As the maritime industry evolves, applications such as real-time navigation, ocean monitoring, and emergency rescue increasingly require reliable communication. They also demand efficient computation offloading to support AI-driven services. However, terrestrial networks offer sparse coverage and unstable links in open-sea environments, severely constraining both connectivity and the execution of computation-intensive tasks. Although low Earth orbit (LEO) satellites extend coverage over oceans, their frequent handovers and high operating costs hinder stable, low-latency communication and efficient computation offloading. Unmanned aerial vehicle-based relays can enhance connectivity, but their limited endurance and environmental vulnerability hinder large-scale deployment. In contrast, high-altitude platform stations (HAPS) offer quasi-stationary positioning, broad coverage, and long operational duration, making them promising intermediaries between LEO satellites and maritime users. Building on this motivation, we design a space–air–ground–sea integrated network architecture in which HAPS function as relay nodes. We model the multi-layer task offloading and resource allocation as a partially observable Markov decision process to capture the uncertainty and dynamics of maritime environments. To solve it, we adopt multi-agent proximal policy optimization, which enables centralized training with decentralized execution. Furthermore, we incorporate an exact potential game mechanism into the reward design to enhance agent coordination and ensure alignment with system-wide objectives. Simulation results show that our method outperforms three representative baselines; under maximum task load, it reduces average latency, energy consumption, and overall cost by 8.7%, 2.0%, and 18.6%, respectively, verifying its effectiveness for computation-intensive maritime services.

The widespread adoption of 5G technology has significantly enhanced the interconnectivity and automation of factory equipment, increasing the complexity of device coordination. Equipment failures or performance degradation can result not only in individual device downtime but also in cascading disruptions across the entire production line, potentially leading to factory-wide shutdowns and decreased production efficiency. Ensuring high equipment reliability—particularly fault tolerance under extreme conditions—is therefore critical to the stable operation of a 5G fully connected factory. To address this challenge, this paper proposes a proactive fault tolerance (FT) framework that enhances system reliability through parallel redundancy strategies. The framework begins by evaluating system reliability based on device responses in the most recent service cycle. If the computed FT value falls below the threshold specified by the Service Level Agreement (SLA), the framework is triggered. Devices are then ranked according to their historical performance and their criticality in the upcoming production cycle. A “1+N” redundancy allocation strategy is applied, and the resulting Integer Nonlinear Programming (INLP) problem is solved using a Redundancy Allocation Algorithm based on Transformer-enhanced Deep Reinforcement Learning (RAA-TDRL). To validate the proposed framework, we conduct simulations using a real-world dataset to emulate performance fluctuations. Experimental results show that the proposed key equipment identification method and RAAT-DRL algorithm achieve superior performance and reduced time complexity compared to baseline approaches.

The growing adoption of AI-Generated Content (AIGC) has made large-scale processing of multiple Generative AI (GAI) training jobs a key strategy for improving cost-efficiency in computing clusters. However, the distributed nature of GAI models, together with inherent network bottlenecks, imposes significant challenges on system performance. Moreover, differences in training purposes, variations in model sizes, and asynchronous lifecycles create a dynamic environment. As a result, the coexistence of multiple GAI training jobs in a computing cluster exacerbates problems such as resource misallocation, fragmentation, and network contention, leading to low resource utilization and inefficient training performance. These motivate us to explore an efficient resource scheduling approach for completing multiple GAI training jobs. Accordingly, we introduce an intrinsic topology-aware scheduling framework designed to ensure flexible scheduling and efficient distributed training of GAI models. To address the trade-off between the number of concurrent jobs and the communication contention they generate, we formulate a multi-objective optimization problem with two objectives: maximizing the utility of GAI jobs and minimizing communication bandwidth. We then propose the Diffusion Model-based AI-Generated Resources Scheduling (DARS) algorithm, designed to capture dynamic, high-dimensional environments and generate optimal resource scheduling decisions. DARS employs a denoising diffusion process to iteratively refine noisy resource allocations into optimized scheduling decisions. Subsequently, we replace the policy network of Deep Reinforcement Learning (DRL) with DARS to address environmental uncertainty and enhance efficiency. Finally, the simulation results confirm that the proposed algorithm outperforms existing approaches.

As an advanced form of IoT technology, integrated sensing and communication (ISAC) deeply integrates communication and perception, enhancing the performance and application range of IoT. At the same time, the space-air-ground integrated network (SAGIN) provides a wider and more efficient connection and information processing support for both. However, the highly dynamic and time-varying characteristics of SAGIN lead to more frequent vertical handovers among heterogeneous wireless networks, which seriously affects the continuity and reliability of services. This motivates us to explore an effective vertical handover method in SAGIN to guarantee the quality of network service. The issue is a typical complex and high-dimensional problem with its online and dynamic characteristics, which provides a particularly favorable scenario for the adaptability of the diffusion model (DM). Accordingly, we propose a novel vertical handover decision algorithm with the aid of DM. First, we innovate a novel vertical handover analytical model that describes handover jitter, load difference, and handover robustness. Then we formulate it as a multiobjective optimization problem. Next, inspired by Generative AI (GAI), we propose a DM-based GAI-empowered handover decision (DGHD) algorithm to capture the time-varying and high-dimensional environments and generate optimal vertical handover decisions. Subsequently, the policy network of multiagent proximal policy optimization (MAPPO) is replaced with the proposed DGHD for addressing environmental uncertainty and enhancing efficiency. Finally, the simulations exhibit that our proposed algorithm outperforms existing algorithms.

This paper proposes a weight-aware deep reinforcement learning (WADRL) approach designed to address the multiobjective vehicle routing problem with time windows (MOVRPTW), aiming to use a single deep reinforcement learning (DRL) model to solve the entire multiobjective optimization problem. The Non-dominated sorting genetic algorithm-II (NSGA-II) method is then employed to optimize the outcomes produced by the WADRL, thereby mitigating the limitations of both approaches. Firstly, we design an MOVRPTW model to balance the minimization of travel cost and the maximization of customer satisfaction. Subsequently, we present a novel DRL framework that incorporates a transformer-based policy network. This network is composed of an encoder module, a weight embedding module where the weights of the objective functions are incorporated, and a decoder module. NSGA-II is then utilized to optimize the solutions generated by WADRL. Finally, extensive experimental results demonstrate that our method outperforms the existing and traditional methods. Due to the numerous constraints in VRPTW, generating initial solutions of the NSGA-II algorithm can be time-consuming. However, using solutions generated by the WADRL as initial solutions for NSGA-II significantly reduces the time required for generating initial solutions. Meanwhile, the NSGA-II algorithm can enhance the quality of solutions generated by WADRL, resulting in solutions with better scalability. Notably, the weight-aware strategy significantly reduces the training time of DRL while achieving better results, enabling a single DRL model to solve the entire multiobjective optimization problem.

In the 6G era, user demand for low-latency, cost-effective extreme services such as extended reality (XR) and holographic communications has significantly increased. Multi-domain networks, known for their vast capacity and coverage, are essential in fulfilling the growing demand for high-performance services. Despite their potential, these networks face challenges with domain isolation, requiring a software defined network (SDN) controller for inter-domain communication. Network function virtualization (NFV) enhances flexibility of service delivery with customizable service function chain (SFC), yet prior research falls short in delivering low-latency, cost-efficient services in multi-domain NFV networks alongside an unreasonable assumption that software on physical nodes can support the execution of all virtualization network functions (VNFs). In this paper, we study the problem of SFC deployment with VNF-dependent software migration (SD-VDSM) in multi-domain networks. Particularly, we first formulate the problem by setting an objective to minimize the end-to-end communication delay and the associated costs of service provisioning, while simultaneously ensuring load balancing across multi-domain networks. However, complexity of the issue escalates to an intractable level due to the intertwined nature of SFC deployment strategies and VNF-dependent software migration tactics, which mutually influence each other intricately. To tackle this issue, we propose an innovative heuristic algorithm, designated as the Joint SFC Deployment with VNF-Dependent Software Migration Algorithm (JSD-VDSMA). Comprising three fundamental steps, this algorithm is crafted to adeptly resolve the complexities of service provisioning across multi-domain networks. A suite of rigorous experimental assessments is detailed, demonstrating the capability of our proposed JSD-VDSMA. Through these comparative analyses, we demonstrate its effectiveness not only to increase the service acceptance rate but also to diminish both the end-to-end communication delay and resource utilization costs in comparison to its counterparts.

Unmanned aerial vehicle (UAV) images are widely applied in real-world applications. However, traditional object detection methods have encountered a sharp performance drop on UAV images, largely due to the presence of numerous small objects with low resolution, sparse textures, and blurred features. In this paper, we propose a novel Adaptive Anchor Detection Network (AADN) that dynamically generate anchors through deep neural network. Subsequently the proposed Temporal And Self-Attention (TAA) module enhances the extracted image features by capturing temporal and internal correlation information. Furthermore, we introduce a data augmentation method based on category k-means clustering, designed to reduce irrelevant background regions and improve the network’s ability to distinguish objects with similar categories. Extensive experiments show the superior performance of AADN on both the VisDrone2019 and TinyPerson dataset, achieving an impressive 2% increase in mean average precision (mAP) compared to state-of-the-art methods.

