developed to assess the risk of collisions in real time while driving. The algorithm can converge fast to a globally optimal solution and adapt to different traffic environments. Third, to maximize the comfort of passengers during the braking process, a graded warning strategy is developed at the prerequisite of guaranteed safety. With the above schemes sewed in to one framework, our proposal can achieve rear-end warning with reduced missing alarm rate, accurate risk assessment and enhanced comfort to passengers. The extensive simulations validate the effectiveness and accuracy of our proposal in terms of relative lane positioning, risk assessment, and collision avoidance.

Large-scale services are generally hosted on multiple application servers to scale out in today's data centers. Load balancers distribute users' requests across these servers. Software load balancer and switch-based load balancer are two typical classes of load balancers. However, most of the existing mechanisms either exhibit high processing latency at load balancers or likely lead to unbalanced requests distribution without considering the disparity of the application servers. In this paper, we study how the disparity of application servers significantly impacts the response time of requests. A fast switch-based Load Balancer considering Application Server states (LBAS) then is proposed to minimize the processing latency at both load balancers and application servers. The data plane of LBAS is well designed to store millions of connections in limited storage capacity without violating per-connection consistency. Besides, a partial dynamic weighting algorithm based on the Ridge Regression theory is designed and implemented to decrease the processing latency at application servers. We implement LBAS using the P4 programming language and conduct a series of extensive experiments to evaluate the performance. The results demonstrate that the proposed LBAS mechanism significantly reduces the response time of requests compared with Uniform random, Static weight, and Spotlight in various scenarios.

Recently, the emergence of blockchain has stirred great interests in the field of Internet of Things (IoT). However, numerous non-trivial problems in the current blockchain system prevent it from being used as a generic platform for large-scale services and applications in IoT. One notable drawback is the scalability problem. Lots of projects and researches have been done to solve this problem. Nevertheless, they do not consider different users' conditions, only using a single consensus protocol as the best fit one, as well as the IoT system is heavily constrained by computing and networking resources. In this article, we study a permissioned blockchain-based IoT architecture. In order to improve the scalability of the blockchain system and meet the needs of different users, we propose a service-oriented permissioned blockchain, where different consensus protocols are launched according to users' quality of service (QoS) requirements. Specially, we quantify a few popular consensus protocols. Additionally, we select block producers, which need a great number of computation resources, as well as dynamically allocate network bandwidth to the blockchain system. We formulate consensus protocols selection, block producers selection, and network bandwidth allocation as a joint optimization problem. We then use a dueling deep reinforcement learning approach to solve the problem. Simulation results demonstrate the effectiveness of our proposed scheme.

The application of blockchain to mobile edge computing (MEC) systems has attracted great interests. However, the design and optimization of blockchain and MEC in most existing works are done separately, which will result in sub-optimal performance. In this paper, we propose a joint optimization framework for blockchain-enabled MEC systems to achieve the optimal trade-off between the performance of the MEC system and the performance of the blockchain system. Specifically, both MEC and blockchain are considered as services in the framework, where energy consumption and delay/time to finality (DTF) are the performance metrics for the MEC system and the blockchain system, respectively. We formulate an optimization problem to achieve the optimal trade-off through jointly optimizing user association, data rate allocation, block producer scheduling, and computational resource allocation. To solve the problem, we decouple the optimization variables for efficient algorithm design. In addition, we develop an iterative algorithm for user association and data rate allocation and a bisection algorithm for computing resource allocation. Simulation results show the convergence of the proposed algorithms, and the proposed scheme can achieve the optimal trade-off between energy consumption and DTF.

The Internet of things (IoT) is becoming more and more flexible and economical with the advancement in information and communication technologies. However, IoT networks will be ultra-dense with the explosive growth of IoT devices. Network function virtualization (NFV) emerges to provide flexible network frameworks and efficient resource management for the performance of IoT networks. In NFV-enabled IoT infrastructure, service function chain (SFC) is an ordered combination of virtual network functions (VNFs) that are related to each other based on the logic of IoT applications. However, the embedding process of SFC to IoT networks is becoming a big challenge due to the dynamic nature of IoT networks and the abundance of IoT terminals. In this paper, we decompose the complex VNFs into smaller virtual network function components (VNFCs) to make more effective decisions since VNF nodes and IoT network devices are usually heterogeneous. In addition, a deep reinforcement learning (DRL) based scheme with experience replay and target network is proposed as a solution that can efficiently handle complex and dynamic SFC embedding scenarios in IoT. Our simulations consider different types of IoT network topologies. The simulation results present the efficiency of the proposed dynamic SFC embedding scheme.

In this paper, we present a blockchain-based mobile edge computing (B-MEC) framework for adaptive resource allocation and computation offloading in future wireless networks, where the blockchain works as an overlaid system to provide management and control functions. In this framework, how to reach a consensus between the nodes while simultaneously guaranteeing the performance of both MEC and blockchain systems is a major challenge. Meanwhile, resource allocation, block size, and the number of consecutive blocks produced by each producer are critical to the performance of B-MEC. Therefore, an adaptive resource allocation and block generation scheme is proposed. To improve the throughput of the overlaid blockchain system and the quality of services (QoS) of the users in the underlaid MEC system, spectrum allocation, size of the blocks, and number of producing blocks for each producer are formulated as a joint optimization problem, where the time-varying wireless links and computation capacity of the MEC servers are considered. Since this problem is intractable using traditional methods, we resort to the deep reinforcement learning approach. Simulation results show the effectiveness of the proposed approach by comparing with other baseline methods.

Both mobile edge cloud (MEC) and software-defined networking (SDN) are technologies for next generation mobile networks. In this paper, we propose to simultaneously optimize energy consumption and quality of experience (QoE) metrics in video streaming over software-defined mobile networks (SDMN) combined with MEC. Specifically, we propose a novel mechanism to jointly consider buffer dynamics, video quality adaption, edge caching, video transcoding and transmission. First, we assume that the time-varying channel is a discrete-time Markov chain (DTMC). Then, based on this assumption, we formulate two optimization problems which can be depicted as a constrained Markov decision process (CMDP) and a Markov decision process (MDP). Then, we transform the CMDP problem into regular MDP by deploying Lyapunov technique. We utilize asynchronous advantage actor-critic (A3C) algorithm, one of the model-free deep reinforcement learning (DRL) methods, to solve the corresponding MDP issues. Simulation results are presented to show that the proposed scheme can achieve the goal of energy saving and QoE enhancement with the corresponding constraints satisfied.

Massive numbers of nodes will be connected in future wireless networks. This brings great difficulty to collect a large amount of data. Instead of collecting the data individually, computation over multi-access channels (CoMAC) provides an intelligent solution by computing a desired function over the air based on the signal-superposition property of wireless channels. To improve the spectrum efficiency in conventional CoMAC, we propose the use of non-orthogonal multiple access (NOMA) for functions in CoMAC. The desired functions are decomposed into several sub-functions, and multiple sub-functions are selected to be superposed over each resource block (RB). The corresponding achievable rate is derived based on sub-function superposition, which prevents a vanishing computation rate for large numbers of nodes. We further study the limiting case when the number of nodes goes to infinity. An exact expression of the rate is derived that provides a lower bound on the computation rate. Compared with existing CoMAC, the NOMA-based CoMAC not only achieves a higher computation rate but also provides an improved non-vanishing rate. Furthermore, the diversity order of the computation rate is derived, which shows that the system performance is dominated by the node with the worst channel gain among these sub-functions in each RB.

Device-to-Device (D2D) caching assists Mobile Edge Computing (MEC) based caching in offloading inter-domain traffic by sharing cached items with nearby users, while its performance relies heavily on caching nodes' sharing willingness. In this paper, a Blockchain-based Cache and Delivery Market (CDM) is proposed as an incentive mechanism for the distributed caching system. Under given incentive mechanisms, both D2D and MEC caching nodes' willingness is guaranteed by satisfying their expected reward for cache sharing. Besides, for the distributed CDM, content delivery related transactions are executed by smart contracts. To achieve consensus on transactions and prevent frauds, a consensus protocol among the smart contract execution nodes (SCENE) is necessary. To minimize the latency of reaching consensus while guaranteeing its confidence level, we propose partial Practical Byzantine Fault Tolerance (pPBFT) protocol. Further, the model of cache sharing and transaction execution consensus is proposed, and we further formulate caching placement and SCENE selection as Markov Decision Process problems. Due to the complexity and dynamics of the problems, a deep reinforcement learning approach is adopted to solve the problem. The simulation results show that the proposed schemes outperform conventional solutions in terms of traffic offloading, content retrieval latency, and consensus latency.

Both mobile edge cloud (MEC) and software-defined networking (SDN) are technologies for next generation mobile networks. In this paper, we simultaneously optimize energy consumption and quality of experience (QoE) in video streaming over software-defined mobile networks (SDMN) with MEC. Specifically, we propose to jointly consider buffer dynamics, video quality adaption, edge caching, video transcoding and transmission. We formulate two optimization problems which can be depicted as a constrained Markov decision process (CMDP) and a Markov decision process (MDP). Then we transform the CMDP problem into regular MDP by deploying Lyapunov technique. We utilize asynchronous advantage actor-critic (A3C) algorithm, one of the deep reinforcement learning (DRL) methods, to solve the corresponding MDP problems. Simulation results are presented to show that the proposed scheme can achieve the goal of energy saving and QoE enhancement with the corresponding constraints satisfied.

In ultra-dense networks, the increasing demand for wireless services has led to severe energy consumption problem. In this paper, we mainly focus on green communication and computation offloading in ultra-dense networks, constituted of different macro base stations and small-cell base stations. This paper jointly considers edge energy consumption and delay under the limited network resource for multiple users.To address this issue, we propose an efficient computation offloading scheme in multi- user multi-task scenario, and a cuckoo search algorithm is invoked for solving the computation offloading problem. To be specific, the global convergence analysis presents the validity of this computation offloading scheme. Finally, experimental results validate that our proposed scheme is conducive to improving the efficiency of entire system in ultra-dense networks.

Due to the easy deployment, low cost and high maneuverability, unmanned aerial vehicles (UAVs) serving as aerial base stations can be efficiently deployed according to realtime situations for providing high-quality coverage, which can improve the communication efficiency and meet the requirements of green communications. However, due to the finite flight energy, a single UAV has limited capability of providing seamless long-term service to ground users. Therefore, the cooperation of multiple drones relying on sophisticated recharging and reshuffling schemes is necessary. In this paper, we investigate an energy- efficient cooperation strategy of multi-UAVs for providing seamless long-term coverage, where the positioning and the flight strategy are jointly considered. We first introduce a novel UAV power model, based on which we derive the cyclic UAV recharging and reshuffling constraint in order to satisfy the seamless long-term coverage requirement. For maximizing the energy-efficiency, we introduce a two-stage joint optimization algorithm for solving both the optimal UAV deployment as well as the cyclic UAV recharging and reshuffling strategy (CRRS). Finally, the efficiency of our proposed algorithm is shown by the simulation results.

Recent advances in Internet of Things (IoT) provide plenty of opportunities for various areas. Nevertheless, the machine-to-machine (M2M) communications-based IoT develops rapidly but suffers from extra energy consumption, large data transmission latency as well as overmuch network cost, because various of machine-type communication devices (MTCDs) are deployed in the network. To meet the requirements of energy efficient M2M communications, in this paper, we introduce a promising technology named as mobile edge computing (MEC), and propose a performance optimization framework with MEC for M2M communications network based on deep reinforcement learning (DRL). According to dynamic decision process by DRL, the appropriate access networks and the computing servers can be determined and selected with the minimum system cost, which includes lower network cost, time cost and energy consumption for data transmission and computing tasks execution. Extensive simulation results with different system parameters show that our proposed framework can effectively improve the system performance for M2M communications compared to the existing schemes.

Built-in caching in Named Data Networking (NDN) promises to provide efficient content delivery, where the dedicated on-path caching scheme is deployed to serve users' requests on the forwarding path. In this work, to utilize limited caching resources to achieve optimal performance, the caching placement decision is made by jointly considering the content popularity, underlying network topology, forwarding strategy and caching service mechanism in NDN. More specifically, we propose a service-aware caching model. In the model, we first define the Cache Service Matrix (CSM), which describes the position where each user's request is served for each piece of content. In order to make CSM comply with the caching placement, underlying topology, forwarding strategy, and on-path caching service mechanism, we propose an algorithm to calculate CSM under the preceding constraints. With CSM, the utility of caching placement could be derived correctly, and we formulate the optimal caching placement into optimization problems. Moreover, the differential grouping co-evolutionary (DG2-E) algorithm is adopted to decompose and solve the NP-hard optimization problems. Simulation results show the proposed scheme outperforms state of the art solutions in terms of inter-domain traffic reducing and request-response accelerating under arbitrary topologies.

Mobile edge computing enabled vehicular networking has appeared as a promising solution to the emerging resource hungry vehicular applications. In this paper, we study the computation offloading in a cognitive vehicular network that reuses the TV white space (TVWS) bands. We propose to maximize the average economical profit of the service provider by jointly considering communication and computation resource allocation, while guaranteeing network stability and the QoS of TVWS primary users. Based on Lyapunov optimization, we design an per-frame algorithm to tackle the joint optimization problem, where we first derive the closed-form solution for computation resource allocation, and then develop a continuous relaxation and Lagrangian dual decomposition based iterative algorithm for radio resource allocation. Simulation results demonstrate that the proposed algorithm can flexibly balance the profit-delay tradeoff, and can improve the economical profit of the service provider significantly as compared with the existing schemes.

The rapid development of Internet of Vehicles (IoV) necessitates a secure and reliable infrastructure to store and share the massive data. Blockchain, a distributed and immutable ledger, is widely considered as a promising solution to ensure data security and privacy for IoV. To deal with the massive IoV data, the scalability of blockchain becomes a critical issue, which should maximize transactional throughput as well as handling the dynamics of IoV scenarios. Therefore, this paper proposes a novel deep reinforcement learning (DRL) based performance optimization framework for blockchain-enabled IoV, where transactional throughput is maximized while guaranteeing the decentralization, latency and security of the underlying blockchain system. In this framework, we first carry out the performance analysis for blockchain systems from the aspects of scalability, decentralization, latency and security. Further, DRL technique is adopted to select block producers and adjust block size and block interval to adapt to the dynamics of IoV scenarios. Simulation results show that our proposed framework can effectively improve the throughput of blockchain-enabled IoV systems without affecting other properties.

Recently, power allocation has attracted more and more attention in order to optimize the performance of non-orthogonal multiple access (NOMA) systems. Different from existing works, the power allocation problems are investigated for cooperative NOMA systems with dedicated amplify-and-forward half-duplex relay (NOMA-HDR) and full-duplex relay (NOMA-FDR). From the fairness standpoint, the power allocation problems are formulated to maximize the minimum achievable user rate in the considered systems. The problems for both NOMA-HDR and NOMA-FDR systems with two-user and M-user are addressed. The closed-form power allocation policy of two-user NOMA-HDR system is obtained. Also, the optimal numerical power allocation policies for two-user NOMA-FDR and M-user NOMA-HDR systems are obtained. In addition, the problem for M-user NOMA-FDR systems is solved in noise-limited environment. Simulation results show that the proposed NOMA-HDR or NOMA-FDR scheme with power adaption clearly outperforms the NOMA-HDR or NOMA-FDR scheme with fixed power allocation. Besides, when the residual self-interference channel gain is small, the performance of NOMA-FDR system is better than the NOMA-HDR system.

In this paper, we focus on a downlink cellular network, where multiple UAVs serve as aerial basestations to provide wireless connectivity to ground users through frequency division multi-access (FDMA) scheme. The UAVs are exclusively powered by a wireless charging station located on the ground following save-then-transmit protocol. In such a cellular network joint optimization for user association, resource allocation and basesation placement is investigated to maximize the downlink sum rate. The problem is formulated as a mixed integer optimization problem and is thus challenging to solve. We propose an efficient solution based on alternate optimization by iteratively solving one of the three subproblems at a time and an algorithm based on penalty method and successive convex optimization to binarize the association indicators. Numerical result shows that the downlink sum rate cannot be always enhanced by deploying more UAVs due to non-negligible tradeoff between energy/communication sources and co-channel interference.

The mobile telecom carriers have deployed massive infrastructures that support or carry the data signal transmission. Most of them are passive devices lacking the ability to actively monitor and automatically report information, which are called “dumb devices.” At present, the dumb device management has problems, such as information incomplete or inaccurate, and lack of dynamic update mechanism. For catering to the construction of smart cities, we design an information management system considering the full lifecycle management for dumb devices to realize real-time or periodical context awareness and information transmission based on narrow band Internet of Things (NB-IoT). Compared to the existing radio frequency identification (RFID)-based solutions, which require RFID readers and electronic tags and have a limited sensing distance, the NB-IoT-based solution for dumb device management has advantages in transmission distance and communication stability. The NB-IoT terminal is attached to the dumb device, and a global positioning system module is installed on it to obtain the positioning information. The NB-IoT terminal is controlled by an real-time clock (RTC) alarm to periodically enter the low power mode and then wake up to automatically collect the location and battery information and upload it to the server. The application objects of this information management system can be extended to dumb devices in other industries.

Mobile edge computing (MEC) plays an important role in next-generation networks. It aims to enhance processing capacity and offer low-latency computing services for Internet of Things (IoT). In this paper, we investigate a resource allocation policy to maximize the available processing capacity (APC) for MEC IoT networks with constrained power and unpredictable tasks. First, the APC which describes the computing ability and speed of a served IoT device is defined. Then its expression is derived by analyzing the relationship between task partitioning and resource allocation. Based on this expression, the power allocation solution for the single-user MEC system with a single subcarrier is studied and the factors that affect the APC improvement are considered. For the multiuser MEC system, an optimization problem of APC with a general utility function is formulated and several fundamental criteria for resource allocation are derived. By leveraging these criteria, a binary-search water-filling algorithm is proposed to solve the power allocation between local CPU and multiple subcarriers, and a suboptimal algorithm is proposed to assign the subcarriers among users. Finally, the validity of the proposed algorithms is verified by Monte Carlo simulation.

With the developments of communication technologies and smart manufacturing, Industrial Internet of Things (IIoT) has emerged. Software-defined networking (SDN), a promising paradigm shift, has provided a viable way to manage IIoT dynamically, called software-defined IIoT (SDIIoT). In SDIIoT, lots of data and flows are generated by industrial devices, where a physically distributed but logically centralized control plane is necessary. However, one of the most intractable problems is how to reach consensus among multiple controllers under complex industrial environments. In this paper, we propose a blockchain (BC)-based consensus protocol in SDIIoT, along with detailed consensus steps and theoretical analysis, where BC works as a trusted third party to collect and synchronize network-wide views between different SDN controllers. Specially, it is a permissioned BC. In order to improve the throughput of this BC-based SDIIoT, we jointly consider the trust features of BC nodes and controllers, as well as the computational capability of the BC system. Accordingly, we formulate view change, access selection, and computational resources allocation as a joint optimization problem. We describe this problem as a Markov decision process by defining state space, action space, and reward function. Due to the fact that it is difficult to solve this joint problem by traditional methods, we propose a novel dueling deep Q-learning approach. Simulation results are presented to show the effectiveness of our proposed scheme.

The cloud-based Internet of Things (IoT) develops rapidly but suffer from large latency and backhaul bandwidth requirement, the technology of fog computing and caching has emerged as a promising paradigm for IoT to provide proximity services, and thus reduce service latency and save backhaul bandwidth. However, the performance of the fog-enabled IoT depends on the intelligent and efficient management of various network resources, and consequently the synergy of caching, computing, and communications becomes the big challenge. This paper simultaneously tackles the issues of content caching strategy, computation offloading policy, and radio resource allocation, and propose a joint optimization solution for the fog-enabled IoT. Since wireless signals and service requests have stochastic properties, we use the actor-critic reinforcement learning framework to solve the joint decision-making problem with the objective of minimizing the average end-to-end delay. The deep neural network (DNN) is employed as the function approximator to estimate the value functions in the critic part due to the extremely large state and action space in our problem. The actor part uses another DNN to represent a parameterized stochastic policy and improves the policy with the help of the critic. Furthermore, the Natural policy gradient method is used to avoid converging to the local maximum. Using the numerical simulations, we demonstrate the learning capacity of the proposed algorithm and analyze the end-to-end service latency.

To promote future intelligent systems, a novel cyber-physical-social smart system (CPS3) powered by the Internet of Things is presented in this paper, where wireless network virtualization is adopted to enhance the diversity and the flexibility of the service operation and the system management. Based on the presented system, a robust energy-efficient resource allocation scheme is proposed to guarantee the outage probability requirements of controllers and actuators while realizing the maximization of the system energy efficiency. Different from the existing works, imperfect channel state information is studied for energy-efficient resource allocation in CPS3. To effectively handle the formulated optimization problem, the concept of virtual devices is introduced to equivalently reformulate the original problem. Afterward, the probabilistic mixed problem is approximately transformed into a nonprobabilistic problem though outage probability analyses. After the transformation, the optimization problem can be decomposed into power allocation and channel allocation, where an iterative algorithm for power allocation is adopted to maximize the system energy, and a heuristic greedy algorithm is presented to schedule sensors and actuators on different subchannels based on the obtained power allocation results. Simulation results demonstrate the convergency of the proposed algorithm and the advantages of the proposed scheme.

Simultaneous wireless information and power transfer (SWIPT) technique offers a potential solution to ease the contradiction between high data rate and long standby time in the fifth generation (5G) mobile communication systems. In this paper, we focus on the SWIPT network design and optimization with 5G new frequencies. To design an efficient SWIPT network, we first investigate the propagation properties of 5G low-frequency (LF) and high-frequency (HF) channels. Specifically, a measurement campaign focusing on 3.5 GHz and 28 GHz is conducted in both outdoor and outdoor-to-indoor scenarios. Motivated by the measurement results, we design a dual-band SWIPT network, where the HF band is used for short-distance information delivery, while the LF band is used for short-distance energy transfer and long-distance information delivery. The designed network has a win-win architecture which can enhance the throughput of cell-edge users and improve the energy-harvesting efficiency of cell-center users. To further boost the network performance, we devise a joint power-and-channel allocation algorithm, which has the advantages of low complexity and fast convergence. Finally, simulation results demonstrate that the designed dual-band network outperforms the conventional single-band network in terms of energy-harvesting efficiency and user fairness, and the proposed algorithm can further upgrade the network performance significantly.

Distributing data aggregation through multiple access channel (MAC) has been challenging in large wireless networks. In order to tackle the challenge, a computing over the MAC (CP-MAC) scheme has been proposed as a promising communication-computation integrated way for function-centric networks. In this paper, we analyze the performance of the CP-MAC scheme, compared with the traditional communication-computation separated way, i.e., a communicating over the MAC (CM-MAC) scheme. Function-centric wireless networks are considered, where the fusion center (FC) does not need the individual data of each node but only the target function. We begin with the ideal uniform-MAC scenarios, where the CP-MAC scheme is always better than the CM-MAC scheme. Then, practical non-uniform MAC scenarios are studied for both homogeneous networks with Rayleigh fading and heterogeneous networks with a different path loss. Closed-form expressions of the achievable function rate are provided using the asymptotic theory of ordered statistics. It is found that the CP-MAC scheme is not always superior to the CM-MAC scheme. Simulation results are provided to verify and illustrate our derived results.

With the capability of bidirectional communications on a single frequency band, the full-duplex (FD) operation can potentially double the spectral efficiency in physical layer. In network layer, nevertheless, it may cause severe mutual interference to the system. In this paper, we exploit interference alignment (IA) to address the interference in small cell networks, where some of the base stations simultaneously serve both uplink and downlink users on the same frequency via FD. Under such scenario, we first derive the feasibility condition for IA from Bezout's theorem and find that IA can be feasible only if a certain size constraint of the network is satisfied. On this basis, we then propose two clustering methods, i.e., minimized spectrum consumption clustering (MSCC) and minimized interference leakage clustering (MILC), both of which can perfectly eliminate the intra-cluster interference with IA. The difference between them is that MSCC aims at minimizing the number of clusters through allocating orthogonal resource blocks (RBs) for each cluster to avert inter-cluster interference, while MILC tries to minimize the aggregated inter-cluster interference with all clusters sharing the same RB. Extensive simulations verify that MSCC can achieve higher system sum rate, but MILC works better in terms of spectral efficiency.

Recent advances in the industrial Internet of things (IIoT) provide plenty of opportunities for various industries. To address the security and efficiency issues of the massive IIoT data, blockchain is widely considered as a promising solution to enable data storing/processing/sharing in a secure and efficient way. To meet the high throughput requirement, this paper proposes a novel deep reinforcement learning (DRL)-based performance optimization framework for blockchain-enabled IIoT systems, the goals of which are threefold: 1) providing a methodology for evaluating the system from the aspects of scalability, decentralization, latency, and security; 2) improving the scalability of the underlying blockchain without affecting the system's decentralization, latency, and security; and 3) designing a modulable blockchain for IIoT systems, where the block producers, consensus algorithm, block size, and block interval can be selected/adjusted using the DRL technique. Simulations results show that our proposed framework can effectively improve the performance of blockchain-enabled IIoT systems and well adapt to the dynamics of the IIoT.

Big data is becoming a research focus in intelligent transportation systems (ITS), which can be seen in many projects around the world. Intelligent transportation systems will produce a large amount of data. The produced big data will have profound impacts on the design and application of intelligent transportation systems, which makes ITS safer, more efficient, and profitable. Studying big data analytics in ITS is a flourishing field. This paper first reviews the history and characteristics of big data and intelligent transportation systems. The framework of conducting big data analytics in ITS is discussed next, where the data source and collection methods, data analytics methods and platforms, and big data analytics application categories are summarized. Several case studies of big data analytics applications in intelligent transportation systems, including road traffic accidents analysis, road traffic flow prediction, public transportation service plan, personal travel route plan, rail transportation management and control, and assets maintenance are introduced. Finally, this paper discusses some open challenges of using big data analytics in ITS.

With an increasing number of machine-type communication devices (MTCDs), machine-to-machine (M2M) communications have attracted great attentions from both academia and industry. Different from traditional communication networks, the data connections with M2M communications are typically small-sized but with high frequency, necessitating the efficiency optimization of both energy consumption and computation. In this paper, we introduce mobile edge computing (MEC) into virtualized cellular networks with M2M communications, to decrease the energy consumption and optimize the computing resource allocation as well as improve computing capability. Moreover, based on different functions and quality of service (QoS) requirements, the physical network can be virtualized into several virtual networks, and then each MTCD selects the corresponding virtual network to access through the embedded-SIM (eSIM) technology. Meanwhile, the random access process of MTCDs is formulated as a partially observable Markov decision process (POMDP) to minimize the system cost, which consists of both the energy consumption and execution time of computing tasks. Furthermore, to facilitate the network architecture integration, software-defined networking (SDN) is introduced to deal with the diverse protocols and standards in the networks. Extensive simulation results with different system parameters reveal that the proposed scheme could significantly improve the system performance compared to the existing schemes.

Non-orthogonal multiple access (NOMA) has been proposed as a promising multiple access approach for 5G mobile systems because of its superior spectrum efficiency. However, the privacy between the NOMA users may be compromised due to the transmission of a superposition of all users' signals to successive interference cancellation (SIC) receivers. In this paper, we propose two schemes based on beamforming optimization for NOMA that can enhance the security of a specific private user while guaranteeing the other users' quality of service (QoS). Specifically, in the first scheme, when the transmit antennas are inadequate, we intend to maximize the secrecy rate of the private user, under the constraint that the other users' QoS is satisfied. In the second scheme, the private user's signal is zero-forced at the other users when redundant antennas are available. In this case, the transmission rate of the private user is also maximized while satisfying the QoS of the other users. Due to the non-convexity of optimization in these two schemes, we first convert them into convex forms, and then, an iterative algorithm based on the Concave-Convex Procedure is proposed to obtain their solutions. The extensive simulation results are presented to evaluate the effectiveness of the proposed schemes.

Blockchain-based video streaming systems aim to build decentralized peer-to-peer networks with flexible monetization mechanisms for video streaming services. On these blockchain-based platforms, video transcoding, which is computationally intensive and time-consuming, is still a major challenge. Meanwhile, the block size of the underlying blockchain has significant impacts on the system performance. Therefore, this paper proposes a novel blockchain-based framework with an adaptive block size for video streaming with mobile edge computing (MEC). First, we design an incentive mechanism to facilitate collaboration among content creators, video transcoders, and consumers. In addition, we present a block size adaptation scheme for blockchain-based video streaming. Moreover, we consider two offloading modes, i.e., offloading to the nearby MEC nodes or a group of device-to-device (D2D) users, to avoid the overload of MEC nodes. Then, we formulate the issues of resource allocation, scheduling of offloading, and adaptive block size as an optimization problem. We employ a low-complexity alternating direction method of the multipliers-based algorithm to solve the problem in a distributed fashion. Simulation results are presented to show the effectiveness of the proposed scheme.

Future networks are expected to connect an enormous number of nodes wirelessly using wide-band transmission. This brings great challenges. To avoid collecting a large amount of data from the massive number of nodes, computation over multi-access channel (CoMAC) is proposed to compute a desired function over the air utilizing the signal-superposition property of wireless channel. Due to frequency-selective fading, wide-band CoMAC is more challenging and has never been studied before. In this paper, we propose the use of orthogonal frequency division multiplexing (OFDM) in wide-band CoMAC to transmit functions in a similar way to bit sequences through division, allocation, and reconstruction of functions. An achievable rate without any adaptive resource allocation is derived. To prevent a vanishing computation rate from the increase in the number of nodes, a novel sub-function allocation of sub-carriers is derived. Furthermore, we formulate an optimization problem considering power allocation. A sponge-squeezing algorithm adapted from the classical water-filling algorithm is proposed to solve the optimal power allocation problem. The improved computation rate of the proposed framework and the corresponding allocation has been verified through both theoretical analysis and simulation.

Recent advances of computing, caching, and communication (3C) can have significant impacts on mobile social networks (MSNs). MSNs can leverage these new paradigms to provide a new mechanism for users to share resources (e.g., information, computation-based services). In this paper, we exploit the intrinsic nature of social networks, i.e., the trust formed through social relationships among users, to enable users to share resources under the framework of 3C. Specifically, we consider the mobile edge computing (MEC), in-network caching and device-to-device (D2D) communications. When considering the trust-based MSNs with MEC, caching and D2D, we apply a novel big data deep reinforcement learning (DRL) approach to automatically make a decision for optimally allocating the network resources. The decision is made purely through observing the network's states, rather than any handcrafted or explicit control rules, which makes it adaptive to variable network conditions. Google TensorFlow is used to implement the proposed deep Q-learning approach. Simulation results with different network parameters are presented to show the effectiveness of the proposed scheme.

The exponential growth of data traffic volume dominates the demand for the next generation mobile networks (5G). The consistent and satisfied quality of experience (QoE) is one of the leading challenges of provisioning services in 5G mobile networks. Thus, in this paper, we propose a novel adaptive prefetching scheme to compensate the undesired transmission conditions in the 5G mobile network by extending the content prefetching concept from the users to the network. Specifically, an optimization problem is proposed for a prefetching scheme that adaptively retrieves users' data to access nodes and (or) user equipments (UEs) before the actual requests according to the network status, QoE status, predicted data rates, and mobility patterns of users. For the sake of tractability, the prefetching problem is transferred to a convex problem that can be solved efficiently. Accordingly, to implement the proposed schemes in the 5G network, system interactions among entities in the network are designed to realize prefetching-related functions. A signaling protocol to support the adaptive prefetching scheme is also presented. Simulation results show that an adaptive prefetching scheme can improve the network performance significantly.

By bringing computing capabilities to the network edge, mobile edge computing (MEC) has emerged as a promising technique to enable low-latency video streaming services. However, due to the rapid growth of the number of devices and the heterogeneous formats of the video streams, the traditionally centralized content delivery schemes are insufficient to provide secure, adaptive video services with low complexity. To achieve a decentralized content market among untruthful parties (e.g., users and operators), in this paper, we propose an effective video transcoding and delivery approach based on MEC and blockchain. In the proposed approach, we envision a set of blockchain-based smart contracts to build an autonomous content delivery market, where all the participants are financially enforced by smart contract terms. Then, users, small base stations (SBSs), and content provider (CP) are able to autonomously adjust their strategies according to the content market statistics. Moreover, we formulate the optimization problem, including resource allocation, determining content price and quality levels of contents, as a three-stage Stackelberg game. We analyze the subgame equilibrium for each stage and the interplays of the three-stage game. Lastly, an iterative algorithm is proposed to obtain the solution. Simulation results are presented to show the effectiveness of the proposed approach.

In this paper, we propose a joint dynamic pilot and data power allocation scheme for time division duplex (TDD) massive multiple-input multiple-output (MIMO) systems, so as to both adaptively mitigate pilot contamination and balance the mutual interference. Due to the unknown of instant channel state information before pilots, we exploit the Gauss-Markov process of temporally-correlated channels and use the Kalman filter to not only filter out the pilot contamination but also provide the priori estimation values. Subsequently, the deterministic approximation of the rate is derived as a function of the priori channel estimation and the priori estimate errors, and accordingly the rate-profile maximization to achieve max-min fairness is formulated. To deal with this optimization coupled across the pilot power and data power as well as the users, we give an iterative alternating rate-suboptimal algorithm composed of two sub-problems, both of which are further solved by introducing the successive convex approximation (SCA) methods and slack variables. Numerical results confirm the improved rate provided by the proposed scheme.

Offloading computationally intensive tasks from user equipments (UEs) to mobile edge computing (MEC) servers is a promising technique to boost up the computational capacity of UEs. However, MEC will incur extra energy consumption and time delays, which motivates the deployment of energy harvesting (EH) small cell networks with MEC in mobile networks. Due to the complexity of such networks, it is challenging to effectively allocate resources for UEs. In this paper, we investigate the offloading decision, wireless and computational resources allocation problem in energy harvesting (EH) small cell networks with MEC. Different from existing literatures, our research focuses on improving mobile operators' revenue by maximizing the amount of the offloaded tasks while decreasing the energy expenditure and time-delays. Besides, queues are created at the MEC server side to store the un-executed tasks in a time slot, which is used as a punishment in our utility function to avoid serious delay. Considering the varying lengths of queues, the states of EH-batteries of small base stations (SBSs) and down-link channels, the above problem is modeled as a Markov decision process (MDP). Since the states and actions in the MDP are infinite, an online and on-policy actor-critic with eligibility traces algorithm is proposed to resolve the problem. Simulation results show the proposed algorithm has superior performances compared with the policy-gradient algorithm and Q-learning.

Vehicular ad hoc networks (VANETs) have become a promising technology in smart transportation systems with rising interest of expedient, safe, and high- efficient transportation. Dynamicity and infrastructure-less of VANETs make it vulnerable to malicious nodes and result in performance degradation. In this paper, we propose a software- defined trust based deep reinforcement learning framework (TDRL-RP), deploying a deep Q-learning algorithm into a logically centralized controller of software-defined networking (SDN). Specifically, the SDN controller is used as an agent to learn the highest routing path trust value of a VANET environment by convolution neural network, where the trust model is designed to evaluate neighbors' behaviour of forwarding packets. Simulation results are presented to show the effectiveness of the proposed TDRL-RP framework.

Mobile edge computing (MEC) has emerged as a potential paradigm to enhance the processing capabilities of mobile user equipments (MUEs), while edge caching has become a promising means of alleviating traffic in the backhual. In this paper, we formulate a stochastic optimization problem to maximize the average economical profit of MEC server by jointly optimizing offloading decision and caching decision making, and the allocation of radio, computing, and caching resources in a cellular network, with network stability taken into account. To tackle this problem, we develop an online algorithm referred to as dynamic joint computation offloading, resource allocation, and content caching algorithm (DJORC) based on Lyapunov optimization theory. Specifically, the proposed DJORC only needs the current states of the system, and without requiring any prior-knowledge. By further using 0-1 integer programming and linear programming, the closed-form solution of the formulated problem is obtained. Simulation results are presented to verify the performance of DJORC under different parameter settings, as well as the performance gains obtained by DJORC over other existing schemes.

Internet of things (IoT) are envisioned to be an essential in our daily lives, but most IoT devices (IDs) are battery powered and have limited resource. Recently, fog computing (FC) has been proposed to support IoT systems, where part or all of the data are offloaded from IDs to fog nodes for processing or computation. In this paper, we propose to optimise the partial computation offloading in an OFDMA based FC IoT system, while ensuring fairness among IoT links with respect to their energy consumption. In particular, we minimize the energy consumption of the worst-case link by jointly optimizing the size of offloaded data and the assignment of subcarriers, while guaranteeing the rate requirement. The formulated min-max energy efficiency optimization problem (MEP) is solved using Lagrangian dual decomposition and subgradient projection, bases on which we propose an iterative algorithm. Our simulation results show that the proposed resource allocation algorithm is more energy efficient than the existing algorithms for FC supported IoT, while achieving fairness among IoT links.

In this paper, we develop computation offloading scheme based on device-to-device (D2D) communications. The scheme is proposed for effective computation execution where some mobile devices (MDs) could offload their computation intensive tasks to appropriate nearby MDs where necessary, with the assistance of the base station. Accordingly, we formulate a stochastic optimization problem to minimize the average expenses (e.g., wireless communication expense, computation service expense) of MDs in task offloading, while considering the computation resource budget constraint to restraint the behavior of the overuse of computation resource and guarantee mobile users' motivation for collaboration not impaired. To solve this problem, we propose an algorithm that does not need any prior- knowledge of available resources of MDs, referred to as the SEEP. To address a couple and mixed combinational subproblem in the SEEP, we decouple optimization variables for suboptimal. By doing so, both task scheduling and subcarrier assignment are obtained in closed forms, while power allocation is solved by developing efficient iterative algorithm that exploits D.C. (difference of convex functions) structure. Simulation results show the convergence of the SEEP, and illustrate SEEP can flexibly coordinate the tradeoff between expenses and delay, and can substantially reduce expenses of MDs against other existing schemes.

With the explosion in the number of connected devices and Internet of Things (IoT) services in smart city, the challenges to meet the demands from both data traffic delivery and information processing are increasingly prominent. Meanwhile, the connected vehicle networks have become an essential part in smart city, bringing massive data traffic as well as significant networking, caching and computing resources. In this paper, we propose a novel vehicle network architecture, mitigating the network congestion with the joint optimization of networking, caching and computing. Cloud computing at the data centers as well as mobile edge computing (MEC) at the evolved node Bs (eNodeBs) and on-board units (OBUs) are taken as the paradigms to provide caching and computing resources. The programmable control principle originated from software-defined networking (SDN) paradigm has been introduced to facilitate the system architecture and resource integration. With the careful modeling of the services, the vehicle mobility and the system state, a joint resource management scheme is proposed and formulated as a partially observable Markov decision process (POMDP) to minimize system cost, which consists of both network overhead and execution time of computing tasks. Extensive simulation results with different system parameters reveal that the proposed scheme could significantly improve the system performance compared to the existing schemes.

Vehicular Ad Hoc Networks (VANETs) have been attracting more and more attention. However, due to the fast movement of vehicles and dynamic topology change, designing an efficient routing protocol in the complex urban environment is quite challenging. In this paper, an Intersection-based Geographic Routing with Transmission Quality guaranteed in Urban VANETs (IGRTQ) is proposed. As a selection guidance of the best route, each road segment is assigned with a weight based on the collected information related to the delay and connectivity of each road segment. Based on the weight information, the road segment can be dynamically selected one by one to form the optimized routing path, avoiding the local maximum and data congestion. An improved greedy strategy is further proposed to forward the packet along the selected road segment, ensuring the fast and reliable packet transmission. Simulation results show that the proposed protocol provides higher packet delivery ratio and lower end-to-end delay compared to the existing protocols.

Driven by the large-scale video traffic, mobile edge computing (MEC) has emerged as a promising technique that extends cloud-computing capabilities to the proximate small base stations (SBSs) in wireless networks, especially in ultra-dense networks (UDNs). With MEC, video transcoding, which processes the adaptive bitrates of a video and provides the adaptive video streaming to users, can significantly release the backhaul burden of networks. However, video transcoding is a time-consuming task, and how to guarantee quality-of- service (QoS) for large video data with MEC is still challenging. To address this issue, in this paper, we propose a joint SBSs selection, tasks scheduling, and resource allocation approach for achieving a delay- optimal transcoding under the constraints of network cost. Specifically, to reduce the delay, a set of SBSs are formed into a Virtual SBSs Group (VSG) to perform the video transcoding and delivering in parallel for a given user. Then, the joint tasks scheduling and feasible resource allocation are performed to minimizing total delay while maintaining a low network cost. The optimization problem is formulated as a mixed integer non- convex programming problem and a three-stage search solution is proposed to solve it. Simulation results show that our proposed approach can significantly improve the transcoding performance while satisfying the resource consumption constraint.

Security is a challenging issue for cognitive radio (CR) networks. Conventionally, interference will degrade the performance of a primary user (PU) when the spectrum is shared with secondary users (SUs). However, when properly designed, SUs can serve as friendly jammers to guarantee the secure transmission of PU. Thus, in this paper, we propose a optimal transceiver design scheme to improve the sum rate of SUs while guaranteeing the secrecy rate of PU. In the scheme, the secondary transceivers are jointly designed to maximize their sum rate while satisfying a threshold on the PU's secrecy rate. Due to the non-convex nature, it is first converted into a convex one and then, an alternating optimization algorithm based on the second-order cone programming is proposed to solve it. Finally, simulation results are presented to verify the effectiveness of the proposed scheme for secure CR networks.

Ultra-dense network (UDN), as well as nonorthogonal multiple access (NOMA), has been emerging as promising techniques to meet the growing demand of data traffic in next-generation wireless networks. However, due to the spectrum sharing among SBSs and users, interference management (IM) is becoming a more important issue in NOMA-based UDN. Moreover, the massive small base stations (SBSs) with various types and overlapped coverage require more intelligent and efficient mechanisms for the IM problem. Thus, in this paper, to reduce interference and improve operation efficiency, we propose a self-optimizing resource allocation (SORA) scheme for IM with joint consideration of the dynamic interference conditions and fierce resource competition among SBSs. Concretely, each SBS constructs the interfering SBSs group adaptively to represent the potential interference from other SBSs. Then, to reduce interference and meet users' requirements, each SBS performs the resource allocation including sub-band and power allocation independently. Moreover, we formulate the problem as a non-cooperation satisfaction game, where a satisfaction function is established for evaluating each SBS's utility. When every SBS's utility is above a preset threshold, the game is considered to reach the satisfaction equilibrium. A distributed algorithm is designed to enable each SBS to learn the satisfaction equilibrium and allocate the resource autonomously. Simulation results show the effectiveness of the proposed scheme compared with the traditional schemes.