In autonomous driving, vision-centric 3D object detection recognizes and localizes 3D objects from RGB images. However, due to high annotation costs and diverse outdoor scenes, training data often fails to cover all possible test scenarios, known as the out-of-distribution (OOD) issue. Training-free image editing offers a promising solution for improving model robustness by training data enhancement without any modifications to pre-trained diffusion models. Nevertheless, inversion-based methods often suffer from limited effectiveness and inherent inaccuracies, while recent rectified-flow-based approaches struggle to preserve objects with accurate 3D geometry. In this paper, we propose DriveFlow, a Rectified Flow Adaptation method for training data enhancement in autonomous driving based on pre-trained Text-to-Image flow models. Based on frequency decomposition, DriveFlow introduces two strategies to adapt noise-free editing paths derived from text-conditioned velocities. 1) High-Frequency Foreground Preservation: DriveFlow incorporates a high-frequency alignment loss for foreground to maintain precise 3D object geometry. 2) Dual-Frequency Background Optimization: DriveFlow also conducts dual-frequency optimization for background, balancing editing flexibility and semantic consistency. Comprehensive experiments validate the effectiveness and efficiency of DriveFlow, demonstrating comprehensive performance improvements on all categories across OOD scenarios.

The recent advancements in large language models (LLMs) have significantly enhanced language understanding and content generation capabilities. However, the deployment of LLMs on resource-constrained Internet of Things (IoT) devices remains challenging due to their substantial computational and storage requirements. To address this issue, we propose a novel LLM pruning method, termed structurally-aware adaptive pruning (SAAP), to reduce computational and storage costs for LLMs while maintaining model performance. Specifically, SAAP first leverages maximum likelihood estimation to calibrate traditional structural importance metrics for LLM pruning. Next, it employs a Bayesian fusion approach to address the predictive uncertainty in multi-granularity metrics, enabling accurate assessments of structural importance for LLMs. Then, SAAP introduces a cross-layer importance alignment mechanism based on quantile mapping, which normalizes layer-wise importance scores to ensure consistent pruning from a global perspective. Furthermore, SAAP develops an efficient block-wise fine-tuning strategy for enhancing the performance of the LLM after pruning. To validate the effectiveness of SAAP, we conduct extensive experiments on nine open-source LLMs across two representative tasks—language modeling and zero-shot classification. Experimental results show that SAAP consistently outperforms several baseline methods, achieving accuracy improvements of 2.5%, 2.63%, and 2.44% on LLaMA-7B, Vicuna-7B, and LLaMA-13B when the pruning ratio is 50%. Finally, SAAP is implemented on a testbed—NVIDIA Jetson AGX Orin 32GB Developer Kit. Test results demonstrate that compared to the foundation LLM, SAAP enhances the inference speed by 86.86% at a pruning ratio of 50%, highlighting its potential for practical deployment on resource-constrained IoT devices.

Semantic communication employs deep learning to transmit semantically meaningful information rather than raw data, thereby improving communication efficiency. To facilitate the adaptation of continuous semantic features to digital transmission systems, vector quantization (VQ) serves as an effective approach for discretizing high-dimensional features into compact codebook indices. However, existing VQ-based systems face a critical dilemma: conventional VQ codebooks demand large index ranges to preserve fidelity, contradicting digital modulation’s need for limited discrete states to ensure noise robustness. To bridge this gap, we design a 2K image transmission framework that jointly considers codebook compactness and transmission robustness. The framework operates in two stages: In Stage 1, we devise MOC-RVQ, a multi-head ordered codebook (MOC) with residual vector quantization (RVQ) to reduce the index range while maintaining image fidelity. In Stage 2, a Swin Transformer-based noise reduction block (NRB) is integrated with feature requantization for further robust restoration. Experiments on 2K-resolution datasets demonstrate that the proposed MOC-RVQ surpasses traditional codecs like BPG, JPEG, and learnable baselines, while maintaining low transmission overhead.

Semantic communication (SC) emerges as a breakthrough paradigm for efficient data transmission in next-generation communication networks. However, SC is still in the infant stage with quite a few limitations, such as insufficient semantic representation capacity, high communication latency, and the susceptibility to channel noise. In this paper, we proposeLiFiSC, a cross-modal fusion based generative semantic communication framework with strong semantic compression capacity and high-fidelity semantic restoration. We then extend it toLiFiSCvv, specifically designed to achieve 3D volumetric video transmission and reconstruction with photorealistic visual quality and pixel-level visual consistency, providing end-to-end live watching experience with acceptable latency. We innovatively incorporate unified vision-language encoding into semantic communication, achieving superior semantic understanding and compression.LiFiSCvvcomprises three key components: (1) Information redundancy reduction through lightweight video analysis and structure-from-motion techniques, decreasing reconstruction cost; (2) Cross-modal fusion learning driven codec mechanisms that enable efficient semantic representation and robust transmission against channel impairments; and (3) Feed-forward vision transformer for rapid volumetric video reconstruction and rendering. Comprehensive evaluation results demonstrate thatLiFiSCvvachieves a live high-fidelity and visually consistent video watching experience, with only seconds of end-to-end latency and around 42× semantic compression, significantly outperforming SOTA methods in general image and volumetric video transmission. The project page is available here: https://inml-tygong.github.io/LiFiSCvv/.

In this paper, we propose a novel physics-informed generative learning approach, named RadioDiff-k2, for accurate and efficient multipath-aware radio map (RM) construction. As future wireless communication evolves towards environment-aware paradigms, the accurate construction of RMs becomes crucial yet highly challenging. Conventional electromagnetic (EM)-based methods, such as full-wave solvers and ray-tracing approaches, exhibit substantial computational overhead and limited adaptability to dynamic scenarios. Although existing neural network (NN) approaches have efficient inferencing speed, they lack sufficient consideration of the underlying physics of EM wave propagation, limiting their effectiveness in accurately modeling critical EM singularities induced by complex multipath environments. To address these fundamental limitations, we propose a novel physics-inspired RM construction method guided explicitly by the Helmholtz equation, which inherently governs EM wave propagation. Specifically, based on the analysis of partial differential equations (PDEs), we theoretically establish a direct correspondence between EM singularities, which correspond to the critical spatial features influencing wireless propagation, and regions defined by negative wave numbers in the Helmholtz equation. We then design an innovative dual diffusion model (DM)-based large artificial intelligence framework comprising one DM dedicated to accurately inferring EM singularities and another DM responsible for reconstructing the complete RM using these singularities along with environmental contextual information. Experimental results demonstrate that the proposed RadioDiff-k2framework achieves state-of-the-art (SOTA) performance in both image-level RM construction and localization tasks, while maintaining inference latency within a few hundred milliseconds. Code is available at https://github.com/UNIC-Lab/RadioDiff-k.

End-to-end (E2E) semantic communication (SemCom) powered by semantic knowledge base (SKB) is an efficient SemCom framework. However, in practical scenarios, SKB discrepancy among multiple SemCom pairs arises due to dynamic environmental changes (e.g., varying source data or tasks) or system-level alterations (e.g., integration of new SemCom pairs with divergent SKBs). Such discrepancy leads to performance disparity in semantic transmission, where underperforming pairs fail to maintain efficient task execution. To address this challenge, this paper introduces a cooperative SKB update policy, which enables collaborative evolution of SKBs to mitigate SKB discrepancy and improve performance of underperforming pairs. For each SemCom pair endowed with SKB-enabled SemCom, partial local SKB is selected out and uploaded to a mobile edge computing (MEC) server for establishing a global SKB. The global SKB aggregates the advantages of each local SKB, and is broadcasted to all SemCom pairs. Then, each SemCom pair updates their local SKB with the assistance of the global SKB. This process makes the local SKBs more sensible and less ambiguous, thereby enhancing the semantic transmission performance. Furthermore, in order to maximize the cooperative gains under limited uplink budgets of SemCom pairs, a knowledge selection optimization problem is formulated for the selection of the uploaded knowledge. Numerical results show that the proposed cooperative SKB update policy obtains significant performance gains, especially for the initially poor-performing pairs, and provide comprehensive performance comparison of the knowledge selection scheme.

Reconstructing 3D scenes and synthesizing novel views from sparse input views is a highly challenging task. Recent advances in video diffusion models have demonstrated strong temporal reasoning capabilities, making them a promising tool for enhancing reconstruction quality under sparse-view settings. However, existing approaches are primarily designed for modest viewpoint variations, which struggle in capturing fine-grained details in close-up scenarios since input information is severely limited. In this paper, we present a diffusion-based framework, called CloseUpShot, for close-up novel view synthesis from sparse inputs via point-conditioned video diffusion. Specifically, we observe that pixel-warping conditioning suffers from severe sparsity and background leakage in close-up settings. To address this, we propose hierarchical warping and occlusion-aware noise suppression, enhancing the quality and completeness of the conditioning images for the video diffusion model. Furthermore, we introduce global structure guidance, which leverages a dense fused point cloud to provide consistent geometric context to the diffusion process, to compensate for the lack of globally consistent 3D constraints in sparse conditioning inputs. Extensive experiments on multiple datasets demonstrate that our method outperforms existing approaches, especially in close-up novel view synthesis, clearly validating the effectiveness of our design.

Semantic communications (SemComs) have emerged as a promising paradigm for joint data and task-oriented transmissions, combining the demands for both the bit-accurate delivery and end-to-end (E2E) distortion minimization. However, current joint source-channel coding (JSCC) in SemComs is not compatible with the existing communication systems and cannot adapt to the variations of the sources or the channels, while separate source-channel coding (SSCC) is suboptimal in the finite blocklength regime. To address these issues, we propose an adaptive source-channel coding (ASCC) scheme for SemComs over parallel Gaussian channels, where the deep neural network (DNN)-based semantic source coding and conventional digital channel coding are separately deployed and adaptively designed. To enable efficient adaptation between the source and channel coding, we first approximate the E2E data and semantic distortions as functions of source coding rate and bit error ratio (BER) via logistic regression, where BER is further modeled as functions of signal-to-noise ratio (SNR) and channel coding rate. Then, we formulate the weighted sum E2E distortion minimization problem for joint source-channel coding rate and power allocation over parallel channels, which is solved by the successive convex approximation. Finally, simulation results demonstrate that the proposed ASCC scheme outperforms typical deep JSCC and SSCC schemes for both the single- and parallel-channel scenarios while maintaining full compatibility with practical digital systems.

This paper studies an over-the-air federated edge learning (Air-FEEL) system with integrated sensing, communication, and computation (ISCC), in which one edge server coordinates multiple edge devices to wirelessly sense the objects and use the sensing data to collaboratively train a machine learning model for recognition tasks. In this system, over-the-air computation (AirComp) is employed to enable one-shot model aggregation from edge devices. Under this setup, we analyze the convergence behavior of the ISCC-enabled Air-FEEL in terms of the loss function degradation, by particularly taking into account the wireless sensing noise during the training data acquisition and the AirComp distortions during the over-the-air model aggregation. The result theoretically shows that sensing, communication, and computation compete for network resources to jointly decide the convergence rate. Based on the analysis, we design the ISCC parameters under the target of maximizing the loss function degradation while ensuring the latency and energy budgets in each round. The challenge lies on the tightly coupled processes of sensing, communication, and computation among different devices. To tackle the challenge, we derive a low-complexity ISCC algorithm by alternately optimizing the batch size control and the network resource allocation. It is found that for each device, less sensing power should be consumed if a larger batch of data samples is obtained and vice versa. Besides, with a given batch size, the optimal computation speed of one device is the minimum one that satisfies the latency constraint. Numerical results based on a human motion recognition task verify the theoretical convergence analysis and show that the proposed ISCC algorithm well coordinates the batch size control and resource allocation among sensing, communication, and computation to enhance the learning performance.

Extremely High Throughput (EHT) and Ultra-high reliability (UHR) are new objectives in Next-Gen Wi-Fi, i.e., Wi-Fi 7 and beyond; however, the data rate within a periodic channel sounding round is expected to significantly deteriorate under time-varying channels with Doppler effect in Downlink Multi-User Multiple-Input Multiple-Output. Therefore, Next-Gen channel sounding must carefully balance the MAC-layer CSI overhead reduction and the PHY-layer channel capacity degradation caused by the Doppler effect for data rate maximization. Despite its critical importance, the cross-layer (PHY + MAC) Wi-Fi channel sounding optimization in time-varying channels remains under-explored. This paper addresses this research gap by proposing a cross-layer optimization problem to find the optimal EHT sounding period that maximizes the average data rate by considering both MAC-layer CSI overhead and PHY-layer channel capacity degradation. This problem is then converted into an equivalent optimization problem that can be solved efficiently using our proposed model driven optimal search algorithm with proven convexity. Afterwards, we introduce a data driven Transformer-based partial CSI prediction framework to alleviate CSI staleness without introducing extra CSI overhead, which further enhances the average data rate. Through simulations, we evaluate the baseline EHT sounding protocol that always uses outdated partial CSI, and then benchmark the baseline against our proposed hybrid data and model driven approach. The numerical results demonstrate that integrating Transformer-based partial CSI prediction with the optimal channel sounding period significantly reduces CSI overhead by up to 25.2%, while increasing the average throughput by up to 30.9%.

This paper addresses the joint transceiver design, including pilot transmission, channel feature extraction and feedback, as well as precoding, for low-overhead downlink massive multiple-input multiple-output (MIMO) communication in frequency-division duplex (FDD) systems. Although deep learning (DL) has shown great potential in tackling this problem, existing methods often suffer from poor scalability in practical systems, as the solution obtained in the training phase merely works for a fixed feedback capacity and a fixed number of users in the deployment phase. To address this limitation, we propose a novel DL-based framework comprised of choreographed neural networks, which can utilize one training phase to generate all the transceiver solutions used in the deployment phase with varying sizes of feedback codebooks and numbers of users. The proposed framework includes a residual vector-quantized variational autoencoder (RVQ-VAE) for efficient channel feedback and an edge graph attention network (EGAT) for robust multi-user precoding. It can adapt to different feedback capacities by flexibly adjusting the RVQ codebook sizes using the hierarchical codebook structure, and scale with the number of users through a feedback module sharing scheme and the inherent scalability of EGAT. Moreover, a progressive training strategy is proposed to further enhance data transmission performance and generalization capability. Numerical results on a real-world dataset demonstrate the superior scalability and performance of our approach over existing methods.

This paper addresses the problem of inverse rendering from photometric images. Existing approaches for this problem suffer from the effects of self-shadows, interreflections, and lack of constraints on the surface reflectance, leading to inaccurate decomposition of reflectance and illumination due to the ill-posed nature of inverse rendering. In this work, we propose a new method for neural inverse rendering. Our method jointly optimizes the light source position to account for the self-shadows in images, and computes indirect illumination using a differentiable rendering layer and an importance sampling strategy. To enhance surface reflectance decomposition, we introduce a new regularization by distilling DINO features to foster accurate and consistent material decomposition. Extensive experiments on synthetic and real datasets demonstrate that our method outperforms the state-of-the-art methods in reflectance decomposition.

Precise vessel segmentation is vital for clinical applications such as diagnosis and surgical planning but remains challenging due to thin, branching geometries and low texture contrast. Although foundation models such as the Segment Anything Model (SAM) show strong performance in general segmentation tasks, they remain suboptimal for vascular structures. In this work, we present VesSAM, a powerful and efficient framework tailored for 2D vessel segmentation. VesSAM integrates three core modules: a convolutional adapter that enhances local texture features, a multi-prompt encoder that fuses anatomical cues via hierarchical cross-attention, and a lightweight mask decoder that reduces jagged artifacts. We also introduce an automated pipeline to generate structured multi-prompt annotations, and curate a diverse benchmark dataset spanning 8 datasets across 5 imaging modalities. Extensive experiments show that VesSAM surpasses state-of-the-art PEFT-based SAM variants by over$\text{1 0 \%}$Dice and 13% IoU, while maintaining competitive accuracy to fully fine-tuned methods with far fewer parameters. VesSAM also generalizes well to out-of-distribution (OoD) settings, outperforming all baselines in average OoD Dice and IoU.

This paper introduces VisionPAD, a novel self-supervised pre-training paradigm designed for vision-centric algorithms in autonomous driving. In contrast to previous approaches that employ neural rendering with explicit depth supervision, VisionPAD utilizes more efficient 3D Gaussian Splatting to reconstruct multi-view representations using only images as supervision. Specifically, we introduce a self-supervised method for voxel velocity estimation. By warping voxels to adjacent frames and supervising the rendered outputs, the model effectively learns motion cues in the sequential data. Furthermore, we adopt a multi-frame photometric consistency approach to enhance geometric perception. It projects adjacent frames to the current frame based on rendered depths and relative poses, boosting the 3D geometric representation through pure image supervision. Extensive experiments on autonomous driving datasets demonstrate that VisionPAD significantly improves performance in 3D object detection, occupancy prediction and map segmentation, surpassing state-of-the-art pre-training strategies by a considerable margin.

In autonomous driving, vision-centric 3D detection aims to identify 3D objects from images. However, high data collection costs and diverse real-world scenarios limit the scale of training data. Once distribution shifts occur between training and test data, existing methods often suffer from performance degradation, known as Out-of-Distribution (OOD) problems. To address this, controllable Text-to-Image (T2I) diffusion offers a potential solution for training data enhancement, which is required to generate diverse OOD scenarios with precise 3D object geometry. Nevertheless, existing controllable T2I approaches are restricted by the limited scale of training data or struggle to preserve all annotated 3D objects. In this paper, we present DriveGEN, a method designed to improve the robustness of 3D detectors in Driving via Training-Free Controllable Text-to-Image Diffusion Generation. Without extra diffusion model training, DriveGEN consistently preserves objects with precise 3D geometry across diverse OOD generations, consisting of 2 stages: 1) Self-Prototype Extraction: We empirically find that self-attention features are semantic-aware but require accurate region selection for 3D objects. Thus, we extract precise object features via layouts to capture 3D object geometry, termed self-prototypes. 2) Prototype-Guided Diffusion: To preserve objects across various OOD scenarios, we perform semantic-aware feature alignment and shallow feature alignment during denoising. Extensive experiments demonstrate our effectiveness in improving 3D detection. The code is available at github.com/Hongbin98/DriveGEN.

Driven by the great success of Large Language Models (LLMs) in the 2D image domain, their application in 3D scene understanding has emerged as a new trend. A key difference between 3D and 2D is that the situation of an egocentric observer in 3D scenes can change, resulting in different descriptions (e.g., "left" or "right"). However, current LLM-based methods overlook the egocentric perspective and use datasets from a global viewpoint. To address this issue, we propose a novel approach to automatically generate a situation-aware dataset by leveraging the scanning trajectory during data collection and utilizing Vision-Language Models (VLMs) to produce high-quality captions and question-answer pairs. Furthermore, we introduce a situation grounding module to explicitly predict the position and orientation of the observer’s viewpoint, thereby enabling LLMs to ground situation descriptions in 3D scenes. We evaluate our approach on several benchmarks, demonstrating that our method effectively enhances the 3D situational awareness of LLMs while significantly expanding existing datasets and reducing manual effort.

This paper proposes an adaptive source-channel coding (ASCC) scheme for point-to-point digital semantic communications over parallel Gaussian channels, where the deep neural network (DNN)-based semantic source coding and conventional digital channel coding are separately deployed and adaptively designed. To enable efficient adaptation between the source and channel coding, we first approximate the E2E data and semantic distortions as functions of source coding rate and bit error ratio (BER) via logistic regression, where BER is further modeled as functions of signal-to-noise ratio (SNR) and channel coding rate. Then, we formulate the weighted sum E2E distortion minimization problem for joint source-channel coding rate and power allocation over parallel channels, which is solved by the successive convex approximation. Finally, simulation results demonstrate that the proposed ASCC scheme outperforms typical separate and deep joint source-channel coding schemes while maintaining full compatibility with practical digital systems.

A Radio Environment Map (REM) is pivotal for optimizing wireless communication systems, yet its accuracy is inherently tied to the complex interplay of electromagnetic propagation and landform heterogeneity. Existing REM construction methods often implicitly ignore land features, leading to inaccuracies in REM estimation. In this paper, we introduce a novel Radio Oriented Heterogeneous Multitask Gaussian Process (RO-HMTGP) to address this gap by jointly integrating heterogeneous inputs, including land features and Reference Signal Receiving Power (RSRP). The proposed RO-HMTGP model treats land features as auxiliary knowledge and captures spatially correlated propagation effects across varying landforms while preserving feature-specific attenuation characteristics. RO-HMTGP enhances the construction accuracy of REM while quantifying uncertainty. Empirical evaluations using real-world datasets demonstrate that the land feature-aware RO-HMTGP achieves superior predictive performance compared to existing methods without land feature awareness. RO-HMTGP advances the integration of geospatial analytics into wireless communications, providing a pathway toward land feature-aware cognitive radio systems.

Radio environment mapping in urban scenarios presents significant challenges due to the complex interplay of multipath propagation, shadowing effects, and heterogeneous urban morphology. We present RadioVAE, a novel probabilistic framework that generates high-fidelity radio maps with uncertainty quantification through a Variational Autoencoder (VAE). The model addresses the critical limitations of deterministic neural networks by learning a structured latent space that explicitly encodes the posterior distribution of propagation features conditioned on urban topology and transmitter characteristics. A UNet-based feature extractor enables simultaneous modeling of both fine-scale spatial variations and city-wide propagation patterns, while KL-divergence regularization ensures physically plausible predictions. The generative process incorporates urban morphological constraints and transmitter parameters to accurately reconstruct complex propagation fields. Experimental results demonstrate that RadioVAE significantly outperforms existing deterministic methods in both prediction accuracy and uncertainty calibration, particularly in challenging urban environments with complex spatial correlations.

Score-based diffusion models represent a significant variant within the family of diffusion models and have found extensive application in the increasingly popular domain of generative tasks. Recent investigations have explored the denoising potential of diffusion models in semantic communications. However, in previous paradigms, noise distortion in the diffusion process does not match precisely with digital channel noise characteristics. In this work, we introduce the ScoreBased Channel Denoising Model (SCDM) for Digital Semantic Communications (DSC). SCDM views the distortion of constellation symbol sequences in digital transmission as a score-based forward diffusion process. We design a tailored forward noise corruption to better align digital channel noise properties in the training phase. During the inference stage, the well-trained SCDM can effectively denoise received semantic symbols under various SNR conditions, reducing the difficulty for the semantic decoder in extracting semantic information from the received noisy symbols and thereby enhancing the robustness of the reconstructed semantic information. Experimental results show that SCDM outperforms the baseline model in PSNR, SSIM, and MSE metrics, particularly at low SNR levels. Moreover, SCDM reduces storage requirements by a factor of 7.8. This efficiency in storage, combined with its robust denoising capability, makes SCDM a practical solution for DSC across diverse channel conditions.

This paper considers the joint pilot, feedback, and precoding design for low-overhead downlink multi-user multiple-input multiple-output (MU-MIMO) communication in frequency-division duplex (FDD) systems. Although the deep learning (DL) technique has demonstrated the potential to address this challenging problem, most of the current works along this line suffer from poor scalability because the numbers of users during the training and implementation phases have to be the same. To overcome this limitation, we propose a novel scalable DL framework for joint transceiver design in FDD systems using neural networks (NNs). Specifically, we design a vector-quantized variational autoencoder-based feedback scheme with a common quantization codebook shared by all the users for efficient feedback. Additionally, we employ a graph attention network (GAT) for effective precoding design. Thanks to the shared feedback module and the inherent scalability of GAT, the proposed NN trained for a particular number of users can be applied in practice given any number of users with satisfactory performance. Numerical results using a real-world channel dataset are provided to demonstrate the superior scalability and performance of the proposed approach over the existing baseline methods.

We propose a continuous gathering scheme based on the swept volume to address the challenges involved in planning a tethered robot duo to efficiently collect marine debris. Specifically, we model the tethered robot duo by constructing a double-layer U-shape, and then apply an object-aware optimization approach that leverages the swept volume signed distance field (SVSDF) to guide trajectory optimization, promoting complete object collection while maintaining a continuous and collision-free gathering motion. Existing algorithms either fail to fully address key challenges, such as assuming an unrealistically infinite tether length or incurring high computational costs. In contrast, our proposed method, by adopting the double-layer U-shape technique, effectively manages tether length constraints and preserves the tether shape, ensuring feasible collection. By utilizing the SVSDF technique to guide the trajectory optimization process, we maximize the swept coverage of objects while minimizing that of obstacles. This enables complete object coverage, avoids collisions, and prevents the tether from becoming trapped by obstacles during the collection process. Moreover, we propose a set of metrics for this gathering planning problem and validate the generated trajectories in simulation, using a collision-free multi-UAV information-gathering approach to efficiently estimate the target area. Simulations demonstrate that our proposed method achieves superior, resolution-independent gathering performance compared to existing algorithms.

Large Language Models (LLMs) exhibit remarkable capabilities in the hierarchical decomposition of complex tasks through semantic reasoning. However, their application in embodied systems faces challenges in ensuring reliable execution of subtask sequences and achieving one-shot success in long-term task completion. To address these limitations in dynamic environments, we propose Closed-Loop Embodied Agent (CLEA)—a novel architecture incorporating four specialized open-source LLMs with functional decoupling for closed-loop task management. The framework features two core innovations: (1) Interactive task planner that dynamically generates executable subtasks based on the environmental memory, and (2) Multimodal execution critic employing an evaluation framework to conduct a probabilistic assessment of action feasibility, triggering hierarchical re-planning mechanisms when environmental perturbations exceed preset thresholds. To validate CLEA’s effectiveness, we conduct experiments in a real environment with manipulable objects, using two heterogeneous robots for object search, manipulation, and search-manipulation integration tasks. Across 12 task trials, CLEA outperforms the baseline model, achieving a 67.3% improvement in success rate and a 52.8% increase in task completion rate. These results demonstrate that CLEA significantly enhances the robustness of task planning and execution in dynamic environments. Our code is available at https://sp4595.github.io/CLEA/.

Sparse Perception Models (SPMs) adopt a query-driven paradigm that forgoes explicit dense BEV or volumetric construction, enabling highly efficient computation and accelerated inference. In this paper, we introduce SQS, a novel query-based splatting pre-training specifically designed to advance SPMs in autonomous driving. SQS introduces a plug-in module that predicts 3D Gaussian representations from sparse queries during pre-training, leveraging self-supervised splatting to learn fine-grained contextual features through the reconstruction of multi-view images and depth maps. During fine-tuning, the pre-trained Gaussian queries are seamlessly integrated into downstream networks via query interaction mechanisms that explicitly connect pre-trained queries with task-specific queries, effectively accommodating the diverse requirements of occupancy prediction and 3D object detection. Extensive experiments on autonomous driving benchmarks demonstrate that SQS delivers considerable performance gains across multiple query-based 3D perception tasks, notably in occupancy prediction and 3D object detection, outperforming prior state-of-the-art pre-training approaches by a significant margin (i.e., +1.3 mIoU on occupancy prediction and +1.0 NDS on 3D detection).

With the proliferation of edge computing, efficient artificial intelligence inference on edge devices has become essential for intelligent applications, such as autonomous vehicles and virtual/augmented reality. In this context, we address the problem of efficient remote object recognition by optimizing feature transmission between mobile devices and edge servers. We propose an optimization framework to tackle the challenges posed by dynamic channel conditions and device mobility in end-to-end communication systems. Our approach builds upon existing methods by leveraging a semantic knowledge base to drive multilevel feature transmission, accounting for temporal factors, state transitions, and dynamic elements throughout the transmission process. Additionally, we enhance the multilevel feature transmission policy by introducing an additional fifth-level edge-assisted semantic communication, which maximizes recognition performance by leveraging a large semantic knowledge base on the edge server. Formulated as an online optimization problem, our framework aims to simultaneously minimize semantic loss and adhere to specified transmission latency thresholds. To achieve this, we design a soft actor-critic-based deep reinforcement learning system with a carefully designed reward structure for real-time decision making. This approach overcomes the optimization difficulty of the NP-hard problem while fulfilling the optimization objectives. Numerical results showcase the superiority of our approach compared to traditional greedy methods across various system setups using open-source datasets.

Differing from the conventional communication system paradigm that models information source as a sequence of (i.i.d. or stationary) random variables, the semantic approach aims at extracting and sending the high-level features of the content deeply contained in the source, thereby breaking the performance limits from the statistical information theory. As a pioneering work in this area, the deep learning-enabled semantic communication (DeepSC) constitutes a novel algorithmic framework based on the transformer—which is a deep learning tool widely used to process text numerically. The main goal of this work is to extend the DeepSC approach from the point-to-point link to the multiuser multiple access channel (MAC). The interuser interference has long been identified as the bottleneck of the MAC. In the classic information theory, the successive interference cancellation (SIC) scheme is a common way to mitigate interference and achieve the channel capacity. Our main contribution is to incorporate the SIC scheme into the DeepSC. As opposed to the traditional SIC that removes interference in the digital symbol domain, the proposed semantic SIC works in the domain of the semantic word embedding vectors. Furthermore, to enhance the training efficiency, we propose a pretraining scheme and a partial retraining scheme that quickly adjust the neural network parameters when new users are added to the MAC. We also modify the existing loss function to facilitate training. Finally, we present numerical experiments to demonstrate the advantage of the proposed semantic approach as compared to the existing benchmark methods.

Federated learning (FL) has become a transformative paradigm for distributed machine learning over wireless networks. However, the performance of FL is hindered by the unreliable communication links between resource-constrained Internet of Things (IoT) devices and the central server. To overcome this challenge, we propose a novel framework that employs an unmanned aerial vehicle (UAV) as a mobile server to enhance the FL training process. By capitalizing on the UAV’s mobility, we establish strong line-of-sight connections with IoT devices, thereby enhancing communication reliability and capacity. To maximize training efficiency, we formulate a latency minimization problem that jointly optimizes bandwidth allocation, computing resources, transmit power for both the UAV and IoT devices, and the flight trajectory of the UAV. Subsequently, we analyze the required rounds of the IoT devices training and the UAV aggregation for FL convergence. Based on the convergence constraint, we transform the problem into three subproblems and develop an efficient alternating optimization algorithm to solve this problem. Additionally, we provide a thorough analysis of the algorithm’s convergence and computational complexity. Extensive numerical results demonstrate that the proposed algorithm-based scheme not only surpasses existing benchmark schemes in reducing latency up to 15.29%, but also achieves training efficiency that nearly matches the ideal scenario.

Future sixth-generation (6G) systems are expected to leverage extremely large-scale multiple-input multiple-output (XL-MIMO) technology, which significantly expands the range of the near-field region. The spherical wavefront characteristics in the near field introduce additional degrees of freedom (DoFs), namely distance and angle, into the channel model, which leads to unique challenges in channel estimation (CE). In this paper, we propose a new sensing-enhanced uplink CE scheme for near-field XL-MIMO, which notably reduces the required quantity of baseband samples and the dictionary size. In particular, we first propose a sensing method that can be accomplished in a single time slot. It employs power sensors embedded within the antenna elements to measure the received power pattern rather than baseband samples. A time inversion algorithm is then proposed to precisely estimate the locations of users and scatterers, which offers a substantially lower computational complexity. Based on the estimated locations from sensing, a novel dictionary is then proposed by considering the eigen-problem based on the near-field transmission model, which facilitates efficient near-field CE with less baseband sampling and a more lightweight dictionary. Moreover, we derive the general form of the eigenvectors associated with the near-field channel matrix, revealing their noteworthy connection to the discrete prolate spheroidal sequence (DPSS). Simulation results unveil that the proposed time inversion algorithm achieves accurate localization with power measurements only, and remarkably outperforms various widely-adopted algorithms in terms of computational complexity. Furthermore, the proposed eigen-dictionary considerably improves the accuracy in CE with a compact dictionary size and a drastic reduction in baseband samples by up to 66%.

Task-oriented integrated sensing, communication, and computation (ISCC) is a key technology for achieving low-latency edge inference and enabling efficient implementation of artificial intelligence (AI) in industrial cyber-physical systems (ICPS). However, the constrained energy supply at edge devices has emerged as a critical bottleneck. In this paper, we propose a novel energy-efficient ISCC framework for AI inference at resource-constrained edge devices, where adjustable split inference, model pruning, and feature quantization are jointly designed to adapt to diverse task requirements. A joint resource allocation design problem for the proposed ISCC framework is formulated to minimize the energy consumption under stringent inference accuracy and latency constraints. To address the challenge of characterizing inference accuracy, we derive an explicit approximation for it by analyzing the impact of sensing, communication, and computation processes on the inference performance. Building upon the analytical results, we propose an iterative algorithm employing alternating optimization to solve the resource allocation problem. In each subproblem, the optimal solutions are available by respectively applying a golden section search method and checking the Karush-Kuhn-Tucker (KKT) conditions, thereby ensuring the convergence to a local optimum of the original problem. Numerical results demonstrate the effectiveness of the proposed ISCC design, showing a significant reduction in energy consumption of up to 40% compared to existing methods, particularly in low-latency scenarios.

Distributed Artificial Intelligence-Generated Content (AIGC) has attracted significant attention, but two key challenges remain: maximizing subjective Quality of Experience (QoE) and improving energy efficiency, which are particularly pronounced in widely adopted Generative Diffusion Model (GDM)-based image generation services. In this paper, we propose a novel user-centric Interactive AI (IAI) approach for service management, with a distributed GDM-based AIGC framework that emphasizes efficient and cooperative deployment. The proposed method restructures the GDM inference process by allowing users with semantically similar prompts to share parts of the denoising chain. Furthermore, to maximize the users' subjective QoE, we propose an IAI approach, i.e., Reinforcement Learning With Large Language Models Interaction (RLLI), which utilizes Large Language Model (LLM)-empowered generative agents to replicate users interactions, providing real-time and subjective QoE feedback aligned with diverse user personalities. Lastly, we present the GDM-based Deep Deterministic Policy Gradient (G-DDPG) algorithm, adapted to the proposed RLLI framework, to allocate communication and computing resources effectively while accounting for subjective user traits and dynamic wireless conditions. Simulation results demonstrate that G-DDPG improves total QoE by 15% compared with the standard DDPG algorithm.

Artificial intelligence (AI) is envisioned to be natively integrated into the sixth-generation (6G) mobile networks to support a diverse range of intelligent applications. Federated edge learning (FEEL) emerges as a vital enabler of this vision by leveraging the sensing, communication, and computation capabilities of geographically dispersed edge devices to collaboratively train AI models without sharing raw data. This article explores the pivotal role of FEEL in advancing both the “wireless for AI” and “AI for wireless” paradigms, thereby facilitating the realization of scalable, adaptive, and intelligent 6G networks. We begin with a comprehensive overview of learning architectures, models, and algorithms that form the foundations of FEEL. We, then, establish a novel task-oriented communication principle to examine key methodologies for deploying FEEL in dynamic and resource-constrained wireless environments, focusing on device scheduling, model compression, model aggregation, and resource allocation. Furthermore, we investigate the domain-specific optimizations of FEEL to facilitate its promising applications, ranging from wireless air-interface technologies to mobile and the Internet of Things (IoT) services. Finally, we highlight key future research directions for enhancing the design and impact of FEEL in 6G.

This paper investigates the energy-efficient hybrid beamforming design for a multi-functional integrated sensing, communications, and powering (ISCAP) system. In this system, a base station (BS) with a hybrid analog-digital (HAD) architecture sends unified wireless signals to communicate with multiple information receivers (IRs), sense multiple point targets, and wirelessly charge multiple energy receivers (ERs) at the same time. To facilitate the energy-efficient design, we present a novel HAD architecture for the BS transmitter, which allows dynamic on-off control of its radio frequency (RF) chains and analog phase shifters (PSs) through a switch network. We also consider a practical and comprehensive power consumption model for the BS, by taking into account the power-dependent non-linear power amplifier (PA) efficiency, and the on-off non-transmission power consumption model of RF chains and PSs. We jointly design the hybrid beamforming and dynamic on-off control at the BS, aiming to minimize its total power consumption, while guaranteeing the performance requirements on communication rates, sensing Cramér-Rao bound (CRB), and harvested power levels. The formulation also takes into consideration the per-antenna transmit power constraint and the constant modulus constraints for the analog beamformer at the BS. The resulting optimization problem for ISCAP is highly non-convex due to the binary on-off non-transmission power consumption of RF chains and PSs, the non-linear PA efficiency, and the coupling between analog and digital beamformers. To tackle this problem, we first approximate the binary on-off non-transmission power consumption into a continuous form, and accordingly propose an iterative algorithm to find a high-quality approximate solution with ensured convergence, by employing techniques from alternating optimization (AO), sequential convex approximation (SCA), and semi-definite relaxation (SDR). Then, based on the optimized beamforming weights, we develop an efficient method to determine the binary on-off control of RF chains and PSs, as well as the associated hybrid beamforming solution. Numerical results show that the proposed design achieves an improved energy efficiency for ISCAP than other benchmark schemes without joint design of hybrid beamforming and dynamic on-off control. This validates the benefit of dynamic on-off control in energy reduction, especially when the multi-functional performance requirements become less stringent.

Semantic communication has drawn substantial attention as a promising paradigm to achieve effective and intelligent communications. However, efficient image semantic communication encounters challenges with a lower testing compression ratio (CR) and signal-to-noise ratio (SNR) compared to the training phase. To tackle this issue, we propose an innovative semantic knowledge base (SKB)-enabled generative semantic communication system for image classification task and image generation task. Specifically, a lightweight SKB, comprising class-level information, is exploited to guide the semantic communication process, which enables us to transmit only the relevant indices. This approach promotes the completion of the image classification task at the transmitter and significantly reduces the transmission load. Meanwhile, the class-level knowledge in the SKB facilitates the image generation task by allowing controllable generation, making it possible to generate class-consistent images in resource-constrained and low SNR scenarios. Furthermore, an adaptive CR and mode selection mechanism is designed to automatically adjust the CR and task mode, which allows the proposed system accommodate various CR and SNR conditions. Evaluation results indicate that the proposed method outperforms the benchmarks and achieves superior performance with minimal CR and SNR.

Recent progress in 3-D scene understanding has explored visual grounding [3D visual grounding (3DVG)] to localize a target object through a language description. However, existing methods only consider the dependency between the entire sentence and the target object, ignoring fine-grained relationships between contexts and nontarget ones. In this article, we extend 3DVG to a more fine-grained task, called 3D phrase-aware grounding (3DPAG). The 3DPAG task aims to localize the target objects in a 3-D scene by explicitly identifying all phrase-related objects and then conducting the reasoning according to contextual phrases. To tackle this problem, we manually labeled about 227 K phrase-level annotations using a self-developed platform, from 88 K sentences of widely used 3DVG datasets, i.e., Natural Reference in 3-D (Nr3D), Spatial Reference in 3-D (Sr3D), and ScanRefer. By tapping on our datasets, we can extend previous 3DVG methods to the fine-grained phrase-aware scenario. It is achieved through the proposed novel phrase-object alignment (POA) optimization and phrase-specific pretraining (PSP), boosting conventional 3DVG performance as well. Extensive results confirm significant improvements, i.e., previous state-of-the-art method achieves 3.9%, 3.5%, and 4.6% overall accuracy gains on Nr3D, Sr3D, and ScanRefer, respectively. Our datasets and platform are released in https://github.com/CurryYuan/PhraseRefer.

The widespread adoption of real-time services on the Internet has aroused interest in the study of low-latency and deterministic communications. Deterministic guarantee over wide-area networks (WANs), the primary infrastructure for communications, is essential to achieving end-to-end deterministic transmission. However, applying off-the-shelf deterministic schemes to WANs is challenging due to the statistical multiplexing nature of WANs and the non-periodic nature of WAN traffic. In this paper, we propose a novel deterministic framework for WANs, named DetWAN, which guarantees the timely delivery of WAN traffic via deadline-aware routing and scheduling. We design a coordinated earliest deadline first (CEDF) scheduling scheme in the data plane of the DetWAN, which provides determinism for non-periodic deadline-constrained traffic while following statistical multiplexing. To precisely estimate the capacity of deadline-constrained traffic that the DetWAN can satisfy, we derive an end-to-end deadline satisfiability criterion in the DetWAN by introducing the deadline curve into traffic modeling. Based on the criterion, we formulate the deadline-aware routing and scheduling problem as a stochastic optimization problem to maximize the timely delivery ratio. Furthermore, we propose a distributed admission control algorithm based on multi-agent deep reinforcement learning in the control plane to solve the problem in a highly autonomous manner. The algorithm can jointly determine optimal routes and per-hop deadline budgets for traffic flows in a decentralized mode. Extensive evaluation results validate the deterministic guarantee as well as the high throughput of the DetWAN and show that the proposed admission control algorithm can significantly improve the timely delivery ratio compared with benchmarks in WAN scenarios.

Although existing speech-driven talking face generation methods achieve significant progress, they are far from real-world application due to the avatar-specific training demand and unstable lip movements. To address the above issues, we propose the GSmoothFace, a novel two-stage generalized talking face generation model guided by a fine-grained 3D face model, which can synthesize smooth lip dynamics while preserving the speaker's identity. Our proposed GSmoothFace model mainly consists of the Audio to Expression Prediction (A2EP) module and the Target Adaptive Face Translation (TAFT) module. Specifically, we first develop the A2EP module to predict expression parameters synchronized with the driven speech. It uses a transformer to capture the long-term audio context and learns the parameters from the fine-grained 3D facial vertices, resulting in accurate and smooth lip-synchronization performance. Afterward, the well-designed TAFT module, empowered by Morphology Augmented Face Blending (MAFB), takes the predicted expression parameters and target video as inputs to modify the facial region of the target video without distorting the background content. The TAFT effectively exploits the identity appearance and background context in the target video, which makes it possible to generalize to different speakers without retraining. Both quantitative and qualitative experiments confirm the superiority of our method in terms of realism, lip-synchronization, and visual quality.

This paper considers a multi-functional orthogonal frequency division multiplexing (OFDM) system with integrated sensing, communication, and powering (ISCAP), in which a multi-antenna base station (BS) transmits OFDM signals to simultaneously deliver information to multiple information receivers (IRs), provide energy supply to multiple energy receivers (ERs), and sense potential targets based on the echo signals. To facilitate ISCAP, the BS employs the joint transmit beamforming design by sending dedicated sensing/energy beams jointly with information beams. Furthermore, we consider the beam scanning for sensing, in which the joint beams scan in different directions over time to sense potential targets. In order to ensure the sensing beam scanning performance and meet the communication and powering requirements, it is essential to properly schedule IRs and ERs and design the resource allocation over time, frequency, and space. More specifically, we optimize the joint transmit beamforming over multiple OFDM symbols and subcarriers, with the objective of minimizing the average beampattern matching error of beam scanning for sensing, subject to the constraints on the average communication rates at IRs and the average harvested power at ERs. We find converged high-quality solutions to the formulated problem by proposing efficient iterative algorithms based on advanced optimization techniques. We also develop various heuristic designs based on the principles of zero-forcing (ZF) beamforming, round-robin user scheduling, and time switching, respectively. Numerical results show that our proposed algorithms adaptively generate information and sensing/energy beams at each time-frequency slot to match the scheduled IRs/ERs with the desired scanning beam, significantly outperforming the heuristic designs.

Hybrid analog and digital beamforming (HBF) is a cost-efficient technique to achieve high data rates in millimeterwave (mmWave) communication systems. This paper applies the emerging graph neural networks (GNNs) to HBF by leveraging the topological information in wireless networks for better adaptation to dynamic environments. To address the issue of limited feature extraction capability of the existing single-type GNNs, such as node-GNN or edge-GNN, we model the mmWave communication systems as both homogeneous and heterogeneous graphs, and separate the HBF design into node- and edge-level subtasks. Then, the two graphs are presented by two novel models based on homogeneous graph attention network (GAT) and heterogeneous GAT (HGAT), respectively, and mapped to the desired power allocation, radio frequency precoder and baseband precoder. Both the proposed GAT and HGAT are generalizable in the user variation scenarios, while the HGAT is also generalizable in antenna variation scenarios through the incorporation of a complex embedding layer. Furthermore, we introduce a constraint adaptive layer in the GAT and HGAT to ensure feasible outputs. Extensive numerical results based on the public dataset DeepMIMO are provided to evaluate the GAT and HGAT. The proposed approaches generally outperform existing baselines in terms of adaptability to system settings and generalizability to (unseen) problem parameters/sizes, while the HGAT can even achieve faster and better inference than traditional optimization approaches.

Communication and energy efficiencies are two crucial objectives in the pursuit of edge intelligence in 6G networks, and become increasingly important given the prevalence of large model training. Existing designs typically focus on either communication efficiency or energy efficiency due to the fact that improving one objective generally comes at the expense of the other. Over-the-air federated learning (OTA-FL) has recently emerged as a promising approach to enhance both efficiencies through an integrated communication and computation design. Nevertheless, most previous studies on OTA-FL only consider scenarios where the dataset for the entire FL procedure is collected and available prior to training. In real-world applications, devices continuously collect new data in an online manner. This underscores the significance of sample collection through sensing in a practical FL pipeline. We propose to integrate sensing with communication and computation into a joint design to further boost the communication-and-energy efficiencies of OTA-FL. Specifically, we consider a training latency and energy consumption minimization problem with performance guarantees. To this end, we first derive an average training error (ATE) metric to quantify convergence performance. Then, a joint sensing, communication and computation resource allocation strategy is developed based on a deep reinforcement learning (DRL) algorithm that nests convex optimization with a deep Q-network. Extensive experiments are conducted to validate our theoretical analysis, and demonstrate the effectiveness of the proposed design for communication-and-energy efficient FL.

In some applications, edge learning is experiencing a shift in focus from conventional learning from scratch to two-stage learning combining pre-training and task-specific fine-tuning. This paper considers the problem of joint communication and computation resource management in a two-stage edge learning system. In this system, model pre-training is first conducted at an edge server via centralized learning on local pre-stored general data, and then task-specific fine-tuning is performed at edge devices based on the pre-trained model via federated edge learning. For the two-stage learning model, we first analyze the convergence behavior (in terms of the average squared gradient norm bound), which characterizes the impacts of various system parameters, such as the number of learning rounds and batch sizes in the two stages, on the convergence rate. Based on our analytical results, we then propose a joint communication and computation resource management design to minimize an average squared gradient norm bound, subject to constraints on the transmit power, overall system energy consumption, and training delay. The decision variables include the number of learning rounds, batch sizes, clock frequencies, and transmit power control for both pre-training and fine-tuning stages. Finally, numerical results are provided to evaluate the effectiveness of our proposed design. It is shown that the proposed joint resource management over the pre-training and fine-tuning stages well balances the system performance trade-off among the training accuracy, delay, and energy consumption. The proposed design is also shown to effectively leverage the inherent trade-off between pre-training and fine-tuning, which arises from the differences in data distribution between pre-stored general data versus real-time task-specific data, thus efficiently optimizing overall system performance.

Federated edge learning (FEEL) enables privacy-preserving model training through periodic communication between edge devices and the server. Unmanned Aerial Vehicle (UAV)-mounted edge devices are particularly advantageous for FEEL due to their flexibility and mobility in efficient data collection. In UAV-assisted FEEL, sensing, computation, and communication are coupled and compete for limited onboard resources, and UAV deployment also affects sensing and communication performance. Therefore, the joint design of UAV deployment and resource allocation is crucial to achieving the optimal training performance. In this paper, we address the problem of joint UAV deployment design and resource allocation for FEEL via a concrete case study of human motion recognition based on wireless sensing. We first analyze the impact of UAV deployment on the sensing quality and identify a threshold value for the sensing elevation angle that guarantees a satisfactory quality of data samples. Due to the non-ideal sensing channels, we consider the probabilistic sensing model, where the successful sensing probability of each UAV is determined by its position. Then, we derive the upper bound of the FEEL training loss as a function of the sensing probability. Theoretical results suggest that the convergence rate can be improved if UAVs have a uniform successful sensing probability. Based on this analysis, we formulate a training time minimization problem by jointly optimizing UAV deployment, integrated sensing, computation, and communication (ISCC) resources under a desirable optimality gap constraint. To solve this challenging mixed-integer non-convex problem, we apply the alternating optimization technique, and propose the bandwidth, batch size, and position optimization (BBPO) scheme to optimize these three decision variables alternately. Simulation results demonstrate that our BBPO scheme outperforms other baseline schemes regarding convergence rate and testing accuracy. The simulation implementation is available at https://github.com/TheaSherlock/ISCC-UAV.

This paper studies the unmanned-aerial-vehicle (UAV)-enabled data collection for Internet-of-things (IoT) networks, in which multiple UAVs are dispatched to collect data over their correspondingly designated areas. In particular, we consider that the UAVs need to collect data in a timely manner. We further consider a practical segmented channel model, in which the air-to-ground wireless channel is assumed to follow Rayleigh or Rician fading when the corresponding line-of-sight (LoS) link is blocked or unblocked, respectively. Under this setup, we minimize the average Age-of-Information (AoI) for the IoT devices, by jointly optimizing the UAV trajectory, the collection scheduling, and the completion time. Since the problem is non-convex, and the dimension of optimization variables varies w.r.t. the completion time that needs to be optimized, the conventional methods are not applicable for efficiently solving the problem. To address this issue, we first propose a deep-reinforcement-learning (DRL) based algorithm to solve this problem in the special case with single UAV, and then exploit federated learning over multiple UAVs to efficiently train the model in a collaborative manner while preserving the data privacy. Numerical results verify that the proposed methods achieve significantly better performance than benchmarks in terms of the AoI, energy consumption, and completion time.

Point cloud salient object detection (PCSOD) is a newly proposed task in 3D dense segmentation. However, the acquisition of accurate 3D dense annotations comes at a high cost, severely limiting the progress of PCSOD. To address this issue, we propose the first weakly supervised PCSOD (named WeakPCSOD) model, which relies solely on cheap 3D bounding box annotations. In WeakPCSOD, we extract noise-free supervision from coarse 3D bounding boxes while mitigating shape biases inherent in box annotations. To achieve this, we introduce a novel mask-to-box (M2B) transformation and a color consistency (CC) loss. The M2B transformation, from a shape perspective, disentangles predictions from labels, enabling the extraction of noiseless supervision from labels while preserving object shapes independently of the box bias. From an appearance perspective, we further introduce the CC loss to provide dense supervision, which mitigates the non-unique predictions stemming from weak supervision and substantially reduces prediction variability. Furthermore, we employ a self-training (ST) strategy to enhance performance by utilizing high-confidence pseudo labels. Notably, the M2B transformation, CC loss, and ST strategy are seamlessly integrated into any model and incur no computational costs for inference. Extensive experiments demonstrate the effectiveness of our WeakPCSOD model, even comparable to fully supervised models utilizing dense annotations.

3D occupancy prediction is an emerging task that aims to estimate the occupancy states and semantics of 3D scenes using multi-view images. However, image-based scene perception encounters significant challenges in achieving accurate prediction due to the absence of geometric priors. In this paper, we address this issue by exploring cross-modal knowledge distillation in this task, i.e., we leverage a stronger multi-modal model to guide the visual model during training. In practice, we observe that directly applying features or logits alignment, proposed and widely used in bird's-eye-view (BEV) perception, does not yield satisfactory results. To overcome this problem, we introduce RadOcc, a Rendering assisted distillation paradigm for 3D Occupancy prediction. By employing differentiable volume rendering, we generate depth and semantic maps in perspective views and propose two novel consistency criteria between the rendered outputs of teacher and student models. Specifically, the depth consistency loss aligns the termination distributions of the rendered rays, while the semantic consistency loss mimics the intra-segment similarity guided by vision foundation models (VLMs). Experimental results on the nuScenes dataset demonstrate the effectiveness of our proposed method in improving various 3D occupancy prediction approaches, e.g., our proposed methodology enhances our baseline by 2.2% in the metric of mIoU and achieves 50% in Occ3D benchmark.

Limited by the expensive labeling, polyp segmentation models are plagued by data shortages. To tackle this, we propose the mixed supervised polyp segmentation paradigm (MixPolyp). Unlike traditional models relying on a single type of annotation, MixPolyp combines diverse annotation types (mask, box, and scribble) within a single model, thereby expanding the range of available data and reducing labeling costs. To achieve this, MixPolyp introduces three novel supervision losses to handle various annotations: Subspace Projection loss $\left({{{\mathcal{L}}_{{\mathcal{S}}{\mathcal{P}}}}}\right)$, Binary Minimum Entropy loss $\left({{{\mathcal{L}}_{{\mathcal{B}}{\mathcal{M}}{\mathcal{E}}}}}\right)$, and Linear Regularization loss $\left({{{\mathcal{L}}_{{\mathcal{L}}{\mathcal{R}}}}}\right)$. For box annotations, ${{\mathcal{L}}_{{\mathcal{S}}{\mathcal{P}}}}$ eliminates shape inconsistencies between the prediction and the supervision. For scribble annotations, ${{\mathcal{L}}_{{\mathcal{B}}{\mathcal{M}}{\mathcal{E}}}}$ provides supervision for unlabeled pixels through minimum entropy constraint, thereby alleviating supervision sparsity. Furthermore, ${{\mathcal{L}}_{{\mathcal{L}}{\mathcal{R}}}}$ provides dense supervision by enforcing consistency among the predictions, thus reducing the non-uniqueness. These losses are independent of the model structure, making them generally applicable. They are used only during training, adding no computational cost during inference. Extensive experiments on five datasets demonstrate MixPolyp’s effectiveness.

AI-assisted lesion detection models play a crucial role in the early screening of cancer. However, previous image-based models ignore the inter-frame contextual information present in videos. On the other hand, video-based models capture the inter-frame context but are computationally expensive. To mitigate this contradiction, we delve into Video-to-Image knowledge distillation leveraging DEtection TRansformer (V2I-DETR) for the task of medical video lesion detection. V2I-DETR adopts a teacher-student network paradigm. The teacher network aims at extracting temporal contexts from multiple frames and transferring them to the student network, and the student network is an image-based model dedicated to fast prediction in inference. By distilling multi-frame contexts into a single frame, the proposed V2I-DETR combines the advantages of utilizing temporal contexts from video-based models and the inference speed of image-based models. Through extensive experiments, V2I-DETR outperforms previous state-of-the-art methods by a large margin while achieving the real-time inference speed (30 FPS) as the image-based model.

Generalizable 3D object reconstructionfrom single-view RGB-D images remains a challenging task, particularly with real-world data. Current state-of-the-art methods develop Transformer-based implicit field learning, necessitating an intensive learning paradigm that requires dense query-supervision uniformly sampled throughout the entire space. We propose a novel approach, IPoD, which harmonizes implicit field learning with point diffusion. This approach treats the query points for implicit field learning as a noisy point cloud for iterative denoising, allowing for their dynamic adaptation to the target object shape. Such adaptive query points harness diffusion learning's capability for coarse shape recovery and also enhances the implicit representation's ability to delineate finer details. Besides, an additional self-conditioning mechanism is designed to use implicit predictions as the guidance of diffusion learning, leading to a cooperative system. Experiments conducted on the CO3D-v2 dataset affirm the superiority of IPoD, achieving 7.8% improvement in F-score and 28.6% in Chamfer distance over existing methods. The generalizability of IPoD is also demonstrated on the MVImgNet dataset. Our project page is at https://yushuang-wu.github.io/IPoD.

In this era, the success of large language models and text-to-image models can be attributed to the driving force of large-scale datasets. However, in the realm of 3D vision, while remarkable progress has been made with models trained on large-scale synthetic and real-captured object data like Objaverse and MVImgNet, a similar level of progress has not been observed in the domain of human-centric tasks partially due to the lack of a large-scale human dataset. Existing datasets of high-fidelity 3D human capture continue to be mid-sized due to the significant challenges in acquiring large-scale high-quality 3D human data. To bridge this gap, we present MVHuman-Net, a dataset that comprises multi-view human action sequences of 4,500 human identities. The primary focus of our work is on collecting human data that features a large number of diverse identities and everyday clothing using a multi- view human capture system, which facilitates easily scalable data collection. Our dataset contains 9,000 daily outfits, 60,000 motion sequences and 645 million frames with extensive annotations, including human masks, camera parameters, 2D and 3D keypoints, SMPUSMPLX parameters, and corresponding textual descriptions. To explore the potential of MVHumanNet in various 2D and 3D visual tasks, we conducted pilot studies on view-consistent action recognition, human NeRF reconstruction, text-driven view-unconstrained human image generation, as well as 2D view-unconstrained human image and 3D avatar generation. Extensive experiments demonstrate the performance improvements and effective applications enabled by the scale provided by MVHumanNet. As the current largest-scale 3D human dataset, we hope that the release of MVHu-manNet data with annotations will foster further innovations in the domain of 3D human-centric tasks at scale.

3D Visual Grounding (3DVG) aims at localizing 3D object based on textual descriptions. Conventional supervised methods for 3DVG often necessitate extensive annotations and a predefined vocabulary, which can be restrictive. To address this issue, we propose a novel visual programming approach for zero-shot open-vocabulary 3DVG, leveraging the capabilities of large language models (LLMs). Our approach begins with a unique dialog-based method, engaging with LLMs to establish a foundational understanding of zero-shot 3DVG. Building on this, we design a visual program that consists of three types of modules, i.e., view-independent, view-dependent, and functional modules. These modules, specifically tailored for 3D scenarios, work collaboratively to perform complex reasoning and inference. Furthermore, we develop an innovative language-object correlation module to extend the scope of existing 3D object detectors into open-vocabulary scenarios. Extensive experiments demonstrate that our zero-shot approach can outperform some supervised baselines, marking a significant stride towards effective 3DVG. Code is available at https://curryyuan.github.io/Z5VG3D.

We present a neural radiance field method for urban-scale semantic and building-level instance segmentation from aerial images by lifting noisy 2D labels to 3D. This is a challenging problem due to two primary reasons. Firstly, objects in urban aerial images exhibit substantial variations in size, including buildings, cars, and roads, which pose a significant challenge for accurate 2D segmentation. Secondly, the 2D labels generated by existing segmentation methods suffer from the multi-view inconsistency problem, especially in the case of aerial images, where each image captures only a small portion of the entire scene. To overcome these limitations, we first introduce a scale-adaptive semantic label fusion strategy that enhances the segmentation of objects of varying sizes by combining labels predicted from different altitudes, harnessing the novel-view synthesis capabilities of NeRF. We then introduce a novel cross-view instance label grouping based on the 3D scene representation to mitigate the multi-view inconsistency problem in the 2D instance labels. Furthermore, we exploit multi-view reconstructed depth priors to improve the geometric quality of the reconstructed radiance field, resulting in enhanced segmentation results. Experiments on multiple real-world urban-scale datasets demonstrate that our approach outperforms existing methods, high-lighting its effectiveness. The source code is available at https://github.com/zyqz97/Aerial_lifting.

Large Language Models (LLMs) have achieved significant success in various natural language processing tasks, but how wireless networks can support LLMs has not been extensively studied. In this paper, we propose a wireless distributed LLMs paradigm based on Mixture of Experts (MoE), named WDMoE, through server-device collaboration at the wireless network edge. Specifically, we decompose the MoE layer in LLMs by deploying the gating network and the preceding neural network layer at the edge server of the base station (BS), while distributing the expert networks across the mobile devices. This arrangement leverages the parallel capabilities of expert networks on distributed devices. Moreover, to overcome the instability of wireless communications, we design an expert selection policy by taking into account both the performance of the model and the end-to-end latency, which includes both transmission delay and inference delay. Evaluations conducted across various LLMs and multiple datasets demonstrate that WDMoE not only outperforms existing models, such as Llama 2 with 70 billion parameters, but also significantly reduces end-to-end latency.

Semantic communication has emerged as a promising communication paradigm and there have been extensive research focusing on its applications in the increasingly prevalent multi-user scenarios. However, the knowledge discrepancy among multiple users may lead to considerable disparities in their performance. To address this challenge, this paper proposes a novel multi-pair cooperative semantic knowledge base (SKB) update policy. Specifically, for each pair endowed with SKB-enabled semantic communication, its well-understood knowledge in the local SKB is selected out and uploaded to the server to establish a global SKB, via a score-based knowledge selection scheme. The knowledge selection scheme achieves a balance between the uplink transmission overhead and the completeness of the global SKB. Then, with the assistance of the global SKB, each pair’s local SKB is refined and their performance is improved. Numerical results show that the proposed cooperative SKB update policy obtains significant performance gains with minimal transmission overhead, especially for the initially poor-performing pairs.

Vector quantization-based image semantic communication systems have successfully boosted transmission efficiency, but face challenges with conflicting requirements between code-book design and digital constellation modulation. Traditional codebooks need wide index ranges, while modulation favors few discrete states. To address this, we propose a multilevel generative semantic communication system with a two-stage training framework. In the first stage, we train a high-quality codebook, using a multi-head octonary codebook (MOC) to compress the index range. In addition, a residual vector quantization (RVQ) mechanism is also integrated for effective multilevel communication. In the second stage, a noise reduction block (NRB) based on Swin Transformer is introduced, coupled with the multilevel codebook from the first stage, serving as a high-quality semantic knowledge base (SKB) for generative feature restoration. Finally, to simulate modern image transmission scenarios, we employ a diverse collection of high-resolution 2K images as the test set. The experimental results consistently demonstrate the superior performance of MOC-RVQ over conventional methods such as BPG or JPEG. Additionally, MOC-RVQ achieves comparable performance to an analog JSCC scheme, while needing only one-sixth of the channel bandwidth ratio (CBR) and being directly compatible with digital transmission systems.

This paper revisits the identity detection problem under the current grant-free protocol in massive machine-type communications (mMTC) by asking the following question: for stable identity detection performance, is it enough to permit active devices to transmit preambles without any handshaking with the base station (BS)? Specifically, in the current grant-free protocol, the BS blindly allocates a fixed length of preamble to devices for identity detection as it lacks the prior information on the number of active devices K. However, in practice, K varies dynamically over time, resulting in degraded identity detection performance especially when K is large. Consequently, the current grant-free protocol fails to ensure stable identity detection performance. To address this issue, we propose a two-stage communication protocol which consists of estimation of K in Phase I and detection of identities of active devices in Phase II. The preamble length for identity detection in Phase II is dynamically allocated based on the estimated K in Phase I through a table lookup manner such that the identity detection performance could always be better than a predefined threshold. In addition, we design an algorithm for estimating K in Phase I, and exploit the estimated K to reduce the computational complexity of the identity detector in Phase II. Numerical results demonstrate the effectiveness of the proposed two-stage communication protocol and algorithms.

Conventional data compression methods typically model the information source as an i.i.d. stochastic process, thereby establishing the fundamental limit as entropy for lossless compression and as mutual information for lossy compression. However, the source in the real world (e.g., text, music, and speech) is often statistically ill-defined because of its close connection to human perception. This work aims to exploit the semantic aspect of text as inspired by the puzzle crossword. The main idea is to only compress those semantically important words while masking the rest; the proposed decompressor can recover all the missing words automatically according to context. Experiments show that the proposed semantic approach can achieve much higher compression efficiency than the state-of-the-art semantic compression method.

Unmanned Aerial Vehicle (UAV)-mounted edge devices are particularly advantageous for federated edge learning (FEEL) due to their flexibility and mobility in efficient data collection. In UAV-assisted FEEL, sensing, computation, and communication are coupled and compete for limited onboard resources, and UAV deployment also affects sensing and communication performance. Therefore, the joint design of UAV deployment and resource allocation is crucial to achieving the optimal training performance. In this paper, we address the problem of joint UAV deployment design and resource allocation for FEEL via a concrete case study of human motion recognition based on wireless sensing. Due to the nonideal sensing channels, we consider the probabilistic sensing model. Then, we derive the upper bound of the FEEL training loss as a function of the sensing probability. We formulate a training time minimization problem by jointly optimizing UAV deployment, integrated sensing, computation, and communication (ISCC) resources under a desirable optimality gap constraint. To solve this challenging mixed-integer non-convex problem, we propose our algorithm based on the alternating optimization technique. Simulation results demonstrate that our algorithm outperforms other baselines regarding convergence rate and testing accuracy.

With the proliferation of edge computing, efficient AI inference on edge devices has become essential for intelligent applications such as autonomous vehicles and VR/AR. In this context, we address the problem of efficient remote object recognition by optimizing feature transmission between mobile devices and edge servers. We propose an online optimization framework to address the challenge of dynamic channel conditions and device mobility in an end-to-end communication system. Our approach builds upon existing methods by leveraging a semantic knowledge base to drive multi-level feature transmission, accounting for temporal factors and dynamic elements throughout the transmission process. To solve the online optimization problem, we design a novel soft actor-critic-based deep reinforcement learning system with a carefully designed reward function for real-time decision-making, overcoming the optimization difficulty of the NP-hard problem and achieving the minimization of semantic loss while respecting latency constraints. Numerical results showcase the superiority of our approach compared to traditional greedy methods under various system setups.

Accurate 3D lane estimation is crucial for ensuring safety in autonomous driving. However, prevailing monocular techniques suffer from depth loss and lighting variations, hampering accurate 3D lane detection. In contrast, LiDAR points offer geometric cues and enable precise localization. In this paper, we present DV-3DLane, a novel end-to-end **D**ual-**V**iew multi-modal **3D Lane** detection framework that synergizes the strengths of both images and LiDAR points. We propose to learn multi-modal features in dual-view spaces, *i.e.*, *perspective view* (PV) and *bird's-eye-view* (BEV), effectively leveraging the modal-specific information. To achieve this, we introduce three designs: 1) A bidirectional feature fusion strategy that integrates multi-modal features into each view space, exploiting their unique strengths. 2) A unified query generation approach that leverages lane-aware knowledge from both PV and BEV spaces to generate queries. 3) A 3D dual-view deformable attention mechanism, which aggregates discriminative features from both PV and BEV spaces into queries for accurate 3D lane detection. Extensive experiments on the public benchmark, OpenLane, demonstrate the efficacy and efficiency of DV-3DLane. It achieves state-of-the-art performance, with a remarkable 11.2 gain in F1 score and a substantial 53.5% reduction in errors. Code is available on [github](https://github.com/JMoonr/dv-3dlane).

The vast applications of deep generative models are anchored in three core capabilities---*generating* new instances, *reconstructing* inputs, and learning compact *representations*---across various data types, such as discrete text/protein sequences and continuous images. Existing model families, like variational autoencoders (VAEs), generative adversarial networks (GANs), autoregressive models, and (latent) diffusion models, generally excel in specific capabilities and data types but fall short in others. We introduce *Generalized* ***E****ncoding*-***D****ecoding ****D****iffusion ****P****robabilistic ****M****odels* (EDDPMs) which integrate the core capabilities for broad applicability and enhanced performance. EDDPMs generalize the Gaussian noising-denoising in standard diffusion by introducing parameterized encoding-decoding. Crucially, EDDPMs are compatible with the well-established diffusion model objective and training recipes, allowing effective learning of the encoder-decoder parameters *jointly* with diffusion. By choosing appropriate encoder/decoder (e.g., large language models), EDDPMs naturally apply to different data types. Extensive experiments on text, proteins, and images demonstrate the flexibility to handle diverse data and tasks and the strong improvement over various existing models. Code is available at https://github.com/guangyliu/EDDPM .

While 3D object bounding box (bbox) representation has been widely used in autonomous driving perception, it lacks the ability to capture the precise details of an object's intrinsic geometry. Recently, occupancy has emerged as a promising alternative for 3D scene perception. However, constructing a high-resolution occupancy map remains infeasible for large scenes due to computational constraints. Recognizing that foreground objects only occupy a small portion of the scene, we introduce object-centric occupancy as a supplement to object bboxes. This representation not only provides intricate details for detected objects but also enables higher voxel resolution in practical applications. We advance the development of object-centric occupancy perception from both data and algorithm perspectives. On the data side, we construct the first object-centric occupancy dataset from scratch using an automated pipeline. From the algorithmic standpoint, we introduce a novel object-centric occupancy completion network equipped with an implicit shape decoder that manages dynamic-size occupancy generation. This network accurately predicts the complete object-centric occupancy volume for inaccurate object proposals by leveraging temporal information from long sequences. Our method demonstrates robust performance in completing object shapes under noisy detection and tracking conditions. Additionally, we show that our occupancy features significantly enhance the detection results of state-of-the-art 3D object detectors, especially for incomplete or distant objects in the Waymo Open Dataset.

Many networking tasks now employ deep learning (DL) to solve complex prediction and optimization problems. However, current design philosophy of DL-based algorithms entails intensive engineering overhead due to the manual design of deep neural networks (DNNs) for different networking tasks. Besides, DNNs tend to achieve poor generalization performance on unseen data distributions/environments.

In this paper, we introduce a realistic and challenging domain adaptation problem called Universal Semi-supervised Model Adaptation (USMA), which i) requires only a pre-trained source model, ii) allows the source and target domain to have different label sets, i.e., they share a common label set and hold their own private label set, and iii) requires only a few labeled samples in each class of the target domain. To address USMA, we propose a collaborative consistency training framework that regularizes the prediction consistency between two models, i.e., a pre-trained source model and its variant pre-trained with target data only, and combines their complementary strengths to learn a more powerful model. The rationale of our framework stems from the observation that the source model performs better on common categories than the target-only model, while on target-private categories, the target-only model performs better. We also propose a two-perspective, i.e., sample-wise and class-wise, consistency regularization to improve the training. Experimental results demonstrate the effectiveness of our method on several benchmark datasets.

This paper studies the integration of over-the-air computation (AirComp) and wireless power transfer (WPT) for achieving sustainable wireless data aggregation (WDA). In such wireless powered AirComp system, a multi-antenna hybrid access point (HAP) employs the transmit energy beamforming to charge multiple single-antenna low-power wireless devices (WDs) in the downlink, and the WDs utilize their harvested energy to simultaneously send messages to the HAP for AirComp in the uplink. Under this setup, our objective is to minimize the computation mean square error (MSE) by jointly optimizing the transmit en-ergy beamforming and the receive AirComp beamforming at the HAP, as well as the transmit power control at the WDs, subject to the wireless energy harvesting constraints at individual WDs. To tackle the non-convex computation MSE minimization problem, we present an efficient algorithm to find a converged high-quality solution by using the alternating optimization technique, in which the transmit energy beamforming (together with WDs' power control) and the receive beamforming are alternately optimized. Simulation results show that the proposed joint WPT-AirComp scheme significantly decreases the system's MSE, as compared to conventional designs without such joint optimization.

Mobile-edge computing (MEC) networks bring computing and storage capabilities closer to edge devices, which reduces latency and improves network performance. However, to further reduce transmission and computation costs while satisfying user-perceived quality of experience, a joint optimization in computing, pushing, and caching is needed. In this article, we formulate the joint-design problem in MEC networks as an infinite-horizon discounted-cost Markov decision process and solve it using a deep reinforcement learning (DRL)-based framework that enables the dynamic orchestration of computing, pushing, and caching. Through the deep networks embedded in the DRL structure, our framework can implicitly predict user future requests and push or cache the appropriate content to effectively enhance system performance. One issue we encountered when considering three functions collectively is the curse of dimensionality for the action space. To address it, we relaxed the discrete action space into a continuous space and then adopted soft actor–critic learning to solve the optimization problem, followed by utilizing a vector quantization method to obtain the desired discrete action. Additionally, an action correction method was proposed to compress the action space further and accelerate the convergence. Our simulations under the setting of a general single-user, single-server MEC network with dynamic transmission link quality demonstrate that the proposed framework effectively decreases transmission bandwidth and computing cost by proactively pushing data on future demand to users and jointly optimizing the three functions. We also conduct extensive parameter tuning analysis, which shows that our approach outperforms the baselines under various parameter settings.

We investigate the coexistence of task-oriented and data-oriented communications in a IoT system that shares a group of channels, and study the scheduling problem to jointly optimize the weighted age of incorrect information (AoII) and throughput, which are the performance metrics of the two types of communications, respectively. This problem is formulated as a Markov decision problem, which is difficult to solve due to the large discrete action space and the time-varying action constraints induced by the stochastic availability of channels. By exploiting the intrinsic properties of this problem and reformulating the reward function based on channel statistics, we first simplify the solution space, state space, and optimality criteria, and convert it to an equivalent Markov game, for which the large discrete action space issue is greatly relieved. Then, we propose a Whittle’s index guided multi-agent proximal policy optimization (WI-MAPPO) algorithm to solve the considered game, where the embedded Whittle’s index module further shrinks the action space, and the proposed offline training algorithm extends the training kernel of conventional MAPPO to address the issue of time-varying constraints. Finally, numerical results validate that the proposed algorithm significantly outperforms state-of-the-art age of information (AoI) based algorithms under scenarios with insufficient channel resources.

This article considers a heterogeneous mobile-edge computing (HMEC) system with multiple energy-limited Internet of Things (IoT) devices and an unmanned aerial vehicle (UAV). The UAV can supply energy to all the IoT devices through wireless power transfer. To maximize the utilization of the communication and computation resources, all the IoT devices are divided into two groups, i.e., the active devices and the idle devices. The UAV and the idle devices assist the active devices in executing computing tasks. We formulate an optimization problem that maximizes the minimum task computation data volume among all the active devices by jointly optimizing the UAV trajectory and the communication and computation resource allocation. Since the problem is nonconvex, we decompose the problem into two subproblems: 1) the UAV trajectory design and the computation resource allocation and 2) the time allocation. We utilize a block coordinate descent approach to solve these two subproblems alternately. Simulation results demonstrate that the proposed algorithm can provide an optimized trajectory robust to different initializations. Additionally, compared to the benchmark algorithms, our proposed algorithm shows superior performance in terms of system efficiency and computation data volume.

This paper investigates the energy efficiency of a multiple-input multiple-output (MIMO) integrated sensing and communications (ISAC) system for Internet of things (IoT), in which one multi-antenna IoT transceiver transmits unified ISAC signals to a multi-antenna communication user (CU) and at the same time use the echo signals to estimate an extended target. We focus on one particular ISAC transmission block and take into account the practical on-off non-transmission power at the IoT transceiver. Under this setup, we minimize the energy consumption at the transceiver while ensuring a minimum average data rate requirement for communication and a maximum Cramér-Rao bound (CRB) requirement for target estimation, by jointly optimizing the transmit covariance matrix and the “on” duration for active transmission. We obtain the optimal solution to the rate-and-CRB-constrained energy minimization problem in a semi-closed form. Interestingly, the obtained optimal solution is shown to unify the spectrum-efficient and energy-efficient communications and sensing designs. In particular, for the special MIMO sensing case with rate constraint inactive, the optimal solution follows the isotropic transmission with shortest “on” duration, in which the IoT transceiver radiates the required sensing energy by using sufficiently high power over the shortest duration. For the general ISAC case, the optimal transmit covariance solution is of full rank and follows the eigenmode transmission based on the communication channel, while the optimal “on” duration is determined based on both the rate and CRB constraints. Numerical results show that the proposed ISAC design achieves significantly reduced energy consumption as compared to the benchmark schemes based on isotropic transmission, always-on transmission, and sensing or communications only designs, especially when the rate and CRB constraints become stringent.

Optimizing edge caching is crucial for the advancement of next-generation (nextG) wireless networks, ensuring high-speed and low-latency services for mobile users. Existing data-driven optimization approaches often lack awareness of the distribution of random data variables and focus solely on optimizing cache hit rates, neglecting potential reliability concerns, such as base station overload and unbalanced cache issues. This oversight can result in system crashes and degraded user experience. To bridge this gap, we introduce a novel digital twin-assisted optimization framework, called D-REC, which integrates reinforcement learning (RL) with diverse intervention modules to ensure reliable caching in nextG wireless networks. We first develop a joint vertical and horizontal twinning approach to efficiently create network digital twins, which are then employed by D-REC as RL optimizers and safeguards, providing ample datasets for training and predictive evaluation of our cache replacement policy. By incorporating reliability modules into a constrained Markov decision process, D-REC can adaptively adjust actions, rewards, and states to comply with advantageous constraints, minimizing the risk of network failures. Theoretical analysis demonstrates comparable convergence rates between D-REC and vanilla data-driven methods without compromising caching performance. Extensive experiments validate that D-REC outperforms conventional approaches in cache hit rate and load balancing while effectively enforcing predetermined reliability intervention modules.

In the sixth-generation (6G) networks, massive low-power devices are expected to sense environment and deliver tremendous data. To enhance the radio resource efficiency, the integrated sensing and communication (ISAC) technique exploits the sensing and communication functionalities of signals, while the simultaneous wireless information and power transfer (SWIPT) techniques utilizes the same signals as the carriers for both information and power delivery. The further combination of ISAC and SWIPT leads to the advanced technology namely integrated sensing, communication, and power transfer (ISCPT). In this paper, a multi-user multiple-input multiple-output (MIMO) ISCPT system is considered, where a base station equipped with multiple antennas transmits messages to multiple information receivers (IRs), transfers power to multiple energy receivers (ERs), and senses a target simultaneously. The sensing target can be regarded as a point or an extended surface. When the locations of IRs and ERs are separated, the MIMO beamforming designs are optimized to improve the sensing performance while meeting the communication and power transfer requirements. The resultant non-convex optimization problems are solved based on a series of techniques including Schur complement transformation and rank reduction. Moreover, when the IRs and ERs are co-located, the power splitting factors are jointly optimized together with the beamformers to balance the performance of communication and power transfer. To better understand the performance of ISCPT, the target positioning problem is further investigated. Simulations are conducted to verify the effectiveness of our proposed designs, which also reveal a performance tradeoff among sensing, communication, and power transfer.

3D single object tracking in LiDAR point clouds (LiDAR SOT) plays a crucial role in autonomous driving. Current approaches all follow the Siamese paradigm based on appearance matching. However, LiDAR point clouds are usually textureless and incomplete, which hinders effective appearance matching. Besides, previous methods greatly overlook the critical motion clues among targets. In this work, beyond 3D Siamese tracking, we introduce amotion-centric paradigmto handle LiDAR SOT from a new perspective. Following this paradigm, we propose a matching-free two-stage trackerM$^{2}$2-Track. At the 1st-stage,$M^{2}$-Track localizes the target within successive frames viamotion transformation. Then it refines the target box throughmotion-assisted shape completion at the 2nd-stage. Due to the motion-centric nature, our method shows its impressive generalizability with limited training labels and provides good differentiability for end-to-end cycle training. This inspires us to explore semi-supervised LiDAR SOT by incorporating a pseudo-label-based motion augmentation and a self-supervised loss term. Under the fully-supervised setting, extensive experiments confirm that$M^{2}$-Track significantly outperforms previous state-of-the-arts on three large-scale datasets while running at57FPS($\sim$∼3%,$\sim$∼11%and$\sim$∼22%precision gains on KITTI, NuScenes, and Waymo Open Dataset respectively). While under the semi-supervised setting, our method performs on par with or even surpasses its fully-supervised counterpart using fewer than half labels from KITTI. Further analysis verifies each component's effectiveness and shows the motion-centric paradigm's promising potential for auto-labeling and unsupervised domain adaptation.

This paper studies the over-the-air computation (AirComp) in an orthogonal frequency division multiplexing (OFDM) system with imperfect channel state information (CSI), in which multiple single-antenna wireless devices (WDs) simultaneously send uncoded signals to a multi-antenna access point (AP) for distributed functional computation over multiple subcarriers. In particular, we consider two scenarios with best-effort and error-constrained computation tasks, with the objectives of minimizing the average computation mean squared error (MSE) and the computation outage probability over the multiple subcarriers, respectively. Towards this end, we jointly optimize the transmit coefficients at the WDs and the receive beamforming vectors at the AP over subcarriers, subject to the maximum transmit power constraints at individual WDs. First, for the special case with a single receive antenna at the AP, we propose the semi-closed-form globally optimal solutions to the two problems using the Lagrange-duality method. It is shown that at each subcarrier, the WDs’ optimized power control policy for average MSE minimization follows a regularized channel inversion structure, while that for computation outage probability minimization follows an on-off regularized channel inversion, with the regularization dependent on the transmit power budget and channel estimation error. Next, for the general case with multiple receive antennas at the AP, we present efficient algorithms based on alternating optimization and convex optimization to find converged solutions to both problems. It is shown that with finite receive antennas at the AP, a non-zero computation MSE for AirComp is inevitable due to the channel estimation errors even when the transmit powers at WDs tend to infinity, while with massive receive antennas, the average MSE and outage probability vanish when the channel vectors are independent and identically distributed. Finally, numerical results are provided to demonstrate the effectiveness of the proposed designs.

This paper studies a multi-intelligent-reflecting-surface-(IRS)-enabled integrated sensing and communications (ISAC) system, in which multiple IRSs are installed to help the base station (BS) provide ISAC services at separate line-of-sight (LoS) blocked areas. We focus on the scenario with semi-passive uniform linear array (ULA) IRSs for sensing, in which each IRS is integrated with dedicated sensors for processing echo signals, and each IRS simultaneously serves one sensing target and multiple communication users (CUs) in its coverage area. We consider two cases with point and extended targets, in which each IRS aims to estimate the target direction-of-arrival (DoA) and the complete target response matrix, respectively. Under this setup, we first derive the closed-form Cramér-Rao bounds (CRBs) for parameter estimation under the two target models. Then, we assume that the BS sends combined information and dedicated sensing signals for ISAC, and accordingly consider two different types of CU receivers that can and cannot cancel the interference from dedicated sensing signals. Under this setup, we minimize the maximum CRB at all IRSs, via jointly optimizing the transmit beamformers at the BS and the reflective beamformers at the multiple IRSs, subject to the minimum signal-to-interference-plus-noise ratio (SINR) constraints at individual CUs, the maximum transmit power constraint at the BS, and the unit-modulus constraints at the multiple IRSs. To tackle the highly non-convex SINR-constrained max-CRB minimization problems, we propose efficient algorithms based on alternating optimization and semi-definite relaxation, to obtain converged solutions. Finally, numerical results are provided to verify the benefits of our proposed designs over various benchmark schemes based on separate or heuristic beamforming designs.

Visible light positioning (VLP) is an accurate and low-cost positioning technique. However, existing VLP algorithms require multiple luminaires or multiple sensors to achieve the desired positioning accuracy, which may not be satisfied in practice. To circumvent this challenge, a novel visual odometry (VO) assisted VLP algorithm (VO-VLP) is proposed, which can achieve accurate positioning using only a single luminaire at the transmitter and a single camera at the receiver. In the considered model, the luminaires are equipped on the ceiling and consistently broadcast coordinate information of the luminaires by visible light communication (VLC). A user equipped with a camera captures photos of the ceiling so as to locate its position via VO-VLP. In particular, VO-VLP first uses the single luminaire’s circle feature and VLC information to obtain the pose and location of the user. However, there are dual solutions due to the limited received information in the single-luminaire scenario and the symmetry of the circular luminaire. Then, we propose a duality elimination method to eliminate the wrong one by introducing VO to exploit the visual features on the ceiling in two consecutive images, which are captured when the user moves. To verify the feasibility of our designed VO-VLP, a prototype is implemented. A cooperative multi-information image processing method is proposed for the prototype to ensure that the VLC information and the visual information of the luminaire and the ceiling can be simultaneously received for real-time positioning. Simulations and experiments are conducted to prove that VO-VLP can achieve accurate positioning with only a single luminaire and a camera without any extra sensors, such as an inertial measurement unit. In particular, simulation results show that the proposed indoor positioning algorithm can achieve a 97% positioning accuracy of around 10 cm, and experimental results show that the average positioning accuracy is less than 10 cm.

In this paper, we address a complex but practical scenario in Active Learning (AL) known as open-set AL, where the unlabeled data consists of both in-distribution (ID) and out-of-distribution (OOD) samples. Standard AL methods will fail in this scenario as OOD samples are highly likely to be regarded as uncertain samples, leading to their selection and wasting of the budget. Existing methods focus on selecting the highly likely ID samples, which tend to be easy and less informative. To this end, we introduce two criteria, namely contrastive confidence and historical divergence, which measure the possibility of being ID and the hardness of a sample, respectively. By balancing the two proposed criteria, highly informative ID samples can be selected as much as possible. Furthermore, unlike previous methods that require additional neural networks to detect the OOD samples, we propose a contrastive clustering framework that endows the classifier with the ability to identify the OOD samples and further enhances the network's representation learning. The experimental results demonstrate that the proposed method achieves state-of-the-art performance on several benchmark datasets.

Accurate polyp detection is critical for early colorectal cancer diagnosis. Although remarkable progress has been achieved in recent years, the complex colon environment and concealed polyps with unclear boundaries still pose severe challenges in this area. Existing methods either involve computationally expensive context aggregation or lack prior modeling of polyps, resulting in poor performance in challenging cases. In this paper, we propose the Enhanced CenterNet with Contrastive Learning (ECC-PolypDet), a two-stage training & end-to-end inference framework that leverages images and bounding box annotations to train a general model and fine-tune it based on the inference score to obtain a final robust model. Specifically, we conduct Box-assisted Contrastive Learning (BCL) during training to minimize the intra-class difference and maximize the inter-class difference between foreground polyps and backgrounds, enabling our model to capture concealed polyps. Moreover, to enhance the recognition of small polyps, we design the Semantic Flow-guided Feature Pyramid Network (SFFPN) to aggregate multi-scale features and the Heatmap Propagation (HP) module to boost the model's attention on polyp targets. In the fine-tuning stage, we introduce the IoU-guided Sample Re-weighting (ISR) mechanism to prioritize hard samples by adaptively adjusting the loss weight for each sample during fine-tuning. Extensive experiments on six large-scale colonoscopy datasets demonstrate the superiority of our model compared with previous state-of-the-art detectors.

The high-accuracy and resource-intensive deep neural networks (DNNs) have been widely adopted by live video analytics (VA), where camera videos are streamed over the network to resource-rich edge/cloud servers for DNN inference. Common video encoding configurations (e.g., resolution and frame rate) have been identified with significant impacts on striking the balance between bandwidth consumption and inference accuracy and therefore their adaption scheme has been a focus of optimization. However, previous profiling-based solutions suffer from high profiling cost, while existing deep reinforcement learning (DRL) based solutions may achieve poor performance due to the usage of fixed reward function for training the agent, which fails to craft the application goals in various scenarios. In this paper, we proposeILCAS, the first imitation learning (IL) based configuration-adaptive VA streaming system. Unlike DRL-based solutions,ILCAStrains the agent with demonstrations collected from the expert which is designed as an offline optimal policy that solves the configuration adaption problem through dynamic programming. To tackle the challenge of video content dynamics,ILCASderives motion feature maps based on motion vectors which allowILCASto visually “perceive” video content changes. Moreover,ILCASincorporates a cross-camera collaboration scheme to exploit the spatio-temporal correlations of cameras for more proper configuration selection. Extensive experiments confirm the superiority ofILCAScompared with state-of-the-art solutions, with 2-20.9% improvement of mean accuracy and 19.9–85.3% reduction of chunk upload lag.

Despite simplicity, stochastic gradient descent (SGD)-like algorithms are successful in training deep neural networks (DNNs). Among various attempts to improve SGD, weight averaging (WA), which averages the weights of multiple models, has recently received much attention in the literature. Broadly, WA falls into two categories: 1) online WA, which averages the weights of multiple models trained in parallel, is designed for reducing the gradient communication overhead of parallel mini-batch SGD and 2) offline WA, which averages the weights of one model at different checkpoints, is typically used to improve the generalization ability of DNNs. Though online and offline WA are similar in form, they are seldom associated with each other. Besides, these methods typically perform either offline parameter averaging or online parameter averaging, but not both. In this work, we first attempt to incorporate online and offline WA into a general training framework termed hierarchical WA (HWA). By leveraging both the online and offline averaging manners, HWA is able to achieve both faster convergence speed and superior generalization performance without any fancy learning rate adjustment. Besides, we also analyze the issues faced by the existing WA methods, and how our HWA addresses them, empirically. Finally, extensive experiments verify that HWA outperforms the state-of-the-art methods significantly.

MVImgNet is a large-scale dataset that contains multi-view images of ~220k real-world objects in 238 classes. As a counterpart of ImageNet, it introduces 3D visual signals via multi-view shooting, making a soft bridge between 2D and 3D vision. This paper constructs the MVImgNet2.0 dataset that expands MVImgNet into a total of ~520k objects and 515 categories, which derives a 3D dataset with a larger scale that is more comparable to ones in the 2D domain. In addition to the expanded dataset scale and category range, MVImgNet2.0 is of a higher quality than MVImgNet owing to four new features: (i) most shoots capture 360° views of the objects, which can support the learning of object reconstruction with completeness; (ii) the segmentation manner is advanced to produce foreground object masks of higher accuracy; (iii) a more powerful structure-from-motion method is adopted to derive the camera pose for each frame of a lower estimation error; (iv) higher-quality dense point clouds are reconstructed via advanced methods for objects captured in 360 ° views, which can serve for downstream applications. Extensive experiments confirm the value of the proposed MVImgNet2.0 in boosting the performance of large 3D reconstruction models. MVImgNet2.0 will be public at luyues.github.io/mvimgnet2 , including multi-view images of all 520k objects, the reconstructed high-quality point clouds, and data annotation codes, hoping to inspire the broader vision community.

Sampling, grouping, and aggregation are three important components in the multi-scale analysis of point clouds. In this paper, we present a novel data-driven sampler learning strategy for point-wise analysis tasks. Unlike the widely used sampling technique, Farthest Point Sampling (FPS), we propose to learn sampling and downstream applications jointly. Our key insight is that uniform sampling methods like FPS are not always optimal for different tasks: sampling more points around boundary areas can make the point-wise classification easier for segmentation. Towards this end, we propose a novel sampler learning strategy that learns sampling point displacement supervised by task-related ground truth information and can be trained jointly with the underlying tasks. We further demonstrate our methods in various point-wise analysis tasks, including semantic part segmentation, point cloud completion, and keypoint detection. Our experiments show that jointly learning of the sampler and task brings better performance than using FPS in various point-based networks.

Visual Question Answering on 3D Point Cloud (VQA-3D) is an emerging yet challenging field that aims at answering various types of textual questions given an entire point cloud scene. To tackle this problem, we propose the CLEVR3D, a large-scale VQA-3D dataset consisting of 171K questions from 8,771 3D scenes. Specifically, we develop a question engine leveraging 3D scene graph structures to generate diverse reasoning questions, covering the questions of objects' attributes (i.e., size, color, and material) and their spatial relationships. Through such a manner, we initially generated 44K questions from 1,333 real-world scenes. Moreover, a more challenging setup is proposed to remove the confounding bias and adjust the context from a common-sense layout. Such a setup requires the network to achieve comprehensive visual understanding when the 3D scene is different from the general co-occurrence context (e.g., chairs always exist with tables). To this end, we further introduce the compositional scene manipulation strategy and generate 127K questions from 7,438 augmented 3D scenes, which can improve VQA-3D models for real-world comprehension. Built upon the proposed dataset, we baseline several VQA-3D models, where experimental results verify that the CLEVR3D can significantly boost other 3D scene understanding tasks.

With the recent advancements in edge artificial intelligence (AI), future sixth-generation (6G) networks need to support new AI tasks such as classification and clustering apart from data recovery. Motivated by the success of deep learning, the semantic-aware and task-oriented communications with deep joint source and channel coding (JSCC) have emerged as new paradigm shifts in 6G from the conventional data-oriented communications with separate source and channel coding (SSCC). However, most existing works focused on the deep JSCC designs for one task of data recovery or AI task execution independently, which cannot be transferred to other unintended tasks. Differently, this paper investigates the JSCC semantic communications to support multi-task services, by performing the image data recovery and classification task execution simultaneously. First, we propose a new end-to-end deep JSCC framework by unifying the coding rate reduction maximization and the mean square error (MSE) minimization in the loss function. Here, the coding rate reduction maximization facilitates the learning of discriminative features for enabling to perform classification tasks directly in the feature space, and the MSE minimization helps the learning of informative features for high-quality image data recovery. Next, to further improve the robustness against variational wireless channels, we propose a new gated deep JSCC design, in which a gated net is incorporated for adaptively pruning the output features to adjust their dimensions based on channel conditions. Finally, we present extensive numerical experiments to validate the performance of our proposed deep JSCC designs as compared to various benchmark schemes. It is shown that our proposed designs simultaneously provide efficient multi-task services, and the proposed gated deep JSCC framework efficiently reduces the communication overhead with only marginal performance loss. It is also shown that performing the classification task on the feature space via coding rate reduction maximization is able to better defend the label corruption than the traditional label-fitting methods.

Remote zero-shot object recognition, which involves offloading the zero-shot recognition task from one mobile device to a remote mobile edge computing (MEC) server or another mobile device, is crucial for 6G. To address this challenge, this paper presents a lightweight semantic knowledge base (SKB)-enabled multi-level feature extractor that projects the image into visual, semantic, and intermediate feature spaces. Then, this paper proposes a novel SKB-enabled multi-level feature transmission framework, which utilizes SKB and multi-level feature extractor at both transmitter and receiver. The semantic loss and required transmission latency at each level are characterized, and a multi-level feature transmission optimization problem is formulated to minimize the semantic loss under transmission latency constraint. However, this optimization problem is a multi-choice knapsack problem, which is challenging to solve optimally. To overcome this issue, an enhanced convex concave procedure is proposed to obtain an efficient solution. Furthermore, this paper theoretically analyzes the effects of SKBs on the communication performance when the feature extractors at both ends are the same. Numerical results demonstrate that the proposed design outperforms the benchmarks and provide insights into the impact of SKBs at both ends on performance as well as the tradeoff between transmission latency and zero-shot classification accuracy.

This paper considers improving wireless communication and computation efficiency in federated learning (FL) via model quantization. In the proposed bitwidth FL scheme, edge devices train and transmit quantized versions of their local FL model parameters to a coordinating server, which, in turn, aggregates them into a quantized global model and synchronizes the devices. The goal is to jointly determine the bitwidths employed for local FL model quantization and the set of devices participating in FL training at each iteration. We pose this as an optimization problem that aims to minimize the training loss of quantized FL under a per-iteration device sampling budget and delay requirement. However, the formulated problem is difficult to solve without (i) a concrete understanding of how quantization impacts global ML performance and (ii) the ability of the server to construct estimates of this process efficiently. To address the first challenge, we analytically characterize how limited wireless resources and induced quantization errors affect the performance of the proposed FL method. Our results quantify how the improvement of FL training loss between two consecutive iterations depends on the device selection and quantization scheme as well as on several parameters inherent to the model being learned. Then, to address the second challenge, we show that the FL training process can be described as a Markov decision process (MDP) and propose a model-based reinforcement learning (RL) method to optimize action selection over iterations. Compared to model-free RL, this model-based RL approach leverages the derived mathematical characterization of the FL training process to discover an effective device selection and quantization scheme without imposing additional device communication overhead. Simulation results show that the proposed FL algorithm can reduce the convergence time by 29% and 63% compared to a model free RL method and the standard FL method, respectively.

This paper investigates the effect of low-resolution analog-to-digital converters (ADCs) on device activity detection in massive machine-type communications (mMTC). The low-resolution ADCs induce two challenges on the device activity detection compared with the traditional setup with the assumption of infinite ADC resolution. First, the codebook design for signal quantization by the low-resolution ADC is particularly important since a good design of the codebook can lead to small quantization error on the received signal, which in turn has significant influence on the activity detector performance. To this end, prior information about the received signal power is needed, which depends on the number of active devicesK. This is sharply different from the activity detection problem in traditional setups, in which the knowledge ofKis not required by the BS as a prerequisite. Second, the covariance-based approach achieves good activity detection performance in traditional setups while it is not clear if it can still achieve good performance in this paper. To solve the above challenges, we propose a communication protocol that consists of an estimator forKand a detector for active device identities: 1) For the estimator, the technical difficulty is that the design of the ADC quantizer and the estimation ofKare closely intertwined and doing one needs the information/execution from the other. We propose a progressive estimator which iteratively performs the estimation ofKand the design of the ADC quantizer; 2) For the activity detector, we propose a custom-designed stochastic gradient descent algorithm to estimate the active device identities. Numerical results demonstrate the effectiveness of the communication protocol.

This paper studies a new multi-device edge artificial-intelligent (AI) system, which jointly exploits the AI model split inference and integrated sensing and communication (ISAC) to enable low-latency intelligent services at the network edge. In this system, multiple ISAC devices perform radar sensing to obtain multi-view data, and then offload the quantized version of extracted features to a centralized edge server, which conducts model inference based on the cascaded feature vectors. Under this setup and by considering classification tasks, we measure the inference accuracy by adopting an approximate but tractable metric, namely discriminant gain, which is defined as the distance of two classes in the Euclidean feature space under normalized covariance. To maximize the discriminant gain, we first quantify the influence of the sensing, computation, and communication processes on it with a derived closed-form expression. Then, an end-to-end task-oriented resource management approach is developed by integrating the three processes into a joint design. This integrated sensing, computation, and communication (ISCC) design approach, however, leads to a challenging non-convex optimization problem, due to the complicated form of discriminant gain and the device heterogeneity in terms of channel gain, quantization level, and generated feature subsets. Remarkably, the considered non-convex problem can be optimally solved based on the sum-of-ratios method. This gives the optimal ISCC scheme, that jointly determines the transmit power and time allocation at multiple devices for sensing and communication, as well as their quantization bits allocation for computation distortion control. By using human motions recognition as a concrete AI inference task, extensive experiments are conducted to verify the performance of our derived optimal ISCC scheme.

Though achieving marvelous progress in various scenarios, existing semantic communication frameworks mainly consider single-input single-output Gaussian channels or Rayleigh fading channels, neglecting the widely-used multiple-input multiple-output (MIMO) channels, which hinders the application into practical systems. One common solution to combat MIMO fading is to utilize feedback MIMO channel state information (CSI). In this paper, we incorporate MIMO CSI into system designs from a new perspective and propose the learnable CSI fusion semantic communication (LCFSC) framework, where CSI is treated as side information by the semantic extractor to enhance the semantic coding. To avoid feature fusion due to abrupt combination of CSI with features, we present a non-invasive CSI fusion multi-head attention module inside the Swin Transformer. With the learned attention masking map determined by both source and channel states, more robust attention distribution could be generated. Furthermore, the percentage of mask elements could be flexibly adjusted by the learnable mask ratio, which is produced based on the conditional variational interference in an unsupervised manner. In this way, CSI-aware semantic coding is achieved through learnable CSI fusion masking. Experiment results testify the superiority of LCFSC over traditional schemes and state-of-the-art Swin Transformer-based semantic communication frameworks in MIMO fading channels.

Signal detection and modulation classification are two crucial tasks in various wireless communication systems. Different from prior works that investigate them independently, this paper studies the joint signal detection and automatic modulation classification (AMC) by considering a realistic and complex scenario, in which multiple signals with different modulation schemes coexist at different carrier frequencies. We first generate a coexisting RADIOML dataset (CRML23) to facilitate the joint design. Different from the publicly available AMC dataset, ignoring the signal detection step and containing only one signal, our synthetic dataset covers the more realistic multiple-signal coexisting scenario. Then, we present a joint framework for detection and classification (JDM) for such a multiple-signal coexisting environment, which consists of two modules for signal detection and AMC, respectively. In particular, these two modules are interconnected using a designated data structure called “proposal”. Finally, we conduct extensive simulations over the newly developed dataset, which demonstrate the effectiveness of our designs. Our code and dataset are now available as open-source resources athttps://github.com/Singingkettle/ChangShuoRadioData.

Measuring and alleviating the discrepancies between the synthetic (source) and real scene (target) data is the core issue for domain adaptive semantic segmentation. Though recent works have introduced depth information in the source domain to reinforce the geometric and semantic knowledge transfer, they cannot extract the intrinsic 3D information of objects, including positions and shapes, merely based on 2D estimated depth. In this work, we propose a novel Geometry-Aware Network for Domain Adaptation (GANDA), leveraging more compact 3D geometric point cloud representations to shrink the domain gaps. In particular, we first utilize the auxiliary depth supervision from the source domain to obtain the depth prediction in the target domain to accomplish structure-texture disentanglement. Beyond depth estimation, we explicitly exploit 3D topology on the point clouds generated from RGB-D images for further coordinate-color disentanglement and pseudo-labels refinement in the target domain. Moreover, to improve the 2D classifier in the target domain, we perform domain-invariant geometric adaptation from source to target and unify the 2D semantic and 3D geometric segmentation results in two domains. Note that our GANDA is plug-and-play in any existing UDA framework. Qualitative and quantitative results demonstrate that our model outperforms state-of-the-arts on GTA5->Cityscapes and SYNTHIA->Cityscapes.

The mainstream of the existing approaches for video prediction builds up their models based on a Single-In-Single-Out (SISO) architecture, which takes the current frame as input to predict the next frame in a recursive manner. This way often leads to severe performance degradation when they try to extrapolate a longer period of future, thus limiting the practical use of the prediction model. Alternatively, a Multi-In-Multi-Out (MIMO) architecture that outputs all the future frames at one shot naturally breaks the recursive manner and therefore prevents error accumulation. However, only a few MIMO models for video prediction are proposed and they only achieve inferior performance due to the date. The real strength of the MIMO model in this area is not well noticed and is largely under-explored. Motivated by that, we conduct a comprehensive investigation in this paper to thoroughly exploit how far a simple MIMO architecture can go. Surprisingly, our empirical studies reveal that a simple MIMO model can outperform the state-of-the-art work with a large margin much more than expected, especially in dealing with long-term error accumulation. After exploring a number of ways and designs, we propose a new MIMO architecture based on extending the pure Transformer with local spatio-temporal blocks and a new multi-output decoder, namely MIMO-VP, to establish a new standard in video prediction. We evaluate our model in four highly competitive benchmarks. Extensive experiments show that our model wins 1st place on all the benchmarks with remarkable performance gains and surpasses the best SISO model in all aspects including efficiency, quantity, and quality. A dramatic error reduction is achieved when predicting 10 frames on Moving MNIST and Weather datasets respectively. We believe our model can serve as a new baseline to facilitate the future research of video prediction tasks. The code will be released.

3D shape completion from point clouds is a challenging task, especially from scans of real-world objects. Considering the paucity of 3D shape ground truths for real scans, existing works mainly focus on benchmarking this task on synthetic data, e.g. 3D computer-aided design models. However, the domain gap between synthetic and real data limits the generalizability of these methods. Thus, we propose a new task, SCoDA, for the domain adaptation of real scan shape completion from synthetic data. A new dataset, ScanSalon, is contributed with a bunch of elaborate 3D models created by skillful artists according to scans. To address this new task, we propose a novel cross-domain feature fusion method for knowledge transfer and a novel volume-consistent self-training framework for robust learning from real data. Extensive experiments prove our method is effective to bring an improvement of 6%~7% mIoU.

Being data-driven is one of the most iconic properties of deep learning algorithms. The birth of ImageNet [24] drives a remarkable trend of ‘learning from large-scale data’ in computer vision. Pretraining on ImageNet to obtain rich universal representations has been manifested to benefit various 2D visual tasks, and becomes a standard in 2D vision. However, due to the laborious collection of real-world 3D data, there is yet no generic dataset serving as a counterpart of ImageNet in 3D vision, thus how such a dataset can impact the 3D community is unraveled. To remedy this defect, we introduce MVImgNet, a large-scale dataset of multi-view images, which is highly convenient to gain by shooting videos of real-world objects in human daily life. It contains 6.5 million frames from 219,188 videos crossing objects from 238 classes, with rich annotations of object masks, camera parameters, and point clouds. The multi-view attribute endows our dataset with 3D-aware signals, making it a soft bridge between 2D and 3D vision. We conduct pilot studies for probing the potential of MVImgNet on a variety of 3D and 2D visual tasks, including radiance field reconstruction, multi-view stereo, and view-consistent image understanding, where MVImgNet demonstrates promising performance, remaining lots of possibilities for future explorations. Besides, via dense reconstruction on MVImgNet, a 3D object point cloud dataset is derived, called MVPNet, covering 87,200 samples from 150 categories, with the class label on each point cloud. Experiments show that MVP-Net can benefit the real-world 3D object classification while posing new challenges to point cloud understanding. MVImgNet and MVPNet will be public, hoping to inspire the broader vision community.

In this work, we tackle the challenging problem of learning-based single-view 3D hair modeling. Due to the great difficulty of collecting paired real image and 3D hair data, using synthetic data to provide prior knowledge for real domain becomes a leading solution. This unfortunately introduces the challenge of domain gap. Due to the inherent difficulty of realistic hair rendering, existing methods typically use orientation maps instead of hair images as input to bridge the gap. We firmly think an intermediate representation is essential, but we argue that orientation map using the dominant filtering-based methods is sensitive to uncertain noise and far from a competent representation. Thus, we first raise this issue up and propose a novel intermediate representation, termed as HairStep, which consists of a strand map and a depth map. It is found that HairStep not only provides sufficient information for accurate 3D hair modeling, but also is feasible to be inferred from real images. Specifically, we collect a dataset of 1,250 portrait images with two types of annotations. A learning framework is further designed to transfer real images to the strand map and depth map. It is noted that, an extra bonus of our new dataset is the first quantitative metric for 3D hair modeling. Our experiments show that HairStep narrows the domain gap between synthetic and real and achieves state-of-the-art performance on single-view 3D hair reconstruction.

Depth completion plays a crucial role in autonomous driving, in which cameras and LiDARs are two complementary sensors. Recent approaches attempt to exploit spatial geometric constraints hidden in LiDARs to enhance image-guided depth completion. However, only low efficiency and poor generalization can be achieved. In this paper, we propose BEV@DC, a more efficient and powerful multi-modal training scheme, to boost the performance of image-guided depth completion. In practice, the proposed BEV@DC model comprehensively takes advantage of LiDARs with rich geometric details in training, employing an enhanced depth completion manner in inference, which takes only images (RGB and depth) as input. Specifically, the geometric-aware LiDAR features are projected onto a unified BEV space, combining with RGB features to perform BEV completion. By equipping a newly proposed point-voxel spatial propagation network (PV-SPN), this auxiliary branch introduces strong guidance to the original image branches via 3D dense supervision and feature consistency. As a result, our baseline model demonstrates significant improvements with the sole image inputs. Concretely, it achieves state-of-the-art on several benchmarks, e.g., ranking Top-1 on the challenging KITTI depth completion benchmark.

Guangyi Liu, Zeyu Feng, Yuan Gao, Zichao Yang, Xiaodan Liang, Junwei Bao, Xiaodong He, Shuguang Cui, Zhen Li, Zhiting Hu. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023.

This paper studies a multi-intelligent-reflecting-surface (IRS)-enabled integrated sensing and communications (ISAC) system, in which multiple IRSs are installed to help a base station (BS) provide ISAC services at the line-of-sight (LoS) blocked areas. In particular, we consider the case with semi-passive uniform linear array (ULA) IRSs each integrated with dedicated sensors for receiving echo signals, in which each IRS simultaneously senses one point target and communicates with one communication user (CU) within its coverage area. Under this setup, we first derive the closed-form Cramér-Rae bound (CRB) for the targets' direction-of-arrival (DoA) estimation at the corresponding IRSs. Then, to achieve fair and optimal sensing performance, we minimize the maximum CRB for targets' DoA estimation at all IRSs, by jointly optimizing the transmit beamformers at the BS and the reflective beamformers at the IRSs, subject to the minimum signal-to-interference-plus-noise ratio (SINR) constraints at individual CUs, the maximum transmit power constraint at the BS, and the unit-modulus constraints at the IRSs. To tackle the highly non-convex SINR-constrained max-CRB minimization problem, we propose an efficient algorithm based on alternating optimization and semi-definite relaxation, to obtain a converged solution. Finally, numerical results are provided to verify the effectiveness of our proposed design over various benchmark schemes based on separate or heuristic beamforming designs.

To support future 6G mobile applications, the mobile edge computing (MEC) network needs to be jointly optimized for computing, pushing, and caching to reduce transmission load and computation cost. To achieve this, we propose a framework based on deep reinforcement learning that enables the dynamic orchestration of these three activities for the MEC network. The framework can implicitly predict user future requests using deep networks and push or cache the appropriate content to enhance performance. To address the curse of dimensionality resulting from considering three activities collectively, we adopt the soft actor-critic reinforcement learning in continuous space and design the action quantization and correction specifically to fit the discrete optimization problem. We conduct simulations in a single-user single-server MEC network setting and demonstrate that the proposed framework effectively decreases both transmission load and computing cost under various configurations of cache size and tolerable service delay.

Remote zero-shot object recognition, i.e., offloading zero-shot object recognition tasks from one mobile device to remote mobile edge computing (MEC) server or another mobile device, has become a common and important task to conquer for 6G. With this goal, this paper first establishes a zero-shot multi-level feature extractor, which projects the image into the visual, semantic, as well as intermediate feature space in a lightweight way. Then, this paper proposes a novel multi-level feature transmission framework powered by a semantic knowledge base (SKB), and characterizes the semantic loss and required transmission latency at each level. Under this setup, this paper formulates the multi-level feature transmission optimization problem to minimize the semantic loss under the end-to-end latency constraint. Such a problem, however, is a multi-choice knapsack problem, and thus very difficult to solve. To resolve this issue, this paper proposes an efficient algorithm based on the convex concave procedure to find an efficient solution. Numerical results show that the proposed design outperforms the benchmarks, and illustrate the tradeoff between the transmission latency and zero-shot classification accuracy, as well as the effects of the SKBs at both the transmitter and receiver on classification accuracy.

Gaussian process state-space model (GPSSM) is a fully probabilistic state-space model that has attracted much attention over the past decade. However, the outputs of the transition function in the existing GPSSMs are assumed to be independent, meaning that the GPSSMs cannot exploit the inductive biases between different outputs and lose certain model capacities. To address this issue, this paper proposes an output-dependent and more realistic GPSSM by utilizing the well-known, simple yet practical linear model of coregionalization (LMC) framework to represent the output dependency. To jointly learn the output-dependent GPSSM and infer the latent states, we propose a variational sparse GP-based learning method that only gently increases the computational complexity. Experiments on both synthetic and real datasets demonstrate the superiority of the output-dependent GPSSM in terms of learning and inference performance.

Sensing, computation, and communication (SC2) are highly coupled processes in federated edge learning (FEEL) and need to be jointly designed in a task-oriented manner for pursuing the best FEEL performance under the stringent resource constraints at edge devices. However, this remains an open problem as there is a lack of theoretical understanding on how the SC2resources jointly affect the FEEL performance. In this paper, we address the problem of joint SC2resource allocation for FEEL via a concrete case study of human motion recognition based on wireless sensing. Specifically, the joint SC2resource allocation problem is cast to maximize the convergence speed of FEEL, under the constraints on training time and energy supply of each edge device. Solving this problem entails solving two subproblems in order: the first one reduces to determining a joint sensing and communication resource allocation that maximizes the total number of samples sensed during the entire training process; the second one concerns the partition of the total number of sensed samples over communication rounds to determine the batch size at each round for convergence speed maximization. Finally, extensive simulation results are provided to validate the superiority of the proposed scheme over several baseline schemes.

This paper studies a new multi-device edge artificial-intelligent (AI) system, which jointly exploits the AI model split inference and integrated sensing and communication (ISAC) to enable low-latency intelligent services at the network edge. In this system, multiple ISAC devices perform radar sensing to obtain multi-view data, and then offload the quantized version of extracted features to a centralized edge server, which conducts model inference based on the cascaded feature vectors. Under this setup and by considering classification tasks, we measure the inference accuracy by adopting an approximate but tractable metric, namely discriminant gain, which is defined as the distance of two classes in the Euclidean feature space under normalized covariance. To maximize the discriminant gain, we first quantify the influence of the sensing, computation, and communication processes on it with a derived closed-form expression. Then, an end-to-end task-oriented resource management approach is developed by designing an optimal integrated sensing, computation, and communication (ISCC) scheme. By using human motions recognition as a concrete AI inference task, extensive experiments are conducted to verify the performance of the proposed scheme.

3D lane detection from monocular images is a fundamental yet challenging task in autonomous driving. Recent advances primarily rely on structural 3D surrogates (e.g., bird’s eye view) built from front-view image features and camera parameters. However, the depth ambiguity in monocular images inevitably causes misalignment between the constructed surrogate feature map and the original image, posing a great challenge for accurate lane detection. To address the above issue, we present a novel LATR model, an end-to-end 3D lane detector that uses 3D-aware front-view features without transformed view representation. Specifically, LATR detects 3D lanes via cross-attention based on query and key-value pairs, constructed using our lane-aware query generator and dynamic 3D ground positional embedding. On the one hand, each query is generated based on 2D lane-aware features and adopts a hybrid embedding to enhance the lane information. On the other hand, 3D space information is injected as positional embedding from an iteratively-updated 3D ground plane. LATR outperforms previous state-of-the-art methods on both synthetic Apollo and realistic OpenLane and ONCE-3DLanes by large margins (e.g., 11.4 gain in terms of F1 score on OpenLane). Code will be released at https://github.com/JMoonr/LATR.

Recent work on Neural Radiance Fields (NeRF) has demonstrated significant advances in high-quality view synthesis. A major limitation of NeRF is its low rendering efficiency due to the need for multiple network forwardings to render a single pixel. Existing methods to improve NeRF either reduce the number of required samples or optimize the implementation to accelerate the network forwarding. Despite these efforts, the problem of multiple sampling persists due to the intrinsic representation of radiance fields. In contrast, Neural Light Fields (NeLF) reduce the computation cost of NeRF by querying only one single network forwarding per pixel. To achieve a close visual quality to NeRF, existing NeLF methods require significantly larger network capacities which limits their rendering efficiency in practice. In this work, we propose a new representation called Neural Radiance Distribution Field (NeRDF) that targets efficient view synthesis in real-time. Specifically, we use a small network similar to NeRF while preserving the rendering speed with a single network forwarding per pixel as in NeLF. The key is to model the radiance distribution along each ray with frequency basis and predict frequency weights using the network. Pixel values are then computed via volume rendering on radiance distributions. Experiments show that our proposed method offers a better trade-off among speed, quality, and network size than existing methods: we achieve a ∼254× speed-up over NeRF with similar network size, with only a marginal performance decline. Our project page is at yushuang-wu.github.io/NeRDF.

Fast generation of high-quality 3D digital humans is important to a vast number of applications ranging from entertainment to professional concerns. Recent advances in differentiable rendering have enabled the training of 3D generative models without requiring 3D ground truths. However, the quality of the generated 3D humans still has much room to improve in terms of both fidelity and diversity. In this paper, we present Get3DHuman, a novel 3D human framework that can significantly boost the realism and diversity of the generated outcomes by only using a limited budget of 3D ground-truth data. Our key observation is that the 3D generator can profit from human-related priors learned through 2D human generators and 3D reconstructors. Specifically, we bridge the latent space of Get3DHuman with that of StyleGAN-Human [13] via a specially-designed prior network, where the input latent code is mapped to the shape and texture feature volumes spanned by the pixel-aligned 3D reconstructor [50]. The outcomes of the prior network are then leveraged as the supervisory signals for the main generator network. To ensure effective training, we further propose three tailored losses applied to the generated feature volumes and the intermediate feature maps. Extensive experiments demonstrate that Get3DHuman greatly outperforms the other state-of-the-art approaches and can support a wide range of applications including shape interpolation, shape re-texturing, and single-view reconstruction through latent inversion.

State-of-the-art multi-agent collision avoidance algorithms face limitations when applied to cluttered public environments, where obstacles may have a variety of shapes and structures. The issue arises because most of these algorithms are agent-level methods. They concentrate solely on preventing collisions between the agents while the obstacles are handled merely out-of-policy. Obstacle-aware policies output an action considering both agents and obstacles. Current obstacle-aware algorithms, mainly based on Lidar sensor data, struggle to handle collision avoidance around complex obstacles. To resolve this issue, this paper investigates how to find a better way to travel around diverse obstacles. In particular, we present a global map assisted collision avoidance algorithm which, following the lead of a high-level goal guide and using an obstacle representation called distance map, considers other agents and obstacles simultaneously. Moreover, our model can be loaded into each agent individually, making it applicable to large maps or more agents. Simulation results indicate that our model outperforms the state-of-the-art algorithms, showing in scenarios with complex obstacles. We present a notion for incorporating global information in decentralized decision-making, along with a method for extending agent-level algorithms to cluttered environments in real-world scenarios.

Point clouds produced by 3D scanning are typically sparse, non-uniform, and noisy. Existing upsampling techniques directly learn the mapping from a sparse point set to a dense point set, which is often under-determined and ill-posed. To reduce the uncertainty and ambiguity of the upsampling mapping, this paper proposes a generic three-stage vector-quantization framework, which incorporates a Codebook lookup Transformer and knowledge distillation for Point Cloud Upsampling, named CPU. The proposed CPU reformulates the upsampling task into a relatively determinate code prediction task within a small, discrete proxy space. Since the traditional vector-quantization methods cannot be directly applied to point cloud upsampling scenarios, we introduce a knowledge distillation training scheme that facilitates efficient codebook learning and ensures full utilization of codebook entries. Specifically, we adopt a teacher-student training paradigm to avoid model collapse during codebook learning. In the first stage, we pre-train a vanilla auto-encoder of the dense point set as the teacher model, which provides rich guidance features to ensure sufficient codebook learning. In the second stage, we train a vector-quantized auto-encoder as a student model to capture high-fidelity geometric priors into a learned codebook with the aid of distillation. In the third stage, we propose a Codebook Lookup Transformer to model the global context of the sparse point set and predict the code indices. Then the coarse features of the sparse point set can be quantized and substituted by looking up the indices in the learned codebook. Benefiting from the expressive codebook priors and the distillation training scheme, the proposed CPU outperforms state-of-the-art methods quantitatively and qualitatively.

Volumetric video (VV) recently emerges as a new form of video application providing a photorealistic immersive 3D viewing experience with 6 degree-of-freedom (DoF), which empowers many applications such as VR, AR, and Metaverse. A key problem therein is how to stream the enormous size VV through the network with limited bandwidth. Existing works mostly focused on predicting the viewport for a tiling-based adaptive VV streaming, which however only has quite a limited effect on resource saving. We argue that the content repeatability in the viewport can be further leveraged, and for the first time, propose a client-side cache-assisted strategy that aims to buffer the repeatedly appearing VV tiles in the near future so as to reduce the redundant VV content transmission. The key challenges exist in three aspects, including (1) feature extraction and mining in 6 DoF VV context, (2) accurate long-term viewing pattern estimation and (3) optimal caching scheduling with limited capacity. In this paper, we propose CaV3, an integrated cache-assisted viewport adaptive VV streaming framework to address the challenges. CaV3 employs a Long-short term Sequential prediction model (LSTSP) that achieves accurate short-term, mid-term and long-term viewing pattern prediction with a multi-modal fusion model by capturing the viewer's behavior inertia, current attention, and subjective intention. Besides, CaV3 also contains a contextual MAB-based caching adaptation algorithm (CCA) to fully utilize the viewing pattern and solve the optimal caching problem with a proved upper bound regret. Compared to existing VV datasets only containing single or co-located objects, we for the first time collect a comprehensive dataset with sufficient practical unbounded 360° scenes. The extensive evaluation of the dataset confirms the superiority of CaV3, which outperforms the SOTA algorithm by 15.6%-43% in viewport prediction and 13%-40% in system utility.

Semantic segmentation of point cloud usually relies on dense annotation that is exhausting and costly, so it attracts wide attention to investigate solutions for the weakly supervised scheme with only sparse points annotated. Existing works start from the given labels and propagate them to highly-related but unlabeled points, with the guidance of data, e.g. intra-point relation. However, it suffers from (i) the inefficient exploitation of data information, and (ii) the strong reliance on labels thus is easily suppressed when given much fewer annotations. Therefore, we propose a novel framework, PointMatch, that stands on both data and label, by applying consistency regularization to sufficiently probe information from data itself and leveraging weak labels as assistance at the same time. By doing so, meaningful information can be learned from both data and label for better representation learning , which also enables the model more robust to the extent of label sparsity . Simple yet effective, the proposed PointMatch achieves the state-of-the-art performance under various weakly-supervised schemes on both ScanNet-v2 and S3DIS datasets, especially on the settings with extremely sparse labels, e.g. surpassing SQN by 21.2% and 17.2% on the 0.01% and 0.1% setting of ScanNet-v2, respectively.

As one of the most important manifestations of virtual reality (VR), 360° panoramic videos in recent years have experienced booming development due to the desire for immersive and interactive experiences. Compared to traditional videos, 360° videos are featured with uncertain user Field of View (FoV), more sensitive delay tolerance, and much higher bandwidth requirement, bringing unprecedented challenges to 360° video streaming. Meanwhile, the development of 5G and mobile edge computing starts to pave the way for high-bandwidth low-latency video streaming. Some preliminary works focus on either individual FoV prediction or multiuser Quality of Experience (QoE) oriented cache strategy design, while how to design a holistic solution toward optimizing the overall user QoE with considerations over fairness and long-term system cost remains a nontrivial problem. In this article, we proposeEbublio, a novel intelligent edge caching framework to address the aforementioned challenges in 360° video streaming.Ebublioconsists of a collaborative FoV prediction (CFP) module and a long-term tile caching optimization (LTO) module to jointly optimize the long-term user QoE and system cost. The former module integrates the features of video content, user trajectory, and other users’ records for combined prediction. The latter one employs the Lyapunov framework and a subgradient optimization approach toward the optimal caching replacement policy. Our trace-driven evaluation demonstrates the superiority of our framework, with about 42% improvement in FoV prediction, and 36% improvement in QoE at similar traffic consumption.

Federated learning (FL) emerges as a new distributed machine learning (ML) paradigm that enables thousands of mobile devices to collaboratively train ML models using local data without compromising user privacy. However, the FL learning quality highly relies on the data contribution from the distributed mobile devices. Therefore, a well-designed incentive mechanism with effectiveness, fairness, and reciprocity is in urgent need to guarantee the stable participation of users. In this article, we propose federated auction bandit (FedAB), an incentive and client selection strategy based on a novel multiattribute reverse auction mechanism and a combinatorial multi-armed bandit (CMAB) algorithm. First, we develop a local contribution evaluation method based on importance sampling in the FL context. We then design a novel payment mechanism that is able to preserve individual rationality and incentive compatibility (truthfulness). At last, we design a UCB-based winner selection algorithm that is proven to achieve the server’s utility maximization with fairness and reciprocity. We have conducted extensive experiments on real data sets. The results demonstrate the superiority ofFedAB, with a 10%–50% improvement in total reward, final accuracy, and convergence speed compared to state-of-the-art solutions.

Federated learning (FL) over mobile devices has fostered numerous intriguing applications/services, many of which are delay-sensitive. In this paper, we propose a service delay efficient FL (SDEFL) scheme over mobile devices. Unlike traditional communication efficient FL, which regards wireless communications as the bottleneck, we find that under many situations, the local computing delay is comparable to the communication delay during the FL training process, given the development of high-speed wireless transmission techniques. Thus, the service delay in FL should be computing delay + communication delay over training rounds. To minimize the service delay of FL, simply reducing local computing/communication delay independently is not enough. The delay trade-off between local computing and wireless communications must be considered. Besides, we empirically study the impacts of local computing control and compression strategies (i.e., the number of local updates, weight quantization, and gradient quantization) on computing, communication and service delays. Based on those trade-off observation and empirical studies, we develop an optimization scheme to minimize the service delay of FL over heterogeneous devices. We establish testbeds and conduct extensive emulations/experiments to verify our theoretical analysis. The results show that SDEFL reduces notable service delay with a small accuracy drop compared to peer designs.

Existing localization methods that intensively leverage the environment-specific received signal strength (RSS) or channel state information (CSI) of wireless signals are rather accurate in certain environments. However, these methods, whether based on pure statistical signal processing or data-driven approaches, often struggle to generalize to new environments, which results in considerable time and effort being wasted. To address this challenge, we propose MetaLoc, which is the first fingerprinting-based localization framework that leverages the Model-Agnostic Meta-Learning (MAML). Specifically, built on a deep neural network with strong representation capabilities, MetaLoc is trained on historical data sourced from well-calibrated environments, employing a two-loop optimization mechanism to obtain the meta-parameters. These meta-parameters act as the initialization for quick adaptation in new environments, reducing the need for much human effort. The framework introduces two paradigms for the optimization of meta-parameters: a centralized paradigm that simplifies the process by sharing data from all historical environments, and a distributed paradigm that maintains data privacy by training meta-parameters for each specific environment separately. Furthermore, the advanced distributed paradigm modifies the vanilla MAML loss function to ensure that the reduction of loss occurs in a consistent direction across various training domains, thus facilitating faster convergence during training. Our experiments on both synthetic and real datasets demonstrate that MetaLoc outperforms baseline methods in terms of localization accuracy, robustness, and cost-effectiveness. The code and datasets used in this study are publicly available at:https://github.com/WU-Dongze/MetaLoc.

In this paper, we formulate and solve a two-stage Bayesian sequential change diagnosis (SCD) problem in a multi-sensor setting. In the considered problem, a change first occurs at a sensor and then propagates across the sensor array gradually. After a change is detected, we are allowed to continue observing more samples so that we can identify the distribution after the change more accurately. Our goal is to minimize the total cost including delay, false alarm, and misdiagnosis probabilities. We first characterize the optimal SCD rule. Moreover, to address the high computational complexity issue of the optimal SCD rule, we propose a low-complexity threshold SCD rule. We further analyze the asymptotic optimality of the threshold SCD rule. In addition, we investigate how increasing the number of sensors can improve the performance of the proposed threshold SCD rule. Our analysis holds for different sensor array structures, including linear sensor arrays and 2D lattice sensor arrays.

This paper studies the feasibility of deploying intelligent reflecting surfaces (IRSs) in massive multiple input multiple-output (MIMO) systems to improve the performance of users in the service dead zone. One question of paramount importance is as follows: if the overhead of channel training and the computational complexity of algorithm design arising from the huge number of IRS reflecting elements and base station (BS) antennas have to be controlled, can we provide reasonable performance to the users with weak direct channels? This paper provides an affirm answer to this question. Specifically, to reduce the channel training overhead, we consider an appealing protocol for the uplink communication in the IRS-assisted massive MIMO systems. Under this protocol, the IRS reflection coefficients are optimized based on the channel covariance matrices, which are generally fixed for many coherence blocks, to boost the long-term performance. Then, given the IRS reflecting coefficients, the BS beamforming vectors are designed in each coherence block based on the effective channel of each user, which is the superposition of its direct and reflected user-IRS-BS channels, to improve the instantaneous performance. Since merely the user effective channels are estimated in each coherence block, the training overhead of this protocol is the same as that in the legacy wireless systems without IRSs. Moreover, in the asymptotic regime that the numbers of IRS elements and BS antennas both go to infinity with a fixed ratio, we manage to first characterize the minimum mean-squared error (MMSE) estimators of the user effective channels and then quantify the closed-form user achievable rates as functions of channel covariance matrices with channel training overhead and estimation error taken into account. Interestingly, it is shown that the properties of channel hardening and favorable propagation still hold for the user effective channels, and satisfactory user rates are thus achievable even if simple BS beamforming solutions, e.g., maximal-ratio combining, are employed. Finally, thanks to the rate characterization, we design a low-complexity algorithm to optimize the IRS reflection coefficients based on channel covariance matrices.

Accurate protein contact map prediction (PCMP) is essential for precise protein structure estimation and further biological studies. Recent works achieve significant performance on this task with high quality multiple sequence alignment (MSA). However, the PCMP accuracy drops dramatically while only poor MSA (e.g., absolute MSA count less than 10) is available. Therefore, in this paper, we propose the Contact-Distil to improve the low homologous PCMP accuracy through knowledge distillation on a self-supervised model. Particularly, two pre-trained transformers are exploited to learn the high quality and low quality MSA representation in parallel for the teacher and student model correspondingly. Besides, the co-evolution information is further extracted from pure sequence through a pretrained ESM-1b model, which provides auxiliary knowledge to improve student performance. Extensive experiments show Contact-Distil outperforms previous state-of-the-arts by large margins on CAMEO-L dataset for low homologous PCMP, i.e., around 13.3% and 9.5% improvements against Alphafold2 and MSA Transformer respectively when MSA count less than 10.

Waste inspection for packaged waste is an important step in the pipeline of waste disposal. Previous methods either rely on manual visual checking or RGB image-based inspection algorithm, requiring costly preparation procedures (e.g., open the bag and spread the waste items). Moreover, occluded items are very likely to be left out. Inspired by the fact that X-ray has a strong penetrating power to see through the bag and overlapping objects, we propose to perform waste inspection efficiently using X-ray images without the need to open the bag. We introduce a novel problem of instance-level waste segmentation in X-ray image for intelligent waste inspection, and contribute a real dataset consisting of 5,038 X-ray images (totally 30,881 waste items) with high-quality annotations (i.e., waste categories, object boxes, and instance-level masks) as a benchmark for this problem. As existing segmentation methods are mainly designed for natural images and cannot take advantage of the characteristics of X-ray waste images (e.g., heavy occlusions and penetration effect), we propose a new instance segmentation method to explicitly take these image characteristics into account. Specifically, our method adopts an easy-to-hard disassembling strategy to use high confidence predictions to guide the segmentation of highly overlapped objects, and a global structure guidance module to better capture the complex contour information caused by the penetration effect. Extensive experiments demonstrate the effectiveness of the proposed method. Our dataset is released at WIXRayNet.

Automatic tooth instance segmentation on 3D dental models is a fundamental task for computer-aided orthodontic treatments. Existing learning-based methods rely heavily on expensive point-wise annotations. To alleviate this problem, we are the first to explore a low-cost annotation way for 3D tooth instance segmentation, i.e., labeling all tooth centroids and only a few teeth for each dental model. Regarding the challenge when only weak annotation is provided, we present a dental arch prior-assisted 3D tooth segmentation method, namely DArch. Our DArch consists of two stages, including tooth centroid detection and tooth instance segmentation. Accurately detecting the tooth centroids can help locate the individual tooth, thus benefiting the segmentation. Thus, our DArch proposes to leverage the dental arch prior to assist the detection. Specifically, we firstly propose a coarse-to-fine method to estimate the dental arch, in which the dental arch is initially generated by Bezier curve regression, and then a graph-based convolutional network (GCN) is trained to refine it. With the estimated dental arch, we then propose a novel Arch-aware Point Sampling (APS) method to assist the tooth centroid proposal generation. Meantime, a segmentor is independently trained using a patch-based training strategy, aiming to segment a tooth instance from a 3D patch centered at the tooth centroid. Experimental results on 4, 773 dental models have shown our DArch can accurately segment each tooth of a dental model, and its performance is superior to the state-of-the-art methods.

3D dense captioning aims to describe individual objects in 3D scenes by natural language, where 3D scenes are usually represented as RGB-D scans or point clouds. However, only exploiting single modal information, e.g., point cloud, previous approaches fail to produce faithful descriptions. Though aggregating 2D features into point clouds may be beneficial, it introduces an extra computational burden, especially in the inference phase. In this study, we investigate a cross-modal knowledge transfer using Transformer for 3D dense captioning, namely X-Trans2Cap. Our proposed X-Trans2Cap effectively boost the performance of single-modal 3D captioning through the knowledge distillation enabled by a teacher-student framework. In practice, during the training phase, the teacher network exploits auxiliary 2D modality and guides the student network that only takes point clouds as input through the feature consistency constraints. Owing to the well-designed cross-modal feature fusion module and the feature alignment in the training phase, X-Trans2Cap acquires rich appearance information embedded in 2D images with ease. Thus, a more faithful caption can be generated only using point clouds during the inference. Qualitative and quantitative results confirm that X-Trans2Cap outperforms previous state-of-the-art by a large margin, i.e., about +21 and +16 CIDEr points on ScanRefer and Nr3D datasets, respectively.

3D single object tracking (3D SOT) in LiDAR point clouds plays a crucial role in autonomous driving. Current approaches all follow the Siamese paradigm based on appearance matching. However, LiDAR point clouds are usually textureless and incomplete, which hinders effective appearance matching. Besides, previous methods greatly overlook the critical motion clues among targets. In this work, beyond 3D Siamese tracking, we introduce a motion-centric paradigm to handle 3D SOT from a new perspective. Following this paradigm, we propose a matching-free two-stage tracker M2-Track. At the 1st-stage, M2-Track localizes the target within successive frames via motion transformation. Then it refines the target box through motion-assisted shape completion at the 2nd-stage. Extensive experiments confirm that M2-Track significantly outperforms previous state-of-the-arts on three large-scale datasets while running at 57FPS (~ 8%, ~ 17% and ~ 22% precision gains on KITTI, NuScenes, and Waymo Open Dataset respectively). Further analysis verifies each component's effectiveness and shows the motioncentric paradigm's promising potential when combined with appearance matching. Code will be made available at https://github.com/Ghostish/Open3DSOT.

We investigate the coexistence of task-oriented and data-oriented communications in a IoT system that shares a group of channels, and study the scheduling problem to jointly optimize the weighted age of incorrect information (AoII) and throughput, which are the performance metrics of the two types of communications, respectively. This problem is formulated as a Markov decision problem, which is difficult to solve due to the large discrete action space and the time-varying action constraints induced by the stochastic availability of channels. By exploiting the intrinsic properties of this problem, we first simplify it and convert it to an equivalent Markov game, for which the large and discrete action space issue is greatly relieved. Then, we propose a Whittle's index guided multi-agent proximal policy optimization (WI-MAPPO) algorithm to solve the considered game, where the embedded Whittle's index module further shrinks the action space, and the proposed offline training algorithm extends the training kernel of the conventional MAPPO to address the issue of time-varying constraints.

This paper considers improving wireless communication and computation efficiency in federated learning (FL) via model quantization. In the proposed bitwidth FL scheme, edge devices train and transmit quantized versions of their local FL model parameters to a coordinating server, which, in turn, aggregates them into a quantized global model and synchronizes the devices. With the goal of jointly determining the set of participating devices in each training iteration and the bitwidths employed at the devices, we pose an optimization problem for minimizing the training loss of quantized FL under a device sampling budget and delay requirement. Our analytical results show that the improvement of FL training loss between two consecutive iterations depends on not only the device selection and quantization scheme, but also on several parameters inherent to the model being learned. As a result, we propose, a model-based reinforcement learning (RL) method to optimize action selection over iterations. Compared to model-free RL, the proposed approach leverages the derived mathematical characterization of the FL training process to discover an effective device selection and quantization scheme without imposing additional device communication overhead. Numerical evaluations show that the proposed FL framework can achieve the same classification performance while reducing the number of training iterations needed for convergence by 20% compared to model-free RL-based FL.

Multitask Gaussian process (MTGP) is powerful for joint learning of multiple tasks with complicated correlation patterns. However, due to the assembling of additive independent latent functions (LFs), all current MTGPs including the salient linear model of coregionalization (LMC) and convolution frameworks cannot effectively represent and learn the hierarchical latent interactions between its LFs. In this paper, we further investigate the interactions in LMC of MTGP and then propose a novel kernel representation of the hierarchical interactions, which ameliorates both the expressiveness and the interpretability of MTGP. Specifically, we express the interaction as a product of function interaction (FI) and coefficient interaction. The FI is modeled by using cross convolution of LFs. The coefficient interaction between the LMCs is described as a free-form coupling coregionalization term. We validate that considering the interactions can promote knowledge transferring in MTGP and compare our approach with some state-of-the-art MTGPs on both synthetic-and real-world datasets.

This paper studies a mobile edge computing (MEC) network to support emerging Internet-of-things applications, where multiple access points (APs), each attached with an MEC server, need to collect data from multiple sensors, process them, and then send computation results to the paired actuators for control. Specifically, we consider a three-phase operation protocol for data uploading, edge computing, and results downloading, where the frequency-division multiple access is implemented to accommodate communications of multiple sensors/actuators. Under this setup, we minimize the end-to-end (E2E) latency of the sensing-communication-computation-actuation loop by properly designing the user association and resource allocation policy, subject to the communication and computation resource constraints. The formulated problem, however, is a mixed-integer non-linear program that is difficult to be optimized. Despite this fact, we first search the optimal solution for user association, and then apply convex optimization for resource allocation given the obtained optimal user association. Finally, numerical results show that the proposed optimal joint design significantly reduces the E2E latency, as compared to conventional designs without such joint designs.

Semi-supervised domain adaptation (SSDA) aims to apply knowledge learned from a fully labeled source domain to a scarcely labeled target domain. In this paper, we propose a Multi-level Consistency Learning (MCL) framework for SSDA. Specifically, our MCL regularizes the consistency of different views of target domain samples at three levels: (i) at inter-domain level, we robustly and accurately align the source and target domains using a prototype-based optimal transport method that utilizes the pros and cons of different views of target samples; (ii) at intra-domain level, we facilitate the learning of both discriminative and compact target feature representations by proposing a novel class-wise contrastive clustering loss; (iii) at sample level, we follow standard practice and improve the prediction accuracy by conducting a consistency-based self-training. Empirically, we verified the effectiveness of our MCL framework on three popular SSDA benchmarks, i.e., VisDA2017, DomainNet, and Office-Home datasets, and the experimental results demonstrate that our MCL framework achieves the state-of-the-art performance.

360° videos in recent years have experienced booming development. Compared to traditional videos, 360° videos are featured with uncertain user behaviors, bringing opportunities as well as challenges. Datasets are necessary for researchers and developers to explore new ideas and conduct reproducible analyses for fair comparisons among different solutions. However, existing related datasets mostly focused on users' field of view (FoV), ignoring the more important eye gaze information, not to mention the integrated extraction and analysis of both FoV and eye gaze. Besides, users' behavior patterns are highly related to videos, yet most existing datasets only contained videos with subjective and qualitative classification from video genres, which lack quantitative analysis and fail to characterize the intrinsic properties of a video scene. To this end, we first propose a quantitative taxonomy for 360° videos that contains three objective technical metrics. Based on this taxonomy, we collect a dataset containing users' head and gaze behaviors simultaneously, which outperforms existing datasets with rich dimensions, large scale, strong diversity, and high frequency. Then we conduct a pilot study on users' behaviors and get some interesting findings such as user's head direction will follow his/her gaze direction with the most possible time interval. A case of application in tile-based 360° video streaming based on our dataset is later conducted, demonstrating a great performance improvement of existing works by leveraging our provided gaze information. Our dataset is available at https://cuhksz-inml.github.io/head_gaze_dataset/

We present PVSeRF, a learning framework that reconstructs neural radiance fields from single-view RGB images, for novel view synthesis. Previous solutions, such as pixelNeRF, rely only on pixel-aligned features and suffer from feature ambiguity issues. As a result, they struggle with the disentanglement of geometry and appearance, leading to implausible geometries and blurry results. To address this challenge, we propose to incorporate explicit geometry reasoning and combine it with pixel-aligned features for radiance field prediction. Specifically, in addition to pixel-aligned features, we further constrain the radiance field learning to be conditioned on i) voxel-aligned features learned from a coarse volumetric grid and ii) fine surface-aligned features extracted from a regressed point cloud. We show that the introduction of such geometry-aware features helps to achieve a better disentanglement between appearance and geometry, i.e. recovering more accurate geometries and synthesizing higher quality images of novel views. Extensive experiments against state-of-the-art methods on ShapeNet benchmarks demonstrate the superiority of our approach for single-image novel view synthesis.

Guangyi Liu, Zichao Yang, Tianhua Tao, Xiaodan Liang, Junwei Bao, Zhen Li, Xiaodong He, Shuguang Cui, Zhiting Hu. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2022.

Although recent point cloud analysis achieves impressive progress, the paradigm of representation learning from single modality gradually meets its bottleneck. In this work, we take a step towards more discriminative 3D point cloud representation using 2D images, which inherently contain richer appearance information, e.g., texture, color, and shade. Specifically, this paper introduces a simple but effective point cloud cross-modality training (PointCMT) strategy, which utilizes view-images, i.e., rendered or projected 2D images of the 3D object, to boost point cloud classification. In practice, to effectively acquire auxiliary knowledge from view-images, we develop a teacher-student framework and formulate the cross-modal learning as a knowledge distillation problem. Through novel feature and classifier enhancement criteria, PointCMT eliminates the distribution discrepancy between different modalities and avoid potential negative transfer effectively. Note that PointCMT efficiently improves the point-only representation without any architecture modification. Sufficient experiments verify significant gains on various datasets based on several backbones, i.e., equipped with PointCMT, PointNet++ and PointMLP achieve state-of-the-art performance on two benchmarks, i.e., 94.4% and 86.7% accuracy on ModelNet40 and ScanObjectNN, respectively.

MOTIVATION: Protein secondary structure prediction (PSSP) is one of the fundamental and challenging problems in the field of computational biology. Accurate PSSP relies on sufficient homologous protein sequences to build the multiple sequence alignment (MSA). Unfortunately, many proteins lack homologous sequences, which results in the low quality of MSA and poor performance. In this article, we propose the novel dynamic scoring matrix (DSM)-Distil to tackle this issue, which takes advantage of the pretrained BERT and exploits the knowledge distillation on the newly designed DSM features. Specifically, we propose the DSM to replace the widely used profile and PSSM (position-specific scoring matrix) features. DSM could automatically dig for the suitable feature for each residue, based on the original profile. Namely, DSM-Distil not only could adapt to the low homologous proteins but also is compatible with high homologous ones. Thanks to the dynamic property, DSM could adapt to the input data much better and achieve higher performance. Moreover, to compensate for low-quality MSA, we propose to generate the pseudo-DSM from a pretrained BERT model and aggregate it with the original DSM by adaptive residue-wise fusion, which helps to build richer and more complete input features. In addition, we propose to supervise the learning of low-quality DSM features using high-quality ones. To achieve this, a novel teacher-student model is designed to distill the knowledge from proteins with high homologous sequences to that of low ones. Combining all the proposed methods, our model achieves the new state-of-the-art performance for low homologous proteins. RESULTS: Compared with the previous state-of-the-art method 'Bagging', DSM-Distil achieves an improvement about 5% and 7.3% improvement for proteins with MSA count ≤30 and extremely low homologous cases, respectively. We also compare DSM-Distil with Alphafold2 which is a state-of-the-art framework for protein structure prediction. DSM-Distil outperforms Alphafold2 by 4.1% on extremely low-quality MSA on 8-state secondary structure prediction. Moreover, we release a large-scale up-to-date test dataset BC40 for low-quality MSA structure prediction evaluation. AVAILABILITY AND IMPLEMENTATION: BC40 dataset: https://drive.google.com/drive/folders/15vwRoOjAkhhwfjDk6-YoKGf4JzZXIMC. HardCase dataset: https://drive.google.com/drive/folders/1BvduOr2b7cObUHy6GuEWk-aUkKJgzTUv. Code: https://github.com/qinwang-ai/DSM-Distil.

The rapidly developing energy harvesting (EH) technology is a promising solution to the durability issue in the battery-powered Internet of Things (IoT) systems. In this article, underlaid device-to-device (D2D) transmission powered by radio signals harvested from cellular systems is studied. By considering the dilemmas among EH, D2D transmission opportunity, and interference management, we propose two transmission policies: 1) Policy 1 requires that the available power in the battery should be no less than the D2D transmission power and 2) Policy 2 not only sets the constraint on available power but also introduces theguard zonerule to protect D2D transmissions from severe interference. The employment of aguard zonein this situation is technically challenging since the original distribution of energy arrival will thereby be changed. We derive expressions in closed or semiclosed forms for the considered D2D transmission performance metrics with the stochastic geometry framework and Poisson hole process. With numerical simulation results, the influences of varying network parameters on D2D performances are illustrated. The results show that by introducing a guard zone, the D2D successful transmission rate can be increased by 41.2%. All the developed D2D frameworks and the summarized useful remarks are used to provide meaningful design insights and guidelines for the deployment strategies of EH-based D2D wireless networks.

This article proposes a hybrid dimming (HD) scheme based on joint light-emitting diode (LED) selection and precoding design (TASP-HD) for multiple-user (MU) multiple-cell (MC) visible light communications (VLCs) systems. In TASP-HD, both the LED selection and the precoding of each cell can be dynamically adjusted to reduce the intra- and inter-cell interferences while satisfying illumination constraints. First, an MU-MC-VLC system model is established, and then a sum-rate maximization problem under the dimming level and illumination uniformity constraints is formulated. In this studied problem, the indices of activated LEDs and precoding matrices are optimized, which result in a complex nonconvex mixed-integer problem. To solve this problem, the original problem is separated into two subproblems. The first subproblem, which maximizes the sum rate of users via optimizing the LED selection with a given precoding matrix, is a mixed-integer problem solved by the penalty method. With the optimized LED selection matrix, the second subproblem which focuses on the maximization of the sum-rate via optimizing the precoding matrix is solved by the Lagrangian dual method. Finally, these two subproblems are iteratively solved to obtain a convergent solution. Simulation results verify that in a typical indoor scenario under a dimming level of 70%, the mean bandwidth efficiency (MBE) of TASP-HD is 4.8 bit/s/Hz and 7.13 bit/s/Hz greater than analog dimming (AD) and digital dimming (DD), respectively.

Over-the-air computation(AirComp) has emerged as a new analog power-domainnon-orthogonal multiple access(NOMA) technique for low-latency model/gradient-updatesaggregation in federated edge learning(FEEL). By integrating communication and computation into a joint design, AirComp can significantly enhance the communication efficiency, but at the cost of aggregation errors caused by channel fading and noise. This paper studies a particular type of FEEL with federated averaging (FedAvg) and AirComp-based model-update aggregation, namelyover-the-airFedAvg (Air-FedAvg). We investigate the transmission power control in Air-FedAvg to combat against the AirComp aggregation errors for enhancing the training accuracy and accelerating the training speed. Towards this end, we first analyze the convergence behavior (in terms of the optimality gap) of Air-FedAvg with aggregation errors at different outer iterations. Then, to enhance the training accuracy, we minimize the optimality gap by jointly optimizing the transmission power control at edge devices and the denoising factors at edge server, subject to a series of power constraints at individual edge devices. Furthermore, to accelerate the training speed, we also minimize the training latency of Air-FedAvg with a given targeted optimality gap, in which learning hyper-parameters including the numbers of outer iterations and local training epochs are optimized jointly with the power control. Finally, numerical results show that the proposed transmission power control policy achieves significantly faster convergence speed for Air-FedAvg, as compared with benchmark policies with fixed power transmission or per-iterationmean squared error(MSE) minimization. It is also shown that the Air-FedAvg achieves an order-of-magnitude shorter training latency than the conventional FedAvg with digitalorthogonal multiple access(OMA-FedAvg).

Over-the-air federated edge learning(Air-FEEL) has emerged as a communication-efficient solution to enable distributed machine learning over edge devices by using their data locally to preserve the privacy. By exploiting the waveform superposition property of wireless channels, Air-FEEL allows the “one-shot” over-the-air aggregation of gradient-updates to enhance the communication efficiency, but at the cost of a compromised learning performance due to the aggregation errors caused by channel fading and noise. This paper investigates the transmission power control to combat against such aggregation errors in Air-FEEL. Different from conventional power control designs (e.g., to minimize the individualmean squared error(MSE) of the over-the-air aggregation at each round), we consider a new power control design aiming at directly maximizing the convergence speed. Towards this end, we first analyze the convergence behavior of Air-FEEL (in terms of the optimality gap) subject to aggregation errors at different communication rounds. It is revealed that if the aggregation estimates are unbiased, then the training algorithm would converge exactly to the optimal point with mild conditions; while if they are biased, then the algorithm would converge with an error floor determined by the accumulated estimate bias over communication rounds. Next, building upon the convergence results, we optimize the power control to directly minimize the derived optimality gaps under the cases without and with unbiased aggregation constraints, subject to a set of average and maximum power constraints at individual edge devices. We transform both problems into convex forms, and obtain their structured optimal solutions, both appearing in a form of regularized channel inversion, by using the Lagrangian duality method. Finally, numerical results show that the proposed power control policies achieve significantly faster convergence for Air-FEEL, as compared with benchmark policies with fixed power transmission or conventional MSE minimization.

This paper considers device-free sensing in an orthogonal frequency division multiplexing (OFDM) cellular network to enable integrated sensing and communication (ISAC). A novel two-phase sensing framework is proposed to localize the passive targets that cannot transmit/receive reference signals to/from the base stations (BSs), where the ranges of the targets are estimated based on their reflected OFDM signals to the BSs in Phase I, and the location of each target is estimated based on its ranges to different BSs in Phase II. Specifically, in Phase I, we design a model-free range estimation approach by leveraging the OFDM channel estimation technique for determining the delay values of all the two-way BS-target-BS paths, which does not rely on any BS-target channel model. In Phase II, we reveal that ghost targets may be falsely detected in some cases as all the targets reflect the same signals to the BSs, which thus do not know how to match each estimated range with the right target. Interestingly, we show that the above data association issue is not a fundamental limitation for device-free sensing: under the ideal case of perfect range estimation in Phase I, the probability for ghost targets to exist is proved to be negligible when the targets are randomly located. Moreover, under the practical case of imperfect range estimation in Phase I, we propose an efficient algorithm for joint data association and target localization in Phase II. Numerical results show that our proposed two-phase framework can achieve very high accuracy in the localization of passive targets, which increases with the system bandwidth.

In this paper, a semantic communication framework is proposed for textual data transmission. In the studied model, a base station (BS) extracts the semantic information from textual data, and transmits it to each user. The semantic information is modeled by a knowledge graph (KG) that consists of a set of semantic triples. After receiving the semantic information, each user recovers the original text using a graph-to-text generation model. To measure the performance of the considered semantic communication framework, a metric of semantic similarity (MSS) that jointly captures the semantic accuracy and completeness of the recovered text is proposed. Due to wireless resource limitations, the BS may not be able to transmit the entire semantic information to each user and satisfy the transmission delay constraint. Hence, the BS must select an appropriate resource block for each user as well as determine and transmit part of the semantic information to the users. As such, we formulate an optimization problem whose goal is to maximize the total MSS by jointly optimizing the resource allocation policy and determining the partial semantic information to be transmitted. To solve this problem, a proximal-policy-optimization-based reinforcement learning (RL) algorithm integrated with an attention network is proposed. The proposed algorithm can evaluate the importance of each triple in the semantic information using an attention network and then, build a relationship between the importance distribution of the triples in the semantic information and the total MSS. Compared to traditional RL algorithms, the proposed algorithm can dynamically adjust its learning rate thus ensuring convergence to a locally optimal solution. Simulation results show that the proposed framework can reduce by 41.3% data that the BS needs to transmit and improve by two-fold the total MSS compared to a standard communication network without using semantic communication techniques.

With the acceleration of a new round of global scientific, technological, and industrial revolution, the next generation of information and communication technology, i.e., 6G, will inject new momentum into industry transformation and upgrading, as well as into economic innovation and development.This will subsequently promote a global industrial integration.Wireless communication will be ubiquitous in all areas of future society, supporting novel applications with various performance requirements, such as immersive-or interactive-experience applications requiring a large bandwidth, autonomous driving and vehicle-to-everything applications requiring ultrahigh reliability and ultra-low latency, and applications for industrial Internet requiring massive machine-type connectivity.Facing the challenges of the post-Moore and post-pandemic era, wireless communication needs breakthroughs in network architecture to improve the intelligence, security, robustness, bandwidth, and heterogeneity.With this background, several important tendencies have emerged in the development of 6G wireless communications

The key challenges in live video multicasting include how to properly form multicast groups, select video versions and allocate wireless resources, in order to guarantee the quality of experience (QoE) while ensuring low latency delivery. To address these challenges, in this paper, a novel multicast framework that leverages the advantages of network-assisted dynamic adaptive streaming over HTTP and cloud radio access networks is proposed, where a multicast assistant server is deployed at the edge of a mobile network. Under this architecture, a joint user grouping, version selection, and bandwidth allocation method is designed to optimize the sum of users’ utilities. In particular, a two-step scheme is proposed to solve this complex problem. The number of multicast groups is first automatically determined and a user clustering method is presented. Then, group-level version selection and spectrum assignment algorithms are performed at different time scales. Simulation results demonstrate that our proposed scheme can improve at least 7% QoE compared to baseline methods.

In the game and film industries, modeling 3D heads plays a very important role in designing characters. Although human head modeling has been researched for a long time, few works have focused on animal-like heads, which are of more diverse shapes and richer geometric details. In this article, we present SAniHead, an interactive system for creating animal-like heads with a mesh representation from dual-view sketches. Our core technical contribution is a view-surface collaborative mesh generative network. Initially, a graph convolutional neural network (GCNN) is trained to learn the deformation of a template mesh to fit the shape of sketches, giving rise to a coarse model. It is then projected into vertex maps where image-to-image translation networks are performed for detail inference. After back-projecting the inferred details onto the meshed surface, a new GCNN is trained for further detail refinement. The modules of view-based detail inference and surface-based detail refinement are conducted in an alternating cascaded fashion, collaboratively improving the model. A refinement sketching interface is also implemented to support direct mesh manipulation. Experimental results show the superiority of our approach and the usability of our interactive system. Our work also contributes a 3D animal head dataset with corresponding line drawings.

Visible light positioning and computer vision-based localization have the potential to be cost-effective technologies for accurate indoor localization. However, the feasibility of existing methods in this domain is limited. In this paper, a novel visible light communication (VLC)-assisted perspective-four-line algorithm (V-P4L) is proposed for practical indoor localization. The basic idea of V-P4L is to jointly use VLC and computer vision techniques to achieve high localization accuracy regardless of LED height differences. In particular, the space-domain information is first exploited to estimate the orientation and coordinate information of a single rectangular LED luminaire in the camera coordinate system based on plane geometry theory and solid geometry theory. Then, by using time-domain information transmitted by VLC and the estimated luminaire information, the proposed V-P4L can estimate the position and pose of the camera using single-view geometry theory and the linear least square (LLS) method. To further mitigate the effect of height differences among LEDs on localization accuracy, a correction algorithm based on the LLS method and a simple optimization method is proposed. Due to the combination of time- and space-domain information, V-P4L can achieve accurate localization using a single luminaire without limitation on the correspondences between the features and their projections in conventional perspective-n-line (PnL) algorithms. Simulation results show that the position error caused by the proposed V-P4L algorithm is always less than 15 cm and the orientation error is always less than 4° using popular indoor luminaires. Experimental results with real hardware show that the average position error is less than 3 cm under both similar and different heights for the LEDs.

This paper studies a mobile edge computing (MEC) network to support emerging applications for Internet-of-things, where multiple access points, each attached with an MEC server, need to collect data from multiple sensors, process them, and then send computation results to the paired actuators for control. Specifically, we consider a three-phase operation protocol for data uploading, edge computing, and result downloading, where the frequency-division multiple access is implemented to accommodate communications of multiple sensors/actuators. Under this setup, we minimize the end-to-end (E2E) latency of the sensing-communication-computation-actuation loop by properly designing the user association and resource allocation policy, subject to the communication and computation resource constraints. The formulated problem, however, is a mixed-integer non-linear program that is difficult to be solved. Despite this fact, we obtain the optimal solution via using the brute-force search for user association, and convex optimization for resource allocation under given user associations. Next, to reduce the computation complexity from the brute-force search, we propose an alternative algorithm, where the user association and resource allocation are optimized iteratively in an alternating manner, via concave-convex procedure and convex optimization, respectively. Finally, numerical results show that the proposed designs significantly reduce the E2E latency, compared with conventional separate designs.

In this paper, the problem of enhancing the quality of virtual reality (VR) services is studied for an indoor terahertz (THz)/visible light communication (VLC) wireless network. In the studied model, small base stations (SBSs) transmit high-quality VR images to VR users over THz bands and light-emitting diodes (LEDs) provide accurate indoor positioning services for them using VLC. Here, VR users move in real time and their movement patterns change over time according to their applications, where both THz and VLC links can be blocked by the bodies of VR users. To control the energy consumption of the studied THz/VLC wireless VR network, VLC access points (VAPs) must be selectively turned on so as to ensure accurate and extensive positioning for VR users. Based on the user positions, each SBS must generate corresponding VR images and establish THz links without body blockage to transmit the VR content. The problem is formulated as an optimization problem whose goal is to maximize the average number of successfully served VR users by selecting the appropriate VAPs to be turned on and controlling the user association with SBSs. To solve this problem, a policy gradient-based reinforcement learning (RL) algorithm that adopts a meta-learning approach is proposed. The proposed meta policy gradient (MPG) algorithm enables the trained policy to quickly adapt to new user movement patterns. In order to solve the problem of maximizing the average number of successfully served users for VR scenarios with large numbers of users, a low-complexity dual method based MPG algorithm (D-MPG) with a low complexity is proposed. Simulation results demonstrate that, compared to a baseline trust region policy optimization algorithm (TRPO), the proposed MPG and D-MPG algorithms yield up to 26.8% and 21.9% improvement in the average number of successfully served users as well as 81.2% and 87.5% gains in the convergence speed, respectively.

Over-the-air computation (AirComp) has emerged as a promising technique in future intelligent wireless networks, which enables swift functional computation among distributed wireless devices (WDs) by exploiting the superposition property of wireless channels. This paper studies a newhierarchicalAirComp network over a large area, in which a set of intermediate relays are exploited to facilitate the massive data aggregation from a large number of WDs. Under this setup, we present a two-phase amplify-and-forward (AF) relaying protocol. In the first phase, the WDs simultaneously send their data to the relays, while in the second phase, the relays amplify the respectively received signals and concurrently forward them to the fusion center (FC) for aggregation. Our objective is to minimize the computational mean squared error (MSE) at the FC, by jointly optimizing the transmit coefficients of the WDs, the AF coefficients of the relays, and the de-noising factor of the FC, subject to their individual transmit power constraints. First, we consider the centralized design with global channel state information (CSI), in which the inter-relay signals can be exploited beneficially for data aggregation. In this case, we develop an alternating-optimization-based algorithm to obtain a high-quality solution to the computational MSE minimization problem. The obtained solution shows that the phase of the transmit coefficient at each WD is opposite to that of the WD-relay-FC channel to ensure the signal phase alignment at the FC, and the transmit power of each WD/relay follows a regularized composite-channel-inversion structure to strike a balance between minimizing the signal-magnitude-misalignment-induced error and the noise-induced error. Next, to reduce the signaling overhead caused by the centralized design, we consider an alternative decentralized design with partial CSI, in which the relays and the FC make their own decisions by only requiring the channel power gain information across different relays. In this case, the relays and FC need to treat the inter-relay signals as harmful interference or noise. Accordingly, we optimize the transmit coefficients of the WDs associated with each relay, and the relay AF coefficients (together with the FC de-noising factor) in an iterative manner, which can be implemented efficiently in a decentralized way. Finally, numerical results show the fast convergence of the proposed centralized and decentralized designs. It is also shown that both designs achieve significant MSE performance gains over benchmark schemes without the joint optimization.

This paper investigates the problem of resource allocation for a wireless communication network with distributed reconfigurable intelligent surfaces (RISs). In this network, multiple RISs are spatially distributed to serve wireless users and the energy efficiency of the network is maximized by dynamically controlling the on-off status of each RIS as well as optimizing the reflection coefficients matrix of the RISs. This problem is posed as a joint optimization problem of transmit beamforming and RIS control, whose goal is to maximize the energy efficiency under minimum rate constraints of the users. To solve this problem, two iterative algorithms are proposed for the single-user case and multi-user case. For the single-user case, the phase optimization problem is solved by using a successive convex approximation method, which admits a closed-form solution at each step. Moreover, the optimal RIS on-off status is obtained by using the dual method. For the multi-user case, a low-complexity greedy searching method is proposed to solve the RIS on-off optimization problem. Simulation results show that the proposed scheme achieves up to 33% and 68% gains in terms of the energy efficiency in both single-user and multi-user cases compared to the conventional RIS scheme and amplify-and-forward relay scheme, respectively.

Federated edge learning (FEEL) is a popular distributed learning framework that allows privacy-preserving collaborative model training via periodic learning-updates communication between edge devices and server. Due to the constrained bandwidth, only a subset of devices can be selected to upload their updates at each training iteration. This has led to an active research area in FEEL studying the optimal device scheduling policy for communication time minimization. However, owing to the difficulty in quantifying the exact communication time, prior work in this area can only tackle the problem partially and indirectly by minimizing either the iteration rounds or per-round latency, while the total communication time is determined by both metrics. To close this research gap, we make the first attempt in this paper to formulate and solve the communication time minimization problem. We first derive a tight bound to approximate the remaining communication time through cross-disciplinary effort that combines the learning theory for convergence rate analysis and communication theory for per-round latency analysis. Building on the novel analytical result, an optimized probabilistic device scheduling policy is derived in closed-form by solving the approximate communication time minimization problem. It is found that the optimized policy gradually turns its priority from suppressing the remaining communication rounds to reducing per-round latency as the training process evolves. Extensive experiments based on real-world dataset and a use case on collaborative 3D objective detection in autonomous driving are provided to verify the superiority of the proposed policy over three benchmark policies based on the indirect solution approaches.

Protein secondary structure prediction (PSSP) is an essential task in computational biology. To achieve the accurate PSSP, the general and vital feature engineering is to use multiple sequence alignment (MSA) for Position-Specific Scoring Matrix (PSSM) extraction. However, when only low-quality PSSM can be obtained due to poor sequence homology, previous PSSP accuracy (merely around 65%) is far from practical usage for subsequent tasks. In this paper, we propose a novel PSSM-Distil framework for PSSP on low-quality PSSM, which not only enhances the PSSM feature at a lower level but also aligns the feature distribution at a higher level. In practice, the PSSM-Distil first exploits the proteins with high-quality PSSM to achieve a teacher network for PSSP in a full-supervised way. Under the guidance of the teacher network, the low-quality PSSM and corresponding student network with low discriminating capacity are effectively resolved by feature enhancement through EnhanceNet and distribution alignment through knowledge distillation with contrastive learning. Further, our PSSM-Distil supports the input from a pre-trained protein sequence language BERT model to provide auxiliary information, which is designed to address the extremely low-quality PSSM cases, i.e., no homologous sequence. Extensive experiments demonstrate the proposed PSSM-Distil outperforms state-of-the-art models on PSSP by 6% on average and nearly 8% in extremely low-quality cases on public benchmarks, BC40 and CB513.

LiDAR point cloud analysis is a core task for 3D computer vision, especially for autonomous driving. However, due to the severe sparsity and noise interference in the single sweep LiDAR point cloud, the accurate semantic segmentation is non-trivial to achieve. In this paper, we propose a novel sparse LiDAR point cloud semantic segmentation framework assisted by learned contextual shape priors. In practice, an initial semantic segmentation (SS) of a single sweep point cloud can be achieved by any appealing network and then flows into the semantic scene completion (SSC) module as the input. By merging multiple frames in the LiDAR sequence as supervision, the optimized SSC module has learned the contextual shape priors from sequential LiDAR data, completing the sparse single sweep point cloud to the dense one. Thus, it inherently improves SS optimization through fully end-to-end training. Besides, a Point-Voxel Interaction (PVI) module is proposed to further enhance the knowledge fusion between SS and SSC tasks, i.e., promoting the interaction of incomplete local geometry of point cloud and complete voxel-wise global structure. Furthermore, the auxiliary SSC and PVI modules can be discarded during inference without extra burden for SS. Extensive experiments confirm that our JS3C-Net achieves superior performance on both SemanticKITTI and SemanticPOSS benchmarks, i.e., 4% and 3% improvement correspondingly.

In federated learning, heterogeneity among the clients’ local datasets results in large variations in the number of local updates performed by each client in a communication round. Simply aggregating such local models into a global model will confine the capacity of the system, that is, the single global model will be restricted from delivering good performance on each client’s task. This paper provides a general framework to analyze the convergence of personalized federated learning algorithms. It subsumes previously proposed methods and provides a principled understanding of the computational guarantees. Using insights from this analysis, we propose PFedAtt, a personalized federated learning method that incorporates attention-based grouping to facilitate similar clients’ collaborations. Theoretically, we provide the convergence guarantee for the algorithm, and empirical experiments corroborate the competitive performance of PFedAtt on heterogeneous clients.

Grounding referring expressions in RGBD image has been an emerging field. We present a novel task of 3D visual grounding in single-view RGBD image where the referred objects are often only partially scanned due to occlusion. In contrast to previous works that directly generate object proposals for grounding in the 3D scenes, we propose a bottom-up approach to gradually aggregate content-aware information, effectively addressing the challenge posed by the partial geometry. Our approach first fuses the language and the visual features at the bottom level to generate a heatmap that coarsely localizes the relevant regions in the RGBD image. Then our approach conducts an adaptive feature learning based on the heatmap and performs the object-level matching with another visio-linguistic fusion to finally ground the referred object. We evaluate the proposed method by comparing to the state-of-the-art methods on both the RGBD images extracted from the ScanRefer dataset and our newly collected SUNRefer dataset. Experiments show that our method outperforms the previous methods by a large margin (by 11.2% and 15.6% [email protected]) on both datasets.

Weakly supervised object localization (WSOL) aims to localize objects by only utilizing image-level labels. Class activation maps (CAMs) are the commonly used features to achieve WSOL. However, previous CAM-based methods did not take full advantage of the shallow features, despite their importance for WSOL. Because shallow features are easily buried in background noise through conventional fusion. In this paper, we propose a simple but effective Shallow feature-aware Pseudo supervised Object Localization (SPOL) model for accurate WSOL, which makes the utmost of low-level features embedded in shallow layers. In practice, our SPOL model first generates the CAMs through a novel element-wise multiplication of shallow and deep feature maps, which filters the background noise and generates sharper boundaries robustly. Besides, we further propose a general class-agnostic segmentation model to achieve the accurate object mask, by only using the initial CAMs as the pseudo label without any extra annotation. Eventually, a bounding box extractor is applied to the object mask to locate the target. Experiments verify that our SPOL outperforms the state-of-the-art on both CUB- 200 and ImageNet-1K benchmarks, achieving 93.44% and 67.15% (i.e., 3.93% and 2.13% improvement) Top-5 localization accuracy, respectively.

In this paper, a semantic communication framework is proposed for wireless networks. In the proposed framework, a base station (BS) extracts the semantic information from textual data, and, transmits it to each user. This semantic information is modeled by a knowledge graph (KG) and hence, the semantic information consists of a set of semantic triples. After receiving the semantic information, each user recovers the original text using a graph-to-text generation model. To measure the performance of the studied semantic communication system, a metric of semantic similarity (MSS) that jointly captures the semantic accuracy and completeness of the recovered text is proposed. Due to wireless resource limitations, the BS can only transmit partial semantic information to each user so as to satisfy the transmission delay constraint. Hence, the BS must select an appropriate resource block for each user and determine partial semantic information to be transmitted. This problem is formulated as an optimization problem whose goal is to maximize the total MSS by optimizing the resource allocation policy and determining the partial semantic information to be transmitted. To solve this problem, a policy gradient-based reinforcement learning (RL) algorithm integrated with the attention network is proposed. The proposed algorithm can evaluate the importance of each triple in the semantic information using an attention network and then, build a relationship between the importance distribution of the triples in the semantic information and the total MSS. Simulation results demonstrate that the proposed semantic communication framework can reduce the size of data that the BS needs to transmit by up to 46% and yield a two-fold improvement in the total MSS compared to a standard communication network that does not consider semantic communications.

Federated learning (FL) is widely used in privacy sensitive applications for isolated data islands, with the aim of achieving distributed model training, privacy enhancement and model sharing. Electromyographic (EMG) signals are a type of data collected from wearable sensors of subjects which are distributed on multiple devices, highly personalized and play an important role in several applications including prosthetic hand control, sign languages, grasp recognition, etc. This paper utilizes the FL method to detect single and combined finger movements based on EMG signals. The existing research on FL for wearable healthcare faces challenges of variable probability distributions of data, the need for prerequisite knowledge of server model and computational burdens in parameter transmission. To address these problems, this paper proposes a communication efficient FL framework in which each device only needs to transmit the weight matrices of local models to the server for model aggregation. To further reduce the FL transmission delay, a joint learning and resource allocation problem is formulated via optimizing transmit power of each device, time allocation, and user selection. To solve the delay minimization problem, the objective function is first converted to a tractable expression and then the difference of two convex functions programming is adopted. Simulation results using real EMG signals show that the proposed FL framework with personalized training process successfully detects single and combined finger movements for distributed users. Two public EMG datasets with 10 and 15 different finger movements are employed. Over 98% overall test accuracy is achieved in both datasets which surpasses the conventional learning framework by 1.6% and 0.5% on average. Different scenarios with regard to access points and users are investigated and the convexity of the proposed model is discussed.

This paper investigates the deployment of federated learning (FL) over an over-the-air computation (AirComp) and intelligent reflecting surface (IRS) based wireless network. In the considered system, devices transmit locally trained machine learning (ML) models to the base station (BS) which aggregates the received ML models and generates a shared global ML model. The devices can directly transmit ML models to the BS or using IRS. Meanwhile, AirComp is used to aggregate ML models that are transmitted from the devices to the BS. To minimize the energy consumption of devices, an energy minimization problem is formulated, which jointly optimizes the device selection, phase shift matrix, decoding vector, and power control. To seek the solution, the original optimization problem is divided into four sub-problems. Then the fractional program, greedy algorithm, matrix derivation, and weighted minimum mean square error methods are used to compute the phase shift matrix, device selection vector, decoding vector, and transmit power, respectively. Simulation results show that the proposed algorithm can reduce 11.2% energy consumption of devices compared to an FL algorithm that is implemented at a network without any IRSs.

In this paper, layered min-sum (MS) iterative decoding is formulated as a customized neural network following the sequential scheduling of check node (CN) updates. By virtue of the lifting structure of protograph low-density parity-check (LDPC) codes, identical network parameters are shared among all derived edges originating from the same edge in the protograph, which makes the number of learn- able parameters manageable. The proposed neural layered MS decoder can support arbitrary codelengths consequently. Moreover, an iteration-wise greedy training method is proposed to tune the parameters such that it avoids the vanishing gradient problem and accelerates the decoding convergence.

In this paper, the implementation of a distributed primal-dual algorithm over realistic wireless networks is investigated. In the considered model, the users and one base station (BS) cooperatively perform a distributed primal-dual algorithm for controlling and optimizing wireless networks. In particular, each user must locally update the primal and dual variables and send the updated primal variables to the BS. The BS aggregates the received primal variables and broadcasts the aggregated variables to all users. Since all of the primal and dual variables as well as aggregated variables are transmitted over wireless links, the imperfect wireless links will affect the solution achieved by the distributed primal-dual algorithm. Therefore, it is necessary to study how wireless factors such as transmission errors affect the implementation of the distributed primal-dual algorithm and how to optimize wireless network performance to improve the solution achieved by the distributed primal-dual algorithm. To address these challenges, the convergence rate of the primal-dual algorithm is first derived in a closed form while considering the impact of wireless factors such as data transmission errors. Based on the derived convergence rate, the optimal transmit power and resource block allocation schemes are designed to minimize the gap between the target solution and the solution achieved by the distributed primal-dual algorithm. Simulation results show that the proposed distributed primal-dual algorithm can reduce the gap between the target and obtained solution by up to 52% compared to the distributed primal-dual algorithm without considering imperfect wireless transmission.

Over-the-air federated edge learning (Air-FEEL) is a communication-efficient solution for privacy-preserving distributed learning over wireless networks. Air-FEEL allows "one-shot" over-the-air aggregation of gradient/model-updates by exploiting the waveform superposition property of wireless channels, and thus promises an extremely low aggregation latency that is independent of the network size. However, such communication efficiency may come at a cost of learning performance degradation due to the aggregation error caused by the non-uniform channel fading over devices and noise perturbation. Prior work adopted channel inversion power control (or its variants) to reduce the aggregation error by aligning the channel gains, which, however, could be highly suboptimal in deep fading scenarios due to the noise amplification. To overcome this issue, we investigate the power control optimization for enhancing the learning performance of Air-FEEL. Towards this end, we first analyze the convergence behavior of the Air-FEEL by deriving the optimality gap of the loss-function under any given power control policy. Then we optimize the power control to minimize the optimality gap for accelerating convergence, subject to a set of average and maximum power constraints at edge devices. The problem is generally non-convex and challenging to solve due to the coupling of power control variables over different devices and iterations. To tackle this challenge, we develop an efficient algorithm by jointly exploiting the successive convex approximation (SCA) and trust region methods. Numerical results show that the optimized power control policy achieves significantly faster convergence than the benchmark policies such as channel inversion and uniform power transmission.

We consider the problem of convergence time minimization for federated learning (FL) implemented in wireless systems. In such setups, each wireless edge device transmits its local FL model parameters to a base station (BS). The BS then uses the received FL parameters to generate a common FL model and broadcasts it to all edge devices. Since the FL parameters must be transmitted over wireless links, the convergence time depends not only on the number of training steps, but also on the FL parameter transmission delay at each training step, which can be substantial when conveying a large number of parameters. In addition, due to limited wireless resources such as spectrum, only a subset of edge devices can participate in each FL training step, which can further increase convergence time. Our goal therefore is to optimize wireless resource management and user selection for FL, as well as limit the volume of transmitted FL parameters. In this paper, three schemes for facilitating communication efficient FL are introduced: First, a probabilistic device selection scheme is designed such that the devices that can significantly improve the convergence speed and training loss have high probabilities for FL parameter transmission. Then, given the subset of participating devices, an efficient wireless resource allocation scheme is developed. Finally, a quantization method is proposed to reduce the data size. Simulation results demonstrate that the proposed FL method can improve handwritten digit identification accuracy and convergence delay by up to 3% and 90% compared to the conventional FL.

Federated learning-based navigation has received much attention in vehicular IoT. The intention is to employ a big number of end-users for data collection along different trajectories and perform local training of a global learning model to substitute the global positioning system (GPS) in urban areas. The prerequisites for its commercialization, however, lie in the location-dependent input data trustworthiness and participants’ privacy preservation. In this paper, we propose a privacy-preserving proof-of-location mechanism using blockchain to meet these conditions. Specifically, the proposed scheme utilizes a Threshold Identity-Based Encryption (TIBE) system for the generation of secret shares, such that each anonymous location proof can only be verified with at least a threshold number of participants. In addition, the proposed scheme exploits a cuckoo filter for the secure and efficient maintenance and dissemination of location proofs. Systematic security analysis is conducted to demonstrate the fulfillment of harsh security requirements. Performance evaluations are carried out to validate the computation efficiency in comparison with an oblivious transfer (OT) protocol, which has been widely adopted for secure data acquisition.

Different from public 4G/5G networks that are dominated by downlink (DL) traffic, emerging 5G non-public networks (NPNs) need to support significant uplink (UL) traffic to enable emerging applications such as industrial Internet of things (IIoT). The UL-DL spectrum sharing is becoming a viable solution to enhance the UL throughput of NPNs, which allows NPNs to perform the UL transmission over the time-frequency resources configured for DL transmission in coexisting public networks. To deal with the severe interference from the DL public base station (BS) transmitter to the coexisting UL non-public BS receiver, we propose an adaptive asymmetric successive interference cancellation (SIC) approach, in which the non-public BS is enabled to have the capability of decoding the DL signals transmitted from the public BS and cancelling them for interference mitigation. In particular, this paper studies a basic UL-DL spectrum sharing scenario when a UL non-public BS and a DL public BS coexist in the same area, each communicating with multiple users via orthogonal frequency-division multiple access (OFDMA). Under this setup, we aim to maximize the common UL throughput of all non-public users, under the condition that the DL throughput of each public user is above a certain threshold. The decision variables include the subcarrier allocation and user scheduling for both non-public and public BSs, the receiver mode of the non-public BS over subcarriers, as well as the rate and power control. Numerical results show that the proposed design significantly improves the common UL throughput as compared to benchmark schemes without such consideration.

In this paper, the problem of enhancing the quality of virtual reality (VR) services is studied for an indoor terahertz (THz)/visible light communication (VLC) wireless network. In the studied model, small base stations (SBSs) transmit high-quality VR images to users over THz bands and light-emitting diodes (LEDs) provide accurate indoor positioning services for VR users using VLC. Here, VR users move in real time and their movement patterns change over time according to their application. Both THz and VLC links can be blocked by the bodies of VR users. To control the energy consumption of the studied THz/VLC wireless VR network, VLC access points (VAPs) must be selectively turned on so as to ensure accurate and extensive positioning for VR users. Based on the user positions, each SBS must generate corresponding VR images and build THz links without body blockage to transmit the VR content. The problem is formulated as an optimization problem whose goal is to maximize the sum successful transmission probability of all VR users by selecting the appropriate VAPs to be turned on and controlling the user association with SBSs. To solve this problem, a policy gradient-based reinforcement learning (RL) algorithm using meta-learning framework is proposed. The proposed algorithm can effectively solve the formulated problem and enable the trained policy to quickly adapt to new user movement patterns. Simulation results demonstrate that, compared to a baseline trust region policy optimization algorithm (TRPO), the proposed meta-learning solution yields a 78% improvement in the convergence speed and about 16.4% improvement in the sum successful transmission probabilities of all VR users.

The key point of language-guided person search is to construct the cross-modal association between visual and textual input. Existing methods focus on designing multimodal attention mechanisms and novel cross-modal loss functions to learn such association implicitly. We propose a representation learning method for language-guided person search based on color reasoning (LapsCore). It can explicitly build a fine-grained cross-modal association bidirectionally. Specifically, a pair of dual sub-tasks, image colorization and text completion, is designed. In the former task, rich text information is learned to colorize gray images, and the latter one requests the model to understand the image and complete color word vacancies in the captions. The two sub-tasks enable models to learn correct alignments between text phrases and image regions, so that rich multimodal representations can be learned. Extensive experiments on multiple datasets demonstrate the effectiveness and superiority of the proposed method.

Compared with the visual grounding on 2D images, the natural-language-guided 3D object localization on point clouds is more challenging. In this paper, we propose a new model, named InstanceRefer1, to achieve a superior 3D visual grounding through the grounding-by-matching strategy. In practice, our model first predicts the target category from the language descriptions using a simple language classification model. Then, based on the category, our model sifts out a small number of instance candidates (usually less than 20) from the panoptic segmentation on point clouds. Thus, the non-trivial 3D visual grounding task has been effectively re-formulated as a simplified instance-matching problem, considering that instance-level candidates are more rational than the redundant 3D object proposals. Subsequently, for each candidate, we perform the multi-level contextual inference, i.e., referring from instance attribute perception, instance-to-instance relation perception, and instance-to-background global localization perception, respectively. Eventually, the most relevant candidate is selected and localized by ranking confidence scores, which are obtained by the cooperative holistic visual-language feature matching. Experiments confirm that our method outperforms previous state-of-the-arts on ScanRefer online benchmark and Nr3D/Sr3D datasets.

Current 3D single object tracking approaches track the target based on a feature comparison between the target template and the search area. However, due to the common occlusion in LiDAR scans, it is non-trivial to conduct accurate feature comparisons on severe sparse and incomplete shapes. In this work, we exploit the ground truth bounding box given in the first frame as a strong cue to enhance the feature description of the target object, enabling a more accurate feature comparison in a simple yet effective way. In particular, we first propose the BoxCloud, an informative and robust representation, to depict an object using the point-to-box relation. We further design an efficient box-aware feature fusion module, which leverages the aforementioned BoxCloud for reliable feature matching and embedding. Integrating the proposed general components into an existing model P2B [27], we construct a superior box-aware tracker (BAT)1. Experiments confirm that our proposed BAT outperforms the previous state-of-the-art by a large margin on both KITTI and NuScenes benchmarks, achieving a 12.8% improvement in terms of precision while running ∼20% faster.

Protein secondary structure prediction (PSSP) is essential for protein function analysis. However, for low homologous proteins, the PSSP suffers from insufficient input features. In this paper, we explicitly import external self-supervised knowledge for low homologous PSSP under the guidance of residue-wise (amino acid wise) profile fusion. In practice, we firstly demonstrate the superiority of profile over Position-Specific Scoring Matrix (PSSM) for low homologous PSSP. Based on this observation, we introduce the novel self-supervised BERT features as the pseudo profile, which implicitly involves the residue distribution in all native discovered sequences as the complementary features. Furthermore, a novel residue-wise attention is specially designed to adaptively fuse different features (i.e., original low-quality profile, BERT based pseudo profile), which not only takes full advantage of each feature but also avoids noise disturbance. Besides, the feature consistency loss is proposed to accelerate the model learning from multiple semantic levels. Extensive experiments confirm that our method outperforms state-of-the-arts (i.e., 4.7% for extremely low homologous cases on BC40 dataset).

Point Cloud Sampling and Recovery (PCSR) is critical for massive real-time point cloud collection and processing since raw data usually requires large storage and computation. This paper addresses a fundamental problem in PCSR: How to downsample the dense point cloud with arbitrary scales while preserving the local topology of discarded points in a case-agnostic manner (i.e., without additional storage for point relationships)? We propose a novel Locally Invertible Embedding (PointLIE) framework to unify the point cloud sampling and upsampling into one single framework through bi-directional learning. Specifically, PointLIE decouples the local geometric relationships between discarded points from the sampled points by progressively encoding the neighboring offsets to a latent variable. Once the latent variable is forced to obey a pre-defined distribution in the forward sampling path, the recovery can be achieved effectively through inverse operations. Taking the recover-pleasing sampled points and a latent embedding randomly drawn from the specified distribution as inputs, PointLIE can theoretically guarantee the fidelity of reconstruction and outperform state-of-the-arts quantitatively and qualitatively.

Recent advances in unsupervised domain adaptation have achieved remarkable performance on semantic segmentation tasks. Despite such progress, existing works mainly focus on bridging the inter-domain gaps between the source and target domain, while only few of them noticed the intra-domain gaps within the target data. In this work, we propose a pixel-level intra-domain adaptation approach to reduce the intra-domain gaps within the target data. Compared with image-level methods, ours treats each pixel as an instance, which adapts the segmentation model at a more fine-grained level. Specifically, we first conduct the inter-domain adaptation between the source and target domain; Then, we separate the pixels in target images into the easy and hard subdomains; Finally, we propose a pixel-level adversarial training strategy to adapt a segmentation network from the easy to the hard subdomain. Moreover, we show that the segmentation accuracy can be further improved by incorporating a continuous indexing technique in the adversarial training. Experimental results show the effectiveness of our method against existing state-of-the-art approaches.

Limited radio frequency (RF) resources restrict the number of users that can participate in federated learning (FL) thus affecting FL convergence speed and performance. In this paper, we first introduce visible light communication (VLC) as a supplement to RF in FL and build a hybrid VLC/RF communication system, in which each indoor user can use both VLC and RF to transmit its FL model parameters. Then, the problem of user selection and bandwidth allocation is studied for FL implemented over a hybrid VLC/RF system aiming to optimize the FL performance. The problem is first separated into two subproblems. The first subproblem is a user selection problem with a given bandwidth allocation, which is solved by a traversal algorithm. The second subproblem is a bandwidth allocation problem with a given user selection, which is solved by a numerical method. The final user selection and bandwidth allocation are obtained by iteratively solving these two subproblems. Simulation results show that the proposed FL algorithm that efficiently uses VLC and RF for FL model transmission can improve the prediction accuracy by up to 10% compared with a conventional FL system using only RF.

Abstract Modeling 3D objects on existing software usually requires a heavy amount of interactions, especially for users who lack basic knowledge of 3D geometry. Sketch‐based modeling is a solution to ease the modelling procedure and thus has been researched for decades. However, modelling a man‐made shape with complex structures remains challenging. Existing methods adopt advanced deep learning techniques to map holistic sketches to 3D shapes. They are still bottlenecked to deal with complicated topologies. In this paper, we decouple the task of sketch2shape into a part generation module and a part assembling module, where deep learning methods are leveraged for the implementation of both modules. By changing the focus from holistic shapes to individual parts, it eases the learning process of the shape generator and guarantees high‐quality outputs. With the learned automated part assembler, users only need a little manual tuning to obtain a desired layout. Extensive experiments and user studies demonstrate the usefulness of our proposed system.

In this article, a joint task, spectrum, and transmit power allocation problem is investigated for a wireless network in which the base stations (BSs) are equipped with mobile-edge computing (MEC) servers to jointly provide computational and communication services to users. Each user can request one computational task from three types of computational tasks. Since the data size of each computational task is different, as the requested computational task varies, the BSs must adjust their resource (subcarrier and transmit power) and task allocation schemes to effectively serve the users. This problem is formulated as an optimization problem whose goal is to minimize the maximal computational and transmission delay among all users. A multistack reinforcement learning (RL) algorithm is developed to solve this problem. Using the proposed algorithm, each BS can record the historical resource allocation schemes and users’ information in its multiple stacks to avoid learning the same resource allocation scheme and users’ states, thus improving the convergence speed and learning efficiency. The simulation results illustrate that the proposed algorithm can reduce the number of iterations needed for convergence and the maximal delay among all users by up to 18% and 11.1% compared to the standard$Q$-learning algorithm.

In this article, the problem of minimizing energy and time consumption for task computation and transmission in mobile-edge computing-enabled balloon networks is investigated. In the considered network, high-altitude balloons (HABs), acting as flying wireless base stations, can use their powerful computational capabilities to process the computational tasks offloaded from their associated users. Since the data size of each user’s computational task varies over time, the HABs must dynamically adjust their resource allocation schemes to meet the users’ needs. This problem is posed as an optimization problem, whose goal is to minimize the energy and time consumption for task computation and transmission by adjusting the user association, service sequence, and task allocation schemes. To solve this problem, a support vector machine (SVM)-based federated learning (FL) algorithm is proposed to determine the user association proactively. The proposed SVM-based FL method enables HABs to cooperatively build an SVM model that can determine all user associations without any transmissions of either user historical associations or computational tasks to other HABs. Given the predictions of the optimal user association, the service sequence and task allocation of each user can be optimized so as to minimize the weighted sum of the energy and time consumption. Simulations with real-city cellular traffic data show that the proposed algorithm can reduce the weighted sum of the energy and time consumption of all users by up to 15.4% compared to a conventional centralized method.

The recent success of single-agent reinforcement learning (RL) in Internet of Things (IoT) systems motivates the study of multiagent RL (MARL), which is more challenging but more useful in large-scale IoT. In this article, we consider a voting-based MARL problem, in which the agents vote to make group decisions and the goal is to maximize the globally averaged returns. To this end, we formulate the MARL problem based on the linear programming form of the policy optimization problem and propose a primal-dual algorithm to obtain the optimal solution. We also propose a voting mechanism through which the distributed learning achieves the same sublinear convergence rate as centralized learning. In other words, the distributed decision making does not slow down the process of achieving global consensus on optimality. Finally, we verify the convergence of our proposed algorithm with numerical simulations and conduct case studies in practical multiagent IoT systems.

The recent development of deep learning methods provides a new approach to optimize the belief propagation (BP) decoding of linear codes.However, the limitation of existing works is that the scale of neural networks increases rapidly with the codelength, thus they can only support short to moderate codelengths.From the point view of practicality, we propose a high-performance neural min-sum (MS) decoding method that makes full use of the lifting structure of protograph low-density parity-check (LDPC) codes.By this means, the size of the parameter array of each layer in the neural decoder only equals the number of edge-types for arbitrary codelengths.In particular, for protograph LDPC codes, the proposed neural MS decoder is constructed in a special way such that identical parameters are shared by a bundle of edges derived from the same edge-type.To reduce the complexity and overcome the vanishing gradient problem in training the proposed neural MS decoder, an iteration-byiteration (i.e., layer-by-layer in neural networks) greedy training method is proposed.With this, the proposed neural MS decoder tends to be optimized with faster convergence, which is aligned with the early termination mechanism widely used in practice.To further enhance the generalization ability of the proposed neural MS decoder, a codelength/rate compatible training method is proposed, which randomly selects samples from a set of codes lifted from the same base code.As a theoretical performance evaluation tool, a trajectory-based extrinsic information transfer (T-EXIT) chart is developed for various decoders.Both T-EXIT and simulation results show that the optimized MS decoding can provide faster convergence and up to 1dB gain compared with the plain MS decoding and its variants with only slightly increased complexity.In addition, it can even outperform the sum-product algorithm for some short codes.

In this paper, the problem of the trajectory design for a group of energy-constrained drones operating in dynamic wireless network environments is studied. In the considered model, a team of drone base stations (DBSs) is dispatched to cooperatively serve clusters of ground users that have dynamic and unpredictable uplink access demands. In this scenario, the DBSs must cooperatively navigate in the considered area to maximize coverage of the dynamic requests of the ground users. This trajectory design problem is posed as an optimization framework whose goal is to find optimal trajectories that maximize the fraction of users served by all DBSs. To find an optimal solution for this non-convex optimization problem under unpredictable environments, a value decomposition based reinforcement learning (VD-RL) solution coupled with a meta-training mechanism is proposed. This algorithm allows the DBSs to dynamically learn their trajectories while generalizing their learning to unseen environments. Analytical results show that, the proposed VD-RL algorithm is guaranteed to converge to a locally optimal solution of the non-convex optimization problem. Simulation results show that, even without meta-training, the proposed VD-RL algorithm can achieve a 53.2% improvement of the service coverage and a 30.6% improvement in terms of the convergence speed, compared to baseline multi-agent algorithms. Meanwhile, the use of the meta-training mechanism improves the convergence speed of the VD-RL algorithm by up to 53.8% when the DBSs must deal with a previously unseen task.

This paper considers the blind recognition of the type and the encoding parameters of channel codes from the Gaussian noisy signals. Specifically, based on the recurrent neural network (RNN), the attention mechanism, and the residual neural network (ResNet), three universal recognizers are proposed to identify the type, rate, and length of the target channel codes, with a training set generated by a small portion of all the possible code parameters. The proposed architectures need near zero a priori knowledge about the target channel code, and only require the length of the received signal to be dozen times of the codeword length. Numerical experiments show that the proposed deep learning methods own strong generalization to identify channel codes from the testing samples not generated by the encoding parameters utilized for the training set.

Federated learning-based automotive navigation has recently received considerable attention, as it can potentially address the issue of weak global positioning system (GPS) signals under severe blockages, such as in downtowns and tunnels. Specifically, the data-driven navigation framework combines the position estimation offered by the high-sampling inertial measurement units and the position calibration provided by the low-sampling GPS signals. Despite its promise, the privacy preservation and flexibility of the participating users in the federated learning process are still problematic. To address these challenges, in this article, we propose an efficient, flexible, and privacy-preserving model aggregation scheme under a federated learning-based navigation framework named FedLoc. Specifically, our proposed scheme efficiently protects the locally trained model updates, flexibly supports the fluctuation of participants, and is robust against unregistered malicious users by exploiting a homomorphic threshold cryptosystem, together with the bounded Laplace mechanism and the skip list. We perform a detailed security analysis to demonstrate the security properties in terms of privacy preservation and dishonest user detection. In addition, we evaluate and compare the computational efficiency with two traditional schemes, and the simulation results show that our scheme greatly improves the computational efficiency during participant fluctuation. To validate the effectiveness of our scheme, we also show that only part of the model update is excluded from aggregation in the case of a dishonest user.

Along with increasingly popular virtual reality applications, the three-dimensional (3D) point cloud has become a fundamental data structure to characterize 3D objects and surroundings. To process 3D point clouds efficiently, a suitable model for the underlying structure and outlier noises is always critical. In this work, we propose a hypergraph-based new point cloud model that is amenable to efficient analysis and processing. We introduce tensor-based methods to estimate hypergraph spectrum components and frequency coefficients of point clouds in both ideal and noisy settings. We establish an analytical connection between hypergraph frequencies and structural features. We further evaluate the efficacy of hypergraph spectrum estimation in two common applications of sampling and denoising of point clouds for which we provide specific hypergraph filter design and spectral properties. Experimental results demonstrate the strength of hypergraph signal processing as a tool in characterizing the underlying properties of 3D point clouds.

With the advances of Internet of Things (IoT) solutions in intelligent transportation systems, collected vehicle data can produce insights on emerging vehicular phenomenon, and further contribute to the further improvement of innovative and efficient vehicular systems. Particularly, by leveraging data collected from vehicle sensors and maintenance models constructed from operation and repair history, predictive maintenance aims to detect the anomalies of vehicles and provide early warnings before the occurrence of failure. However, privacy preservation still remains as one of the top concerns for vehicle owners in predictive maintenance, as the sensory data could potentially violate their location and identity privacy. To address this challenge, in this article, we propose a privacy-preserving and verifiable continuous data collection scheme with the intent of predictive maintenance in vehicular fog, which gathers and organizes the sensor data of each individual vehicle on a sliding window basis. Specifically, our proposed scheme exploits the homomorphic Paillier cryptosystem and truncated α-geometric technique to protect the content of each individual piece of sensory data. Meanwhile, our proposed scheme also aggregates and authenticates the collected sensory data reports on a time-series sliding window basis, which achieves the continuous observation of the recently collected vehicular sensory data. Detailed security analysis is carried out to demonstrate the security properties of our proposed scheme, including confidentiality, authentication and privacy preservation. In performance evaluations, we also compare our proposed scheme with a traditional scheme, and our scheme shows great improvement in terms of communication and computation overheads. Furthermore, to show the feasibility of our proposed scheme, we also compare and discuss the expected squared error introduced by the differential privacy mechanism.

Medical imaging technologies, including computed tomography (CT) or chest X-Ray (CXR), are largely employed to facilitate the diagnosis of the COVID-19. Since manual report writing is usually too time-consuming, a more intelligent auxiliary medical system that could generate medical reports automatically and immediately is urgently needed. In this article, we propose to use the medical visual language BERT (Medical-VLBERT) model to identify the abnormality on the COVID-19 scans and generate the medical report automatically based on the detected lesion regions. To produce more accurate medical reports and minimize the visual-and-linguistic differences, this model adopts an alternate learning strategy with two procedures that are knowledge pretraining and transferring. To be more precise, the knowledge pretraining procedure is to memorize the knowledge from medical texts, while the transferring procedure is to utilize the acquired knowledge for professional medical sentences generations through observations of medical images. In practice, for automatic medical report generation on the COVID-19 cases, we constructed a dataset of 368 medical findings in Chinese and 1104 chest CT scans from The First Affiliated Hospital of Jinan University, Guangzhou, China, and The Fifth Affiliated Hospital of Sun Yat-sen University, Zhuhai, China. Besides, to alleviate the insufficiency of the COVID-19 training samples, our model was first trained on the large-scale Chinese CX-CHR dataset and then transferred to the COVID-19 CT dataset for further fine-tuning. The experimental results showed that Medical-VLBERT achieved state-of-the-art performances on terminology prediction and report generation with the Chinese COVID-19 CT dataset and the CX-CHR dataset. The Chinese COVID-19 CT dataset is available at https://covid19ct.github.io/.

In this paper, the convergence time of federated learning (FL), when deployed over a realistic wireless network, is studied. In particular, a wireless network is considered in which wireless users transmit their local FL models (trained using their locally collected data) to a base station (BS). The BS, acting as a central controller, generates a global FL model using the received local FL models and broadcasts it back to all users. Due to the limited number of resource blocks (RBs) in a wireless network, only a subset of users can be selected to transmit their local FL model parameters to the BS at each learning step. Moreover, since each user has unique training data samples, the BS prefers to include all local user FL models to generate a converged global FL model. Hence, the FL training loss and convergence time will be significantly affected by the user selection scheme. Therefore, it is necessary to design an appropriate user selection scheme that can select the users who can contribute toward improving the FL convergence speed more frequently. This joint learning, wireless resource allocation, and user selection problem is formulated as an optimization problem whose goal is to minimize the FL convergence time and the FL training loss. To solve this problem, a probabilistic user selection scheme is proposed such that the BS is connected to the users whose local FL models have significant effects on the global FL model with high probabilities. Given the user selection policy, the uplink RB allocation can be determined. To further reduce the FL convergence time, artificial neural networks (ANNs) are used to estimate the local FL models of the users that are not allocated any RBs for local FL model transmission at each given learning step, which enables the BS to improve the global model, the FL convergence speed, and the training loss. Simulation results show that the proposed approach can reduce the FL convergence time by up to 56% and improve the accuracy of identifying handwritten digits by up to 3%, compared to a standard FL algorithm.

In this article, the problem of training federated learning (FL) algorithms over a realistic wireless network is studied. In the considered model, wireless users execute an FL algorithm while training their local FL models using their own data and transmitting the trained local FL models to a base station (BS) that generates a global FL model and sends the model back to the users. Since all training parameters are transmitted over wireless links, the quality of training is affected by wireless factors such as packet errors and the availability of wireless resources. Meanwhile, due to the limited wireless bandwidth, the BS needs to select an appropriate subset of users to execute the FL algorithm so as to build a global FL model accurately. This joint learning, wireless resource allocation, and user selection problem is formulated as an optimization problem whose goal is to minimize an FL loss function that captures the performance of the FL algorithm. To seek the solution, a closed-form expression for the expected convergence rate of the FL algorithm is first derived to quantify the impact of wireless factors on FL. Then, based on the expected convergence rate of the FL algorithm, the optimal transmit power for each user is derived, under a given user selection and uplink resource block (RB) allocation scheme. Finally, the user selection and uplink RB allocation is optimized so as to minimize the FL loss function. Simulation results show that the proposed joint federated learning and communication framework can improve the identification accuracy by up to 1.4%, 3.5% and 4.1%, respectively, compared to: 1) An optimal user selection algorithm with random resource allocation, 2) a standard FL algorithm with random user selection and resource allocation, and 3) a wireless optimization algorithm that minimizes the sum packet error rates of all users while being agnostic to the FL parameters.

With the advent of deep neural networks, learning-based single-view reconstruction has gained popularity. However, in 3D, there is no absolutely dominant representation that is both computationally efficient and accurate yet allows for reconstructing high-resolution geometry of arbitrary topology. After all, the accurate implicit methods are time-consuming due to dense sampling and inference, while volumetric approaches are fast but limited to heavy memory usage and low accuracy. In this paper, we propose VIPNet, an end-to-end hybrid representation learning for fast and accurate single-view reconstruction under sparse implicit point guidance. Given an image, it first generates a volumetric result. Meanwhile, a corresponding implicit shape representation is learned. To balance the efficiency and accuracy, we adopt PointGenNet to learn some representative points for guiding the voxel refinement with the corresponding sparse implicit inference. A strategy of patch-based synthesis with global-local features under implicit guidance is also applied for reducing memory consumption required to generate high-resolution output. Extensive experiments demonstrate the effectiveness of our method both qualitatively and quantitatively, which indicates that our proposed hybrid learning outperforms separate representation learning. Specifically, our network not only runs 60 times faster than implicit methods but also contributes to accuracy gains. We hope it will inspire a re-thinking of hybrid representation learning.

We introduce FPConv, a novel surface-style convolution operator designed for 3D point cloud analysis. Unlike previous methods, FPConv doesn't require transforming to intermediate representation like 3D grid or graph and directly works on surface geometry of point cloud. To be more specific, for each point, FPConv performs a local flattening by automatically learning a weight map to softly project surrounding points onto a 2D grid. Regular 2D convolution can thus be applied for efficient feature learning. FPConv can be easily integrated into various network architectures for tasks like 3D object classification and 3D scene segmentation, and achieve comparable performance with existing volumetric-type convolutions. More importantly, our experiments also show that FPConv can be a complementary of volumetric convolutions and jointly training them can further boost overall performance into state-of-the-art results.

Raw point clouds data inevitably contains outliers or noise through acquisition from 3D sensors or reconstruction algorithms. In this paper, we present a novel end-to-end network for robust point clouds processing, named PointASNL, which can deal with point clouds with noise effectively. The key component in our approach is the adaptive sampling (AS) module. It first re-weights the neighbors around the initial sampled points from farthest point sampling (FPS), and then adaptively adjusts the sampled points beyond the entire point cloud. Our AS module can not only benefit the feature learning of point clouds, but also ease the biased effect of outliers. To further capture the neighbor and long-range dependencies of the sampled point, we proposed a local-nonlocal (L-NL) module inspired by the nonlocal operation. Such L-NL module enables the learning process insensitive to noise. Extensive experiments verify the robustness and superiority of our approach in point clouds processing tasks regardless of synthesis data, indoor data, and outdoor data with or without noise. Specifically, PointASNL achieves state-of-the-art robust performance for classification and segmentation tasks on all datasets, and significantly outperforms previous methods on real-world outdoor SemanticKITTI dataset with considerate noise.

Neural network pruning is widely applied to various mobile applications. Previous pruning methods mainly leverage ad-hoc criteria to evaluate channel importance. In this paper, we propose an effective metaheuristic sub-structure selection (MetaSelection) method for neural network pruning. MetaSelection exploits evolutionary algorithm (EA) to search the proper sub-structure satisfying the resource constraints. In comparison with previous AutoML based methods, MetaSelection can automatically achieve the pruning rate and channel selection at the same time instead of hand-crafted criteria in a cascaded way. Regarding the tremendous search space of channel selection as a combinatorial optimization problem, we further utilize a coarse-to-fine strategy and the novel probability distribution crossover (PDC) to speed up the search procedure. Besides, MetaSelection prunes the network globally rather than in a layer-by-layer way. We evaluate MetaSelection on several appealing deep neural networks, achieving superior results with adaptive depth and width. Concretely, on ImageNet, MetaSelection achieves a top-1 accuracy of 71.5% on MobileNetV2 under 70% FLOPs constraint and a FLOPs reduction of 30% with 76.4% top-1 accuracy for ResNet50.

We implement a wireless indoor tracking system based on the variational Gaussian process state-space model (GPSSM) with smartphone-collected WiFi received signal strength and inertial measurement unit readings. We adapt the existing variational GPSSM framework to wireless tracking scenarios, and provide a practical learning procedure for the variational GPSSM. The proposed system explores both the expressive power of the non-parametric Gaussian process model and its natural mechanism for integrating the state-of-the-art tracking techniques designed upon state-space model. Experimental results obtained from a real office environment validate the outstanding performance of the variational GPSSM in comparison with the traditional parametric state-space model in terms of tracking accuracy.

This paper investigates the problem of resource allocation for a wireless communication network with distributed reconfigurable intelligent surfaces (RISs). In this network, multiple RISs are spatially distributed to serve wireless users and the energy efficiency of the network is maximized by dynamically controlling the on-off status of each RIS as well as optimizing the reflection coefficient matrix of the RISs. This problem is posed as a joint optimization problem of transmit power and RIS control, whose goal is to maximize the energy efficiency under minimum rate constraints of the users. To solve this problem, an alternating algorithm is proposed by solving two sub-problems iteratively. The phase optimization sub-problem is solved by using a successive convex approximation method, which admits a closed-form solution at each step. Moreover, the RIS on-off optimization sub-problem is solved by using the dual method. Simulation results show that the proposed scheme achieves up to 27% and 68% gains in terms of the energy efficiency compared to the conventional RIS scheme and amplify-and-forward relay scheme, respectively.

In this paper, the design of an optimal trajectory for an energy-constrained drone operating in dynamic network environments is studied. In the considered model, a drone base station (DBS) is dispatched to provide uplink connectivity to ground users whose demand is dynamic and unpredictable. In this case, the DBS's trajectory must be adaptively adjusted to satisfy the dynamic user access requests. To this end, a metalearning algorithm is proposed in order to adapt the DBS's trajectory when it encounters novel environments, by tuning a reinforcement learning (RL) solution. The meta-learning algorithm provides a solution that adapts the DBS in novel environments quickly based on limited former experiences. The meta-tuned RL is shown to yield a faster convergence to the optimal coverage in unseen environments with a considerably low computation complexity, compared to the baseline policy gradient algorithm. Simulation results show that, the proposed meta-learning solution yields a 25% improvement in the convergence speed, and about 10% improvement in the DBS' communication performance, compared to a baseline policy gradient algorithm. Meanwhile, the probability that the DBS serves over 50% of user requests increases about 27%, compared to the baseline policy gradient algorithm.

In this paper, the problem of minimizing the weighted sum of age of information (AoI) and total energy consumption of Internet of Things (IoT) devices is studied. In particular, each IoT device monitors a physical process that follows nonlinear dynamics. As the dynamic of the physical process varies over time, each device must sample the real-time status of the physical system and send the status information to a base station (BS) so as to monitor the physical process. The dynamics of the realistic physical process will influence the sampling frequency and status update scheme of each device. In particular, as the physical process varies rapidly, the sampling frequency of each device must be increased to capture these physical dynamics. Meanwhile, changes in the sampling frequency will also impact the energy usage of the device. Thus, it is necessary to determine a subset of devices to sample the physical process at each time slot so as to accurately monitor the dynamics of the physical process using minimum energy. This problem is formulated as an optimization problem whose goal is to minimize the weighted sum of AoI and total device energy consumption. To solve this problem, a machine learning framework based on the repeated update Q-learning (RUQL) algorithm is proposed. The proposed method enables the BS to overcome the biased action selection problem (e.g., an agent always takes a subset of actions while ignoring other actions), and hence, dynamically and quickly finding a device sampling and status update policy so as to minimize the sum of AoI and energy consumption of all devices. Simulations with real data of PM 2.5 pollution in Beijing from the Center for Statistical Science at Peking University show that the proposed algorithm can reduce the sum of AoI by up to 26.9% compared to the conventional Q-learning method.

Traditional deep learning models are trained on centralized servers using labeled sample data collected from edge devices. This data often includes private information, which the users may not be willing to share. Federated learning (FL) is an emerging approach to train such learning models without requiring the users to share their possibly private labeled data. In FL, each user trains its copy of the learning model locally. The server then collects the individual updates and aggregates them into a global model. A major challenge that arises in this method is the need of each user to efficiently transmit its learned model over the throughput limited uplink channel. In this work, we tackle this challenge using tools from quantization theory. In particular, we identify the unique characteristics associated with conveying trained models over rate-constrained channels, and characterize a suitable quantization scheme for such setups. We show that combining universal vector quantization methods with FL yields a decentralized training system, which is both efficient and feasible. We also derive theoretical performance guarantees of the system. Our numerical results illustrate the substantial performance gains of our scheme over FL with previously proposed quantization approaches.

In this paper, the convergence time of federated learning (FL), when deployed over a realistic wireless network, is studied. In particular, with the considered model, wireless users transmit their local FL models (trained using their locally collected data) to a base station (BS). The BS, acting as a central controller, generates a global FL model using the received local FL models and broadcasts it back to all users. Due to the limited number of resource blocks (RBs) in a wireless network, only a subset of users can be selected and transmit their local FL model parameters to the BS at each learning step. Meanwhile, since each user has unique training data samples and the BS must wait to receive all users' local FL models to generate the global FL model, the FL performance and convergence time will be significantly affected by the user selection scheme. In consequence, it is necessary to design an appropriate user selection scheme that enables all users to execute an FL scheme and efficiently train it. This joint learning, wireless resource allocation, and user selection problem is formulated as an optimization problem whose goal is to minimize the FL convergence time while optimizing the FL performance. To address this problem, a probabilistic user selection scheme is proposed using which the BS will connect to the users, whose local FL models have large effects on its global FL model, with high probabilities. Given the user selection policy, the uplink RB allocation can be determined. To further reduce the FL convergence time, artificial neural networks (ANNs) are used to estimate the local FL models of the users that are not allocated any RBs for local FL model transmission, which enables the BS to include more users' local FL models to generate the global FL model so as to improve the FL convergence speed and performance. Simulation results show that the proposed ANN-based FL scheme can reduce the FL convergence time by up to 53.8%, compared to a standard FL algorithm.

Learning and processing of signals over hypergraph models have gained substantial traction owing to the ability of hypergraphs in characterizing multilateral interactions. In this work, we explore hypergraph spectral analysis and provide alternative definitions of frequency domain operations that are practically useful in image processing. We analyze hypergraph spectral properties and present several application examples, including compression, edge detection and segmentation. Successful experiment results demonstrate the effectiveness and the future prospect of the proposed hypergraph frequency operations in image processing.

In this paper, we formulate and solve a two-stage Bayesian sequential change diagnosis problem. Different from the one-stage sequential change diagnosis problem considered in the existing work, after a change has been detected, we can continue to collect samples so that we can identify the distribution after change more accurately. The goal is to minimize the total cost including delay, false alarm and mis-diagnosis probabilities. We first convert the two-stage sequential change diagnosis problem into a two-ordered optimal stopping time problem. Using tools from multiple optimal stopping time problems, we obtain the optimal change detection and distribution identification rules.

We present a novel framework for human video motion transfer. Deviating from recent studies that use only single source image, we propose to allow users to supply multiple source images by simply imitating some poses in the desired target video. To aggregate the appearance from multiple input images, we propose a JAFPro framework that incorporates two modules: an appearance fusion module that adaptively fuses the information in the supplied images and an appearance propagation module that propagates textures through flow-based warping to further improve the result. An attractive feature of JAFPro is that the quality of its results progressively improves as more imitating images are supplied. Furthermore, we build a new dataset containing a large variety of dancing videos in the wild. Extensive experiments conducted on this dataset demonstrate JAFPro outperforms state-of-the-art methods both qualitatively and quantitatively. We will release our code and dataset upon publication of this work.

Point completion refers to complete the missing geometries of objects from partial point clouds. Existing works usually estimate the missing shape by decoding a latent feature encoded from the input points. However, real-world objects are usually with diverse topologies and surface details, which a latent feature may fail to represent to recover a clean and complete surface. To this end, we propose a skeleton-bridged point completion network (SK-PCN) for shape completion. Given a partial scan, our method first predicts its 3D skeleton to obtain the global structure, and completes the surface by learning displacements from skeletal points. We decouple the shape completion into structure estimation and surface reconstruction, which eases the learning difficulty and benefits our method to obtain on-surface details. Besides, considering the missing features during encoding input points, SK-PCN adopts a local adjustment strategy that merges the input point cloud to our predictions for surface refinement. Comparing with previous methods, our skeleton-bridged manner better supports point normal estimation to obtain the full surface mesh beyond point clouds. The qualitative and quantitative experiments on both point cloud and mesh completion show that our approach outperforms the existing methods on various object categories.

In this paper, a power efficient deployment for visible light communication (VLC)-enabled unmanned aerial vehicles (UAVs) is studied. In the studied model, each UAV can provide communication service for ground users and illumination builds the VLC links between UAVs and users. Hence, each UAV’s signal transmission and illumination will affect other UAVs’ signal transmission and illumination. Therefore, to deploy VLC-enabled UAVs so as to efficiently service the ground users, the interference caused by the signal transmission and illumination of UAVs must be considered. This problem is formulated as an optimization problem whose goal is to optimize the deployment of UAVs so as to minimize the power consumption for signal transmission and illumination. An iterative algorithm is first proposed to transform the optimization problem into a series of interdependent subproblems, and the transformed problems are then solved by the Lagrangian dual method. In addition, convergence and complexity of the algorithm are also analyzed. Numerical results show that the proposed scheme can reduce at least 53.7% power consumption when compared to the baselines with UAVs at the center of each sub-area.

Inspired by the recent discoveries in neuroscience, the study of the sparse binary projection model started to attract people's attention, shedding new light on image retrieval. Different from the classical work that tries to reduce the dimension of the data for faster retrieval speed, the model projects dense input samples into a higher-dimensional space and outputs sparse binary data representations after winner-take-all competition. Following the work along this line, this paper designed a new algorithm which obtains a high-quality sparse binary projection matrix through unsupervised training. Simple as it is, the algorithm reported significantly improved results over the state-of-the-art methods in both search accuracy and retrieval speed in a series of empirical evaluations on large-scale image retrieval tasks, which exhibited its promising potential in industrial applications.

We propose a novel Gaussian process kernel that takes advantage of a deep neural network (DNN) structure but retains good interpretability. The resulting kernel is capable of addressing four major issues of the previous works of similar art, i.e., the optimality, explainability, model complexity, and sample efficiency. Our kernel design procedure comprises three steps: (1) Derivation of an optimal kernel with a non-stationary dot product structure that minimizes the prediction/test mean-squared-error (MSE); (2) Decomposition of this optimal kernel as a linear combination of shallow DNN subnetworks with the aid of multi-way feature interaction detection; (3) Updating the hyper-parameters of the subnetworks via an alternating rationale until convergence. The designed kernel does not sacrifice interpretability for optimality. On the contrary, each subnetwork explicitly demonstrates the interaction of a set of features in a transformation function, leading to a solid path toward explainable kernel learning. We test the proposed kernel with both synthesized and real-world data sets, and the proposed kernel is superior to its competitors in terms of prediction performance in most cases. Moreover, it tends to maintain the prediction performance and be robust to data over-fitting issue, when reducing the number of samples.

In the intelligent reflecting surface (IRS) assisted communication systems, the acquisition of channel state information (CSI) is a crucial impediment for achieving the passive beamforming gain of IRS because of the considerable overhead required for channel estimation. Specifically, under the current beamforming design for IRS-assisted communications, KMN + KM channel coefficients should be estimated if the passive IRS cannot estimate its channels with the base station (BS) and users due to its lack of radio frequency (RF) chains, where K, N and M denote the numbers of users, reflecting elements of the IRS, and antennas at the BS, respectively. These numbers can be extremely large in practice considering the current trend of massive MIMO (multiple-input multiple-output), i.e., a large M, and massive connectivity, i.e., a large K. To accurately estimate such a large number of channel coefficients within a short time interval, we devote our endeavour in this paper to investigating the efficient pilot-based channel estimation method in IRS-assisted uplink communications. Building upon the observation that each IRS element reflects the signals from all the users to the BS via the same channel, we analytically verify that a time duration consisting of K+N+max(K-1, [(K-1)N/M)] pilot symbols is sufficient for the BS to perfectly recover all the KMN + KM channel coefficients for the case without receiver noise. In contrast to the conventional uplink communications without IRS in which the minimum pilot sequence length is independent with the number of receive antennas, our study reveals the significant role of massive MIMO in reducing the channel training time for IRS-assisted communications.

The ultradense network (UDN) has been identified as a promising technology to address the challenge of the ever increasing demands on data rates or massive accesses, especially for Internet of Things (IoT)-oriented applications. However, the severe co-channel interference (CCI) generated by densely deployed femtocells in UDN poses a critical issue. The conflict graph is widely recognized as an effective representation of the underlying interference constraints in the network and a powerful tool for interference management. Different from most prior studies that construct conflict graphs based on accurate geographical distance information, which is usually hard to obtain in reality, an accurate and practical machine-learning-based conflict graph construction approach is proposed in this article. Based on the constructed graph, the throughput maximization problem, which is NP-hard, is decoupled into a user clustering subproblem and a subchannel allocation subproblem. The former is solved by proposing a low complexity user clustering algorithm with modified balanced Mink-Cut, which identifies low-interference entities (i.e., clusters) for spectrum reuse; and the latter is solved by presenting a subchannel allocation algorithm with accumulative intercluster interference considered, which could further reduce the interference caused by spectrum reuse. Moreover, to further improve the spectrum efficiency, a supplementary allocation algorithm is deployed to allocate the remaining subchannels. The simulation results show that the proposed approach improves the aggregate throughput by up to 186.68%, compared with the other existing methods.

Signal processing over graphs has recently attracted significant attention for dealing with the structured data. Normal graphs, however, only model pairwise relationships between nodes and are not effective in representing and capturing some high-order relationships of data samples, which are common in many applications, such as Internet of Things (IoT). In this article, we propose a new framework of hypergraph signal processing (HGSP) based on the tensor representation to generalize the traditional graph signal processing (GSP) to tackle high-order interactions. We introduce the core concepts of HGSP and define the hypergraph Fourier space. We then study the spectrum properties of hypergraph Fourier transform (HGFT) and explain its connection to mainstream digital signal processing. We derive the novel hypergraph sampling theory and present the fundamentals of hypergraph filter design based on the tensor framework. We present HGSP-based methods for several signal processing and data analysis applications. Our experimental results demonstrate significant performance improvement using our HGSP framework over some traditional signal processing solutions.

This article considers the design of the transmitter and receiver in a noncoherent massive single-input and multiple-output (SIMO) system over a multipath channel, representing a typical Internet-of-Things (IoT) scenario that consists of multiple single-antenna transmitters and one receiver with a large number of antennas. In particular, the autoencoders, which consist of multiple independent neural networks (NNs), are adopted at the transmitters and the receiver and are trained jointly, while working separately. To avoid the delicate design for mitigating the intersymbol interference (ISI) caused by multipath channels, the modulation schemes at the transmitters and the demodulation rule at the receiver are learned by the NNs over a limited number of channel samples. Moreover, the relationship between the number of channel samples and the performance of the trained transceiver is analyzed. The simulation results show that the proposed method achieves a lower error probability in comparison with the conventional optimization-based methods under typical channel conditions.

Hypergraph spectral analysis has emerged as an effective tool processing complex data structures in data analysis. The surface of a three-dimensional (3D) point cloud, and the multilateral relationship among their points can be naturally captured by the high-dimensional hyperedges. This work investigates the power of hypergraph spectral analysis in unsupervised segmentation of 3D point clouds. We estimate, and order the hypergraph spectrum from observed point cloud coordinates. By trimming the redundancy from the estimated hypergraph spectral space based on spectral component strengths, we develop a clustering-based segmentation method. We apply the proposed method to various point clouds, and analyze their respective spectral properties. Our experimental results demonstrate the effectiveness and efficiency of the proposed segmentation method.

Cyber-physical systems (CPS) are characterized by deep and complex intertwining among cyber components and physical components. Due to the fast increase in system complexities, the operations of CPS involve sensing, processing and storage of massive amount of data. This nature of “big data” imposes fundamental challenges on the design and management of CPS in multiple aspects such as performance, energy efficiency, security, privacy, reliability, sustainability, fault tolerance, scalability and flexibility. Tackling these challenges necessitates innovative big data techniques for handling massive data in CPS. The articles in this special section include a few selected state-of-the-art research results on the topic of big data sensing, processing and storage for CPS, and stimulates a broad range of researchers to participate in the interdisciplinary CPS research in the future. This special issue has received a significant number of submissions while only a small portion of them are selected for publications. The selected papers showcase how interesting data analytics techniques can be leveraged to optimize different metrics in CPS, such as timing, efficiency, schedulability, power, reliability, and security, etc.

In this paper, we propose the first sketching system for interactively personalized and photorealistic face caricaturing. Input an image of a human face, the users can create caricature photos by manipulating its facial feature curves. Our system first performs exaggeration on the recovered 3D face model, which is conducted by assigning the laplacian of each vertex a scaling factor according to the edited sketches. The mapping between 2D sketches and the vertex-wise scaling field is constructed by a novel deep learning architecture. Our approach allows outputting different exaggerations when applying the same sketching on different input figures in term of their different geometric characteristics, which makes the generated results "personalized". With the obtained 3D caricature model, two images are generated, one obtained by applying 2D warping guided by the underlying 3D mesh deformation and the other obtained by re-rendering the deformed 3D textured model. These two images are then seamlessly integrated to produce our final output. Due to the severe stretching of meshes, the rendered texture is of blurry appearances. A deep learning approach is exploited to infer the missing details for enhancing these blurry regions. Moreover, a relighting operation is invented to further improve the photorealism of the result. These further make our results "photorealistic". The qualitative experiment results validated the efficiency of our sketching system.

To meet the ever-increasing demands for mobile video streaming, proactive caching over the network edge has been proposed as a promising solution for next generation wireless networks. In this paper, we consider the trace-driven cache-enabled video streaming design in the scenario of a metropolis to boost the spectral efficiency on the system side and the quality of experience (QoE) on the user side. A novel scheme to jointly provide proactive caching, power allocation, user association and adaptive video streaming is designed via the formation of a QoE-aware throughput maximization problem. Specifically, the caches are refreshed in the content placement phase according to the resource status and expected traffic, which is obtained by exploring the traces collected over a big city. In addition, users need to be associated with a proper small base station (SBS) in the content delivering phase to provide the highest attainable rate. We demonstrate the effectiveness of the proposed scheme via experiments conducted over real user trace datasets.

In intelligent reflecting surface (IRS) assisted communication systems, the acquisition of channel state information is a crucial impediment for achieving the beamforming gain of IRS because of the considerable overhead required for channel estimation. Specifically, under the current beamforming design for IRS-assisted communications, in total KMN+KM channel coefficients should be estimated, where K, N and M denote the numbers of users, IRS reflecting elements, and antennas at the base station (BS), respectively. For the first time in the literature, this paper points out that despite the vast number of channel coefficients that should be estimated, significant redundancy exists in the user-IRS-BS reflected channels of different users arising from the fact that each IRS element reflects the signals from all the users to the BS via the same channel. To utilize this redundancy for reducing the channel estimation time, we propose a novel three-phase pilot-based channel estimation framework for IRS-assisted uplink multiuser communications, in which the userBS direct channels and the user-IRS-BS reflected channels of a typical user are estimated in Phase I and Phase II, respectively, while the user-IRS-BS reflected channels of the other users are estimated with low overhead in Phase III via leveraging their strong correlation with those of the typical user. Under this framework, we analytically prove that a time duration consisting of K + N + max(K - 1, [(K - 1)N/M]) pilot symbols is sufficient for perfectly recovering all the KMN + KM channel coefficients under the case without receiver noise at the BS. Further, under the case with receiver noise, the user pilot sequences, IRS reflecting coefficients, and BS linear minimum mean-squared error channel estimators are characterized in closed-form.

This paper studies a single-user wireless powered mobile edge computing (MEC) system, in which one multi-antenna energy transmitter (ET) employs energy beamforming for wireless power transfer (WPT) towards the user, and the user relies on the harvested energy to locally execute a portion of tasks and offload the other portion to an access point (AP) integrated with an MEC server for remote execution. Different from prior works considering static wireless channels and computation tasks at the user, this paper considers both energy and task causality constraints due to the channel fluctuations and dynamic task arrivals over time. Towards an energy-efficient joint-WPT-MEC design, we minimize the total transmission energy consumption at the ET over a particular finite horizon while ensuring the user's successful task execution, by jointly optimizing the transmission energy allocation at the ET for WPT and the task allocation at the user for local computing and offloading over a particular finite horizon. First, in order to characterize the fundamental performance limit, we consider the offline optimization by assuming that the perfect knowledge of channel state information (CSI) and task state information (TSI) (i.e., task arrival timing and amounts) is known a-priori. In this case, we obtain the well-structured optimal solution to the energy minimization problem by using convex optimization techniques. The optimal solution shows that in the scenario with static channels, the ET should allocate the transmission energy uniformly over time, and the user should employ staircase task allocation for both local computing and offloading, with the number of executed task input-bits monotonically increasing over time. It also shows that in the scenario with time-varying channels, the ET should transmit energy sporadically at slots with causally dominating channel power gains, and the user should apply the staircase task allocation for local computing and staircase water-filling task allocation for offloading with monotonically increasing computation levels over time. Next, inspired by the structured offline solutions obtained above, we develop heuristic online designs for the joint energy and task allocation when the knowledge of CSI/TSI is only causally known. Finally, numerical results show that the proposed joint energy and task allocation designs achieve significantly smaller energy consumption than benchmark schemes with only local computing or full offloading at the user, and the proposed heuristic online designs perform close to the optimal offline solutions and considerably outperform the conventional myopic designs.

We present a deep reinforcement learning method of progressive view inpainting for 3D point scene completion under volume guidance, achieving high-quality scene reconstruction from only a single depth image with severe occlusion. Our approach is end-to-end, consisting of three modules: 3D scene volume reconstruction, 2D depth map inpainting, and multi-view selection for completion. Given a single depth image, our method first goes through the 3D volume branch to obtain a volumetric scene reconstruction as a guide to the next view inpainting step, which attempts to make up the missing information; the third step involves projecting the volume under the same view of the input, concatenating them to complete the current view depth, and integrating all depth into the point cloud. Since the occluded areas are unavailable, we resort to a deep Q-Network to glance around and pick the next best view for large hole completion progressively until a scene is adequately reconstructed while guaranteeing validity. All steps are learned jointly to achieve robust and consistent results. We perform qualitative and quantitative evaluations with extensive experiments on the SUNCG data, obtaining better results than the state of the art.

In this paper, the problem of training federated learning (FL) algorithms over a realistic wireless network is studied. In particular, in the considered model, wireless users perform an FL algorithm that trains their local FL models using their own data and send the trained local FL models to a base station (BS) that will generate a global FL model and send it back to the users. Since all training parameters are transmitted over wireless links, the quality of the training will be affected by wireless factors such as packet errors and availability of wireless resources. Meanwhile, due to the limited wireless bandwidth, the BS must select an appropriate subset of users to execute the FL learning algorithm so as to build a global FL model accurately. This joint learning, wireless resource allocation, and user selection problem is formulated as an optimization problem whose goal is to minimize an FL loss function that captures the performance of the FL algorithm. To address this problem, a closed-form expression for the expected convergence rate of the FL algorithm is first derived to quantify the impact of wireless factors on FL. Then, based on the expected convergence rate of the FL algorithm, the optimal transmit power for each user is derived, under a given user selection and uplink resource block (RB) allocation scheme. Finally, the user selection and uplink RB allocation is optimized so as to minimize the FL loss function. Simulation results show that the proposed joint federated learning and communication framework can reduce the FL loss function value by up to 10% and 16%, respectively, compared to 1) an optimal user selection algorithm with random resource allocation and 2) a random user selection and resource allocation algorithm.

Energy harvesting (EH) technique has attracted great attention in Internet of things (IoT) system as it may significantly increase the network lifetime by using renewable energy sources. In this paper, we consider a simple uplink system composed of one base station (BS) and multiple EH user equipments (UEs), where the system control is modeled as a Markov decision process without any prior knowledge assumed on the energy dynamics. The central controller is the BS, which is in charge of scheduling a subset of UEs to access the limited orthogonal channels and regulating transmission power for the scheduled UEs. We propose an actor-critic deep Q-network based (DQN) reinforcement learning (RL) algorithm to handle such a technically challenging problem with continuous state and action spaces. Experiment results show that the proposed RL algorithm can achieve better performances compared with the existing benchmarks.

In this paper, we investigate the adversarial robustness of subspace learning problems. Different from the scenario addressed by classic robust algorithms that assume fractions of data are corrupted, we consider a more powerful adversary who can observe the whole data and modify all of them. The goal of the adversary is to maximize the distance between the subspace learned from the original data set and that learned from the modified data. We characterize the optimal rank-one attack strategy and show that the optimal strategy depends on the smallest singular value of the original data matrix and the adversary's energy budget.

Gaussian process (GP) for machine learning has been well studied over the past two decades and is now widely used in many sectors. However, the design of low-complexity GP models still remains a challenging research problem. In this paper, we propose a novel scalable GP regression model for processing big datasets, using a large number of parallel computation units. In contrast to the existing methods, we solve the classic maximum likelihood based hyper-parameter optimization problem by a carefully designed distributed alternating direction method of multipliers (ADMM). The proposed method is parallelizable over a large number of computation units. Simulation results confirm the benefits of the proposed scalable GP model over the state-of-the-art distributed methods.

The mobility load balancing (MLB) in self-organizing networks (SONs) is designed to automatically resolve the mismatch between network resource distribution and network traffic demand. In this paper, we propose an off-policy deep reinforcement learning (DRL) based MLB framework to balance the load distribution among all the cells. Our main contribution is three-fold. First, we propose to use off-policy RL with multiple behavior policies to autonomously learn the optimal MLB policy without any prior knowledge over the underlying wireless environments. Second, we propose a corresponding DRL-based MLB model by using deep neural networks as the function approximators to improve the generalization ability over complex system states. Third, we propose an asynchronous parallel learning framework for MLB to improve the training efficiency in a collaborative manner. Experimental results show that our proposed DRL-based MLB model can outperform the existing approaches considerably.

Distributed Gaussian Process (GP) is a scalable Bayesian method that is promising for handling big data. Our contribution in applying GP for traffic prediction is two-fold. First, to the best of our knowledge, this paper is the first to empower GP regression with the alternating direction method of multipliers (ADMM) for distributed hyper-parameter optimization in the training phase, where the ADMM training framework well balances local estimation and information consensus in a principled way. Second, in the prediction phase, we fuse local predictions obtained from distributed computing units via a cross-validation based optimal strategy, which demonstrates itself to be reliable and robust for general regression tasks. Moreover, the cross-validation based optimal fusion strategy is built upon a well acknowledged probabilistic model to retain the valuable closed-form GP prediction properties. Experimental results show that our proposed distributed GP model can outperform the state-of-the-art distributed GP models considerably, in terms of wireless traffic prediction performance.

In this paper, we propose a scheduling scheme called GreenFlowing to reduce the electricity cost for a cloud data center by leveraging heterogeneous ESDs. In our model, the data center can be powered by intermittent green energy like wind and solar, and the electricity with time-varying prices from the power grid. The energy from different sources can also choose to flow into long-term or short-term ESDs for later use. Note that, the former can sustain energy for a long time but with low charging/discharging rate, while the latter can charge/discharge very fast but with high energy leakage that can sustain energy only for a few hours. By combining them together, the energy cost can further be reduced. However, it is hard to decide when and how much energy from different sources should be used to power the data center directly or charged into different types of ESDs. We formulate our scheduling into a large-scale linear programming (LP) problem, which can be solved using CPLEX. Numerical experiments show that our scheduling can significantly reduce the total electricity cost for a cloud data center.

This paper proposes a novel user cooperation approach in both computation and communication for mobile edge computing (MEC) systems to improve the energy efficiency for latency-constrained computation. We consider a basic three-node MEC system consisting of a user node, a helper node, and an access point (AP) node attached with an MEC server, in which the user has latency-constrained and computation-intensive tasks to be executed. We consider two different computation offloading models, namely, the partial and binary offloading, respectively. For partial offloading, the tasks at the user are divided into three parts that are executed at the user, helper, and AP, respectively; while for binary offloading, the tasks are executed as a whole only at one of three nodes. Under this setup, we focus on a particular time block and develop an efficient four-slot transmission protocol to enable the joint computation and communication cooperation. Besides the local task computing over the whole block, the user can offload some computation tasks to the helper in the first slot, and the helper cooperatively computes these tasks in the remaining time; while in the second and third slots, the helper works as a cooperative relay to help the user offload some other tasks to the AP for remote execution in the fourth slot. For both cases with partial and binary offloading, we jointly optimize the computation and communication resources allocation at both the user and the helper (i.e., the time and transmit power allocations for offloading, and the central process unit frequencies for computing), so as to minimize their total energy consumption while satisfying the user's computation latency constraint. Although the two problems are nonconvex in general, we develop efficient algorithms to solve them optimally. Numerical results show that the proposed joint computation and communication cooperation approach significantly improves the computation capacity and energy efficiency at the user and helper, as compared to other benchmark schemes without such a joint design.

Energy harvesting (EH) is a promising technique to fulfill the long-term and self-sustainable operations for Internet of Things (IoT) systems. In this paper, we study the joint access control and battery prediction problems in a small-cell IoT system including multiple EH user equipments (UEs) and one base station (BS) with limited uplink access channels. Each UE has a rechargeable battery with finite capacity. The system control is modeled as a Markov decision process without complete prior knowledge assumed at the BS, which also deals with large sizes in both state and action spaces. First, to handle the access control problem assuming causal battery and channel state information, we propose a scheduling algorithm that maximizes the uplink transmission sum rate based on reinforcement learning (RL) with deep Q -network enhancement. Second, for the battery prediction problem, with a fixed round-robin access control policy adopted, we develop an RL-based algorithm to minimize the prediction loss (error) without any model knowledge about the energy source and energy arrival process. Finally, the joint access control and battery prediction problem is investigated, where we propose a two-layer RL network to simultaneously deal with maximizing the sum rate and minimizing the prediction loss: the first layer is for battery prediction, the second layer generates the access policy based on the output from the first layer. Experiment results show that the three proposed RL algorithms can achieve better performances compared with existing benchmarks.

The Internet of Things (IoT) application has a crucial need for long-term and self-sustainable operations. Energy harvesting (EH) technique has attracted great attention in IoT as it may significantly increase the network lifetime by using renewable energy sources. In this paper, we study a simple IoT system composed of one base station (BS) and multiple EH user equipments (UEs), where the system control is modeled as a Markov decision process without any prior knowledge assumed on the energy dynamics. The central controller, i.e., the BS, is in charge of scheduling a subset of UEs to access the limited orthogonal channels and regulating transmission power for the scheduled UEs. Applying reinforcement learning (RL) methods in this situation is technically challenging since the state and action spaces are continuous. With a long short-term memory (LSTM)-based algorithm to predict the UEs' battery states, we propose an actor-critic deep Q-network (DQN) RL algorithm to simultaneously deal with the access and continuous power control problem, by considering both the sum rate and prediction loss. The experimental results show that the proposed RL algorithm can achieve better performances when compared with the existing benchmarks.

In this article, we propose a deep reinforcement learning (DRL)-based mobility load balancing (MLB) algorithm along with a two-layer architecture to solve the large-scale load balancing problem for ultradense networks (UDNs). Our contribution is threefold. First, this article proposes a two-layer architecture to solve the large-scale load balancing problem in a self-organized manner. The proposed architecture can alleviate the global traffic variations by dynamically grouping small cells into self-organized clusters according to their historical loads, and further adapt to local traffic variations through intracluster load balancing afterwards. Second, for the intracluster load balancing, this article proposes an off-policy DRL-based MLB algorithm to autonomously learn the optimal MLB policy under an asynchronous parallel learning framework, without any prior knowledge assumed over the underlying UDN environments. Moreover, the algorithm enables joint exploration with multiple behavior policies, such that the traditional MLB methods can be used to guide the learning process thereby improving the learning efficiency and stability. Third, this article proposes an offline-evaluation-based safeguard mechanism to ensure that the online system can always operate with the optimal and well-trained MLB policy, which not only stabilizes the online performance but also enables the exploration beyond current policies to make full use of machine learning in a safe way. Empirical results verify that the proposed framework outperforms the existing MLB methods in general UDN environments featured with irregular network topologies, coupled interferences, and random user movements, in terms of the load balancing performance.

The cloud radio access network (C-RAN) is a promising paradigm to meet the stringent requirements of the fifth generation (5G) wireless systems. Meanwhile, the wireless traffic prediction is a key enabler for C-RANs to improve both the spectrum efficiency and energy efficiency through load-aware network managements. This paper proposes a scalable Gaussian process (GP) framework as a promising solution to achieve large-scale wireless traffic prediction in a cost-efficient manner. Our contribution is three-fold. First, to the best of our knowledge, this paper is the first to empower GP regression with the alternating direction method of multipliers (ADMM) for parallel hyper-parameter optimization in the training phase, where such a scalable training framework well balances the local estimation in baseband units (BBUs) and information consensus among BBUs in a principled way for large-scale executions. Second, in the prediction phase, we fuse local predictions obtained from the BBUs via a cross-validation-based optimal strategy, which demonstrates itself to be reliable and robust for general regression tasks. Moreover, such a cross-validation-based optimal fusion strategy is built upon a well acknowledged probabilistic model to retain the valuable closed-form GP inference properties. Third, we propose a C-RAN-based scalable wireless prediction architecture, where the prediction accuracy and the time consumption can be balanced by tuning the number of the BBUs according to the real-time system demands. The experimental results show that our proposed scalable GP model can outperform the state-of-the-art approaches considerably, in terms of wireless traffic prediction performance.

Hyperparameter optimization still remains the core issue in Gaussian processes (GPs) for machine learning. The classical hyperparameter optimization scheme based on maximum likelihood estimation is impractical for big data processing, as its computational complexity is cubic in terms of the number of data points. With the rapid development of efficient parallel data processing on ever cheaper and more powerful hardware, distributed models and algorithms will become ubiquitous. In this letter, we propose an alternative distributed GP hyperparameter optimization scheme using the efficient proximal alternating direction method of multipliers, proposed by Honget al.in 2016, and we derive the closed-form solution for the local sub-problems. In contrast to the existing schemes of similar kind, our proposed one well balances the computational load on each local machine and the communication overhead required for global consensus of the local hyperparameter estimates. The proposed scheme can work in either a synchronous or an asynchronous manner, thus very flexible to be adopted in different computing facilities. Experimental results with both synthetic and real datasets validate the outstanding performance of the proposed scheme.

In this paper, we study an uplink wireless system including N energy harvesting (EH) user equipments (UEs) and one base station (BS) with limited access channels. Each UE has a rechargeable battery with finite capacity. The system control is modeled as a Markov decision process without complete prior knowledge assumed at the BS, which also deals with large sizes in both state and action spaces. To handle such an access control problem, we propose a scheduling algorithm that maximizes the expected sum discounted uplink transmission rate based on reinforcement learning (RL) with deep Q-network (DQN) enhancement. Different from the traditional access control solutions that usually assume strong model knowledges, our goal is to achieve a more stable and balanced transmission over a long time horizon in a data-driven fashion. Finally, experiment results show that the proposed RL algorithm can achieve better performances compared with existing benchmarks.

Timely and accurate knowledge of channel state information (CSI) is necessary to support scheduling operations at both physical and network layers. In order to support pilot-free channel estimation in cell sleeping scenarios, we propose to adopt a channel database that stores the CSI as a function of geographic locations. Such a channel database is generated from historical user records, which usually can not cover all the locations in the cell. Therefore, we develop a two-step interpolation method to infer the channels at the uncovered locations. The method firstly applies the K-nearest-neighbor method to form a coarse database and then refines it with a deep convolutional neural network. When applied to the channel data generated by ray tracing software, our method shows a great advantage in performance over the conventional interpolation methods.

We investigate a nonparametric hypothesis testing problem, in which we assume a testing data stream is generated by one of a set of distributions (hypotheses), and the goal is to test which one of the multiple distributions generates the testing data stream, i.e., which hypothesis occurs. We assume that some distributions in the set are unknown with only training sequences generated by the corresponding distributions are given. We construct the generalized likelihood (GL) test, and characterize the error exponent of the maximum error probability. We show that the error exponent is captured by the Chernoff distance between each pair of distributions as well as the KL divergence between the approximated distributions (via training sequences) and the true distributions. We also show that the ratio between the lengths of training and testing sequences plays an important role in determining the error decay behavior.

In a social network, influence maximization is the problem of identifying a set of users that own the maximum influence ability across the network. In this paper, a novel credit distribution (CD) based model, termed as the multi-action CD (mCD) model, is introduced to quantify the influence ability of each user. Compared to existing models, the new model can work with practical datasets where one type of action is recorded for multiple times. Based on this model, influence maximization is formulated as a submodular maximization problem under a knapsack constraint, which is NP-hard. An efficient streaming algorithm is developed to achieve$(\frac{1}{3}-\epsilon)$approximation of the optimality. Experiments conducted on real Twitter dataset demonstrate that the mCD model enjoys high accuracy compared to the conventional CD model in estimating the total number of people who get influenced in a social network. Furthermore, compared to the greedy algorithm, the proposed single-pass streaming algorithm achieves similar performance in terms of influence maximization, while running several orders of magnitude faster.

In this paper, an asynchronous multi-user deep reinforcement learning scheme is developed to control the handover (HO) processes across multiple user equipments (UEs), in the goal of lowering the HO rate while ensuring certain system throughput. In this scheme, we use a deep neural network (DNN) as an HO controller learned by each UE via reinforcement learning in a collaborative fashion. Moreover, we use supervised learning in initializing the DNN controller before the execution of reinforcement learning to exploit what we already know with traditional HO schemes and to mitigate the negative effects of random exploration at the initial stage. Furthermore, we show that the adopted global-parameter-based framework enables us to train faster with more UEs, which could nicely address the scalability issue to support large systems. Finally, simulation results demonstrate that the proposed framework can achieve better performance than the state-of-art on-line schemes, in terms of HO rates.

Similarity search is a fundamental problem in computing science with various applications and has attracted significant research attention, especially in large-scale search with high dimensions. Motivated by the evidence in biological science, our work develops a novel approach for similarity search. Fundamentally different from existing methods that typically reduce the dimension of the data to lessen the computational complexity and speed up the search, our approach projects the data into an even higher-dimensional space while ensuring the sparsity of the data in the output space, with the objective of further improving precision and speed. Specifically, our approach has two key steps. Firstly, it computes the optimal sparse lifting for given input samples and increases the dimension of the data while approximately preserving their pairwise similarity. Secondly, it seeks the optimal lifting operator that best maps input samples to the optimal sparse lifting. Computationally, both steps are modeled as optimization problems that can be efficiently and effectively solved by the Frank-Wolfe algorithm. Simple as it is, our approach has reported significantly improved results in empirical evaluations, and exhibited its high potentials in solving practical problems.

This paper proposes a joint computation and communication cooperation approach in mobile edge computing (MEC) systems for improving the energy efficiency in mobile computing. In particular, we consider a basic three-node MEC system that consists of a user node, a helper node, and an access point (AP) node attached with an MEC server. We focus on the user's latency-constrained computation over a finite-length block and develop a four-slot protocol for implementing the joint computation and communication cooperation. Under this setup, we jointly optimize the task partition and time allocation, and the transmit power for offloading and central processing unit (CPU) frequencies of local computing at the user and the helper, so as to minimize their total energy consumption subject to the user's computation latency constraint. This problem is optimally solved via convex optimization techniques. Numerical results show that the proposed approach significantly improves the computation capacity and the energy efficiency for the user, as compared to other benchmark schemes without such a joint design.

In recent years, the explosive development of mobile communications and networking, together with the wave of Internet of Things (IoT), has led to super-complex systems, which are difficult to model and manage. At the same time, such systems are generating a large amount of data on a real-time basis, from both the user and network sides. How to utilize such data to relieve the dependence on restrictive, sometimes even unrealistic, system models are the key leading to more efficient and effective future networks, especially when under various resource constraints as in IoT systems. Fortunately, recent advancements in artificial intelligence (AI), empowered by modern machine learning algorithms, have demonstrated remarkable success in a variety of fields and are stimulating numerous data-driven approaches as well as applications. Combining the availability of big data in complex IoT communication networks and the recent advancements in AI, it now comes the time to renovate how we resolve network management issues to more efficiently and effectively fulfill the dynamic demands of network subscribers, especially in the presence of stringent network resource constraints. With the fuel (IoT data) and the engine (AI), data-driven network management will enable us to dynamically and adaptively meet the spatio-temporal network demands in the most resource-aware and resource-smart manner.

We propose a distributed Bayesian quickest detection algorithm for sensor networks, based on a random gossip inter-sensor communication structure. Without a control or fusion center, each sensor executes its local change detection procedure in a parallel and distributed fashion, interacting with its neighboring sensors via random inter-sensor communications to propagate information. By modeling the information propagation dynamics in the network as a Markov process, a two-layer large deviation analysis is presented to analyze the performance of the proposed algorithm. The first-layer analysis shows that the relation between the probability of false alarm and the conditional averaged detection delay satisfies the large deviation principle, where the distributed Kullback-Leibler information number is established as a crucial factor. The second-layer analysis studies the probability that not all observations are available at one sensor. It shows that this probability decays exponentially fast to zero as the averaged rounds of communication increases. The large deviation upper and lower bounds for the converge rate are then derived. Finally, we show that the performance of the distributed algorithm converges exponentially fast to that of the centralized optimal one.one.

In this paper, we propose a two-layer framework to learn the optimal handover (HO) controllers in possibly large-scale wireless systems supporting mobile Internet-of-Things users or traditional cellular users, where the user mobility patterns could be heterogeneous. In particular, our proposed framework first partitions the user equipments (UEs) with different mobility patterns into clusters, where the mobility patterns are similar in the same cluster. Then, within each cluster, an asynchronous multiuser deep reinforcement learning (RL) scheme is developed to control the HO processes across the UEs in each cluster, in the goal of lowering the HO rate while ensuring certain system throughput. In this scheme, we use a deep neural network (DNN) as an HO controller learned by each UE via RL in a collaborative fashion. Moreover, we use supervised learning in initializing the DNN controller before the execution of RL to exploit what we already know with traditional HO schemes and to mitigate the negative effects of random exploration at the initial stage. Furthermore, we show that the adopted global-parameter-based asynchronous framework enables us to train faster with more UEs, which could nicely address the scalability issue to support large systems. Finally, simulation results demonstrate that the proposed framework can achieve better performance than the state-of-art online schemes, in terms of HO rates.

Mobile-edge computing (MEC) and wireless power transfer (WPT) have been recognized as promising techniques in the Internet of Things era to provide massive low-power wireless devices with enhanced computation capability and sustainable energy supply. In this paper, we propose a unified MEC-WPT design by considering a wireless powered multiuser MEC system, where a multiantenna access point (AP) (integrated with an MEC server) broadcasts wireless power to charge multiple users and each user node relies on the harvested energy to execute computation tasks. With MEC, these users can execute their respective tasks locally by themselves or offload all or part of them to the AP based on a time-division multiple access protocol. Building on the proposed model, we develop an innovative framework to improve the MEC performance, by jointly optimizing the energy transmit beamforming at the AP, the central processing unit frequencies and the numbers of offloaded bits at the users, as well as the time allocation among users. Under this framework, we address a practical scenario where latency-limited computation is required. In this case, we develop an optimal resource allocation scheme that minimizes the AP's total energy consumption subject to the users' individual computation latency constraints. Leveraging the state-of-the-art optimization techniques, we derive the optimal solution in a semiclosed form. Numerical results demonstrate the merits of the proposed design over alternative benchmark schemes.

Wireless traffic prediction can effectively reduce the uncertainty in network demand and supply, and thus is a key enabler of smart management in next-generation wireless networks. To the best of our knowledge, this paper is the first to establish a wireless traffic prediction model by applying the Gaussian Process (GP) method based on real 4G traffic data. Our work is two-fold: First, based on the observed wireless traffic patterns, the kernel in our proposed GP model is designed accordingly to capture both the periodic trend and dynamic deviations; second, by leveraging the Toeplitz structure in the covariance matrix, the computational complexity of hyperparameter learning is significantly reduced from O(n3) to O(n2) and that of inference is reduced from O(n3) to O(n \log n), without any loss of prediction accuracy. Experimental results show that the proposed GP model can attain up to 97% prediction accuracy, and outperform the state-of-the-art algorithms considerably.

This paper presents HC-LBR routing, the first location-based routing (LBR) mechanism that simultaneously retains high communication efficiency and privacy for Internet of Things (IoT). Existing LBR schemes were originally designed for wireless sensor networks (WSNs). Although they offer attractive efficiency and scalability, their privacy performance presents a vulnerability when being used in IoT. This is because, unlike a conventional WSN where all nodes are owned by the same user, an IoT network has an open architecture in which nodes owned by different users are mixed and work together. Therefore, when LBR is directly used in an IoT network, the sharing of location information among alien nodes allows one user to peek into the communication privacy of other users. HC-LBR overcomes this privacy weakness by computing efficient geographic routes directly based on Hilbert-Curve-encrypted location information. HC-LBR consists of two components: A Kademlia-tree-based routing algorithm that supports efficient geographic routing in HC encrypted space, and Rand-Mix, a lightweight traffic mixer that uses multiple traffic flows to enhance the privacy of HC routing. Our simulation-based evaluation verifies the high efficiency and privacy of the proposed methods.

Integrating mobile-edge computing (MEC) and wireless power transfer (WPT) is a promising technique in the Internet of Things (IoT) era. It can provide massive low-power mobile devices with enhanced computation capability and sustainable energy supply. In this paper, we consider a wireless powered multiuser MEC system, where a multi-antenna access point (AP) (integrated with an MEC server) broadcasts wireless power to charge multiple users and each user node relies on the harvested energy to execute latency-sensitive computation tasks. With MEC, these users can execute their respective tasks locally by themselves or offload all or part of the tasks to the AP based on a time division multiple access (TDMA) protocol. Under this setup, we pursue an energy-efficient MEC-WPT system design by jointly optimizing the transmit energy beamformer at the AP, the central processing unit (CPU) frequencies and the offloaded bits at each user, as well as the time allocation among different users. In particular, we minimize the energy consumption at the AP over a particular time block subject to the computation latency and energy harvesting constraints per user. By formulating this problem into a convex framework and employing the Lagrange duality method, we obtain its optimal solution in a semi-closed form. Numerical results demonstrate the merits of the proposed joint design over alternative benchmark schemes.

Mobile cloud storage (MCS) is being extensively used nowadays to provide data access services to various mobile platforms such as smart phones and tablets. For cross-platform mobile apps, MCS is a foundation for sharing and accessing user data as well as supporting seamless user experience in a mobile cloud computing environment. However, the mobile usage of smart phones or tablets is quite different from legacy desktop computers, in the sense that each user has his/her own mobile usage pattern. Therefore, it is challenging to design an efficient MCS that is optimized for individual users. In this paper, we investigate a distributed MCS system whose performance is optimized by exploiting the fine-grained context information of every mobile user. In this distributed system, lightweight storage servers are deployed pervasively, such that data can be stored closer to its user. We systematically optimize the data access efficiency of such a distributed MCS by exploiting three types of user context information: mobility pattern, network condition, and data access pattern. We propose two optimization formulations: a centralized one based on mixed-integer linear programming (MILP), and a distributed one based on stable matching. We then develop solutions to both formulations. Comprehensive simulations are performed to evaluate the effectiveness of the proposed solutions by comparing them against their counterparts under various network and context conditions.

In the past decade, the smart phone evolution has accelerated the proliferation of the mobile Internet and spurred a new wave of mobile applications, leading to an unprecedented mobile data volume generated from the mobile devices, content servers, and network operators, which are mainly nonstructured. In this big data era, such nonstructured data fragments are pieced together such that, drastically differing from the traditional practice where services determine and define the data, data is becoming a proactive entity that may drive and even create new services. Compared with the so-termed 5V characteristics of generic big data, namely volume, variety, velocity, veracity, and value, mobile big data is distinct in its unique multidimensional, personalized, multisensory, and real-time features. In this survey, we provide in-depth and comprehensive coverage on the features, sources and applications of mobile big data, as well as the current state-of-the-art, challenges and opportunities for research and development in this field, with an emphasis on the user modeling, infrastructure supporting, data management, and knowledge discovery aspects.

Energy harvesting communication has raised great research interests due to its wide applications and its feasibility of commercialization. In this paper, the multiuser energy diversity is investigated in energy harvesting communication systems. Considering centralized access schemes, the scaling of the average throughput over the number of transmitters is studied, along with the scaling of corresponding available energy in the batteries that store the harvested energy. It is shown that the throughput gain mainly comes from two aspects: the increase of total available energy harvested over time/space; and the combined dynamics of batteries that lead to the improvement in effective transmission power.

Wideband spectrum sensing is a critical functionality in cognitive radio networks to enable dynamic spectrum sharing, but entails a major implementation challenge in compact commodity radios with restricted energy and computation capabilities. Exploiting jointly sparse nature of multiband signals, this paper proposes an efficient blind sub-Nyquist cooperative wideband spectrum sensing scheme, which reduces energy consumption in wideband signal acquisition, processing and transmission, with performance guarantee. In contrast to traditional sub-Nyquist approaches where a wideband signal or its power spectrum is first reconstructed from compressed samples, the proposed scheme locates occupied channels by recovering the signal support jointly from multiple secondary user (SU) measurements. Based on subspace decomposition, the low-dimensional measurement matrix computed at each SU from local sub-Nyquist samples can reduce transmission overhead while improving noise robustness. Numerical analysis and simulation results show that the proposed scheme can achieve good detection performance as well as reduce computation and implementation complexity in comparison with conventional cooperative wideband spectrum sensing schemes.

In distributed inference, cooperation among networked agents can be exploited to enhance the performance of each individual agent. In this paper, we consider signal classification over a network of agents, where each agent observes a certain signal under a particular signal-to-noise ratio (SNR). Each agent produces a statistic that summarizes its observations over a time period and then forwards it to a fusion center for identifying the type of signal in a global manner. A linear cooperation strategy for signal classification is formulated as maximizing the classification probability subject to constrained misclassification probabilities. We show that this problem can be transformed into a convex problem under some conditions and linear cooperation is a simple but effective strategy that can greatly enhance the performance of signal classification over networked agents.

Receiver sensitivity is one of the most important parameters in determining the overall performance of a communication system. Industry has been using exhaustive search in a specified range to identify the achieved receiver sensitivity for wireless devices. Such an exhaustive search scheme is neither efficient in terms of measurement time, nor accurate due to the fixed stepsize. In this paper, we propose a fast sensitivity measurement method for wireless communication systems based on bisection search, where the number of packets needed to estimate the packet error rate (PER) is optimized for the required confidence level. Specifically, we present a new analytic approach to derive the minimum number of packets needed to satisfy the desired confidence level, with the number of test packets further adapted to speed up the measurement of the receiver sensitivity.

In this special issue, we cover some recent results in the following four emerging areas: 5G cellular systems, big data systems, bio/nano/molecular networks, and smart grids. In the past several years, there are various technologies emerging, which are either directly or indirectly related to communication. Some of them are over the evolution of traditional communication systems, while others are over new systems such as smart grids, molecular networks, and big data systems.

Multicast transmission based on real-time network state information is a resource-friendly technique to improve the energy efficiency and reduce the traffic burden for cellular systems. This paper evaluates the effectiveness of this technique for downlink transmissions. In particular, a scenario is considered in which multiple mobile users (MUs) asynchronously request to download one common message locally cached at a base station (BS). Due to the randomness of both the channel conditions and the request arrivals from the MUs, the BS may choose to intelligently hold the arrived requests, especially when the channel conditions are bad or the number of requests is small, and then serve them in one shot later via multicasting. Clearly it is of great interest to balance the delay (incurred by holding the requests) and the energy efficiency (EE, defined as the energy cost per request), and this motivates us to quantify the fundamental tradeoff for the proposed “hold-then-serve” scheme. For the scenario with single channel and unit message sizes, it is shown that for a fixed channel bandwidth, the delay-EE tradeoff reduces to judiciously choosing the optimal stopping rule for when to serve all the arrived requests, where the effect of the bandwidth on the achievable delay-EE region is discussed further. By using optimal stopping theory, it is shown that the optimal stopping rule exists for general Markov channel models and request arrival processes. Particularly, for the hard deadline and proportional delay penalty cases, it is shown that the optimal stopping rule exhibits a threshold structure, and the corresponding threshold in the former case is time varying while in the latter case it is a constant. Finally, for the more general scenario with multiple channels and arbitrary message sizes, the optimal scheduling is formulated as a Markov decision process problem, where some efficient suboptimal scheduling algorithms are proposed.

The rising popularity of wireless services resulting in spectrum shortage has motivated dynamic spectrum sharing to facilitate efficient usage of the underutilized spectrum. Wideband spectrum sensing is a critical functionality to enable dynamic spectrum access by enhancing the opportunities of exploring spectral holes, but entails a major implementation challenge in compact commodity radios that only have limited energy and computation capabilities. In contrast to the traditional sub-Nyquist approaches where a wideband signal or its power spectrum is first reconstructed from compressed samples, this paper proposes a sub-Nyquist wideband spectrum sensing scheme that locates occupied channels blindly by recovering the signal support, based on the jointly sparse nature of multiband signals. Exploiting the common signal support shared among multiple secondary users (SUs), an efficient cooperative spectrum sensing scheme is developed, in which the energy consumption on wideband signal acquisition, processing, and transmission is reduced with detection performance guarantee. Based on subspace decomposition, the low-dimensional measurement matrix, computed at each SU from local sub-Nyquist samples, is deployed to reduce the transmission and computation overhead while improving noise robustness. The theoretical analysis of the proposed sub-Nyquist wideband sensing algorithm is derived and verified by numerical analysis and further tested on real-world TV white space signals. It shows that the proposed scheme can achieve good detection performance as well as reduce the computation and implementation complexity, in comparison with the conventional cooperative wideband spectrum sensing schemes.

This paper considers a heterogeneous ad hoc network with multiple transmitter-receiver pairs, in which all transmitters are capable of harvesting renewable energy from the environment and compete for one shared channel by random access. In particular, we focus on two different scenarios: the constant energy harvesting (EH) rate model where the EH rate remains constant within the time of interest and the i.i.d. EH rate model where the EH rates are independent and identically distributed across different contention slots. To quantify the roles of both the energy state information (ESI) and the channel state information (CSI), a distributed opportunistic scheduling (DOS) framework with two-stage probing and save-then-transmit energy utilization is proposed. Then, the optimal throughput and the optimal scheduling strategy are obtained via one-dimension search, i.e., an iterative algorithm consisting of the following two steps in each iteration: First, assuming that the stored energy level at each transmitter is stationary with a given distribution, the expected throughput maximization problem is formulated as an optimal stopping problem, whose solution is proven to exist and then derived for both models; second, for a fixed stopping rule, the energy level