With the development of mobile services such as autonomous driving and the industrial internet, ultralow latency and pervasive mobility have become key characteristics of the intelligent interconnections among people, machines, and things. As a prevailing mobile network architecture, the network function virtualization (NFV)-based edge-cloud architecture brings the computing and memory resources closer to the end user, significantly reducing service delays and supporting more efficient mobility management. However, the geographically distributed nature of the edge-cloud architecture and the quality of service (QoS) requirements of latency-sensitive services in extreme mobile scenarios make service function chain (SFC) deployment more challenging. In this paper, we investigate a mobility-aware SFC deployment scheme with service migration in an NFV-based edge-cloud system. To properly cope with the mobility pattern of mobile services, a multistage decision-making problem is formulated, aiming to jointly minimize the long-term deployment and migration costs and the average end-to-end service latency while simultaneously satisfying various QoS constraints for services and the physical resource constraints of the edge-cloud system. Then, to address the formulated problem, a deep reinforcement learning (DRL)-based online SFC deployment algorithm is proposed that can automatically detect variations in the widely distributed edge-cloud environment and generate online deployment solutions without human intervention to implement adaptive and fast service provision and also support mobile service migration. Extensive experimental results demonstrate our proposed scheme surpasses its competitors in terms of end-to-end latency and migration cost, with average reductions of 6.26% and 18.77%, respectively, while improving the average service acceptance rate by 19.19%.

Indoor localization is of great significance to a wide range of applications in the era of mobile computing. The maturity of the computer vision techniques and the ubiquity of embedded sensors in commercial off-the-shelf (COTS) smartphones shed the light on the submeter localization services for indoor environment. The state-of-the-art indoor localization works suffer from high-cost deployment and inaccurate results due to the coarse readings from internal measurement units (IMUs) sensors in the smartphones. In this article, we mainly innovate in introducing the WiFi-sensing technology to extract the distance information in a low-cost and device-free manner. Along with the computer vision technology, we model and implement an accurate and easy-to-deploy system for indoor localization. This system enhances indoor localization with multimodal sensing via two images, IMU sensors reading and CSI of WiFi signal. Specifically, we first model and design camera-based, sensor and WiFi-assisted indoor localization and propose several algorithms in this model. We then implement a prototype with smartphones and commercial WiFi devices and evaluate it in several distinct indoor environments. The experimental results show that 92-percentile error is within 0.2 m for indoor targets which sheds light on submeter indoor localization.

Most existing proposals for indoor localization are “unnatural,” as they rely on sensing abilities not available to human beings. While such a mismatch causes complications in human-computer interactions and thus potentially reduces the usability and friendliness of a localization service, it is partially entailed by the need for low-cost/effort sensing with resource-limited mobile devices. Fortunately, recent developments in smart glasses (e.g., Google Glasses) signal a trend toward realistic visual sensing and hence make the sensing ability of mobile devices more compatible to that of human users. Leveraging such front-end developments, we propose VisioMap as a natural indoor localization system that intentionally mimics the human skills in visual localization. VisioMap uses very sparse photograph samples to reconstruct 3-D indoor scenes; this is facilitated by the facts that photographs are taken at the eye-level with high stability and regularity, and that the reconstruction is lightweight as it exploits geometric features rather than image pixels. Localization is in turn performed by matching the geometric features extracted online to the reconstructed 3-D scene, making VisioMap: 1) natural to users as they can see the matched 3-D scene and 2) dispensed with the need for dense fingerprints/POIs toward accurate localization.

Efficient resource allocation which aims to maximize the network utility under energy neural operation is well known as a key issue in energy harvesting wireless sensor networks (EHWSNs). However, as the energy resource is unstable in practical systems, it's challenging to tackle the uncertainty in harvested energy profile. Instead of designing sophisticated harvested energy prediction model, we directly make uncertainty involved in the resource allocation design. Considering the uncertainty of harvested energy profile, a flexible network utility optimization approach is proposed that can achieve high network utility and robustness against uncertain harvested energy. We firstly formulate the network utility maximization problem subject to energy constraints involving uncertainty. We then introduce a flexible uncertainty model to describe the harvested energy and transform the network utility maximization with uncertainties into a traditional optimization problem. Our experimental results demonstrate the proposed approach is able to provide flexible energy allocation and achieve robustness.

Compressed sensing for wireless sensor networks has attracted a lot of research attention in the last decade for its advantages in energy saving, robustness, and so on. Nevertheless, existing solutions mostly focus on the data compression performance while neglecting the energy efficiency. In this paper, we first present the joint resource allocation problem formulation based on compressed sensing. Then a distributed algorithm to compute the sampling rate and routes utilizing local network status is proposed. We conduct extensive experiments based on meteorological wireless sensor networks to verify the merit of our mechanism; it is shown that the proposed mechanism is able to achieve very high efficiency in terms of network lifetime and sensing quality compared with existing approaches.

In this paper, a cost minimization problem is formulated to intelligently schedule energy generations for microgrids equipped with unstable renewable sources and energy storages. In such systems, the uncertain renewable energy will impose unprecedented scheduling challenges. To cope with the fluctuate nature of the renewable energy, an uncertainty model based on renewable energies' moment statistics is developed. Specifically, we obtain the mean vector and second-order moment matrix according to predictions and field measurements and then define uncertainty set to confine the renewable energy generation. The uncertainty model allows the renewable energy generation distributions to fluctuate within the uncertainty set. We develop chance constraint approximations and robust optimization approaches based on a Chebyshev inequality framework to firstly transform and then solve the scheduling problem. Numerical results based on real-world data traces evaluate the performance bounds of the proposed scheduling scheme. It is shown that the temporal-correlation information of the renewable energy within a proper time span can effectively reduce the conservativeness of the solution. Moreover, detailed studies on the impacts of different factors on the proposed scheme provide some interesting insights which shall be useful for the policy making for the future microgrids.

As a key component of building management and security, occupancy inference through smart sensing has attracted a lot of research attentions for nearly two decades. Nevertheless, existing solutions mostly rely on either pre-deployed infrastructures or user device participation, thus hampering their wide adoption. This paper presents CeilingSee, a dedicated occupancy inference system free of heavy infrastructure deployments and user involvements. Building upon existing LED lighting systems, CeilingSee converts part of the ceiling-mounted LED luminaires to act as sensors, sensing the variances in diffuse reflection caused by occupants. In realizing CeilingSee, we first re-design the LED driver to leverage LED's photoelectric effect so as to transform a light emitter to a light sensor. In order to produce accurate occupancy inference, we then engineer efficient learning algorithms to fuse sensing information gathered by multiple LED luminaires. We build a testbed covering a 30m2office area; extensive experiments show that CeilingSee is able to achieve very high accuracy in occupancy inference.

Although Visible Light Communication (VLC) is gaining increasing attention in research, developing a practical VLC system to harness its immediate benefits using Commercial Off-The-Shelf (COTS) devices is still an open issue. To this end, we develop and deploy CeilingTalk as a lightweight wireless broadcast system using COTS LED luminaries as transmitters and smartphone cameras as receivers so that it can be fully hosted in a smartphone and is feasible for all possible indoor environments. CeilingTalk innovates in both encoding and decoding to achieve an adequate throughput for realistic applications. On one hand, it employs Raptor coding to allow multiple LED luminaries to transmit collaboratively so as to benefit both throughput and reliability. On the other hand, it involves a lightweight decoding scheme to handle the asynchrony (both spatial and temporal) in transmissions. Moreover, we analyze the impact of various parameters on the performance of CeilingTalk, in order to derive a model for such VLC systems enabled by COTS devices and hence provide general guidance for future VLC deployments in larger scales. Finally, we conduct extensive field experiments to validate the effectiveness of our LED-camera VLC model, as well as to demonstrate the promising performance of CeilingTalk: up to 1.0 kb/s at a distance of 5 m.

Although Visible Light Communication (VLC) is gaining increasing attentions in research, developing a practical VLC system to harness its immediate benefits using Commercial Off-The-Shelf (COTS) devices is still an open issue. To this end, we develop and deploy CeilingCast as a location-bound wireless broadcast system using COTS LEDs as transmitters and smartphone cameras as receivers. CeilingCast innovates in its effective coding and efficient decoding schemes, so that it can be fully hosted in a smartphone and is feasible for all possible indoor environments. Moreover, we analyze the impact of various parameters on the performance of CeilingCast, in order to derive a model for such VLC systems enabled by COTS devices and hence provide general guidance for future VLC deployments in larger scales. Finally, we conduct extensive field experiments to validate the effectiveness of our LED-camera VLC model, as well as to demonstrate the promising performance of CeilingCast under various parameters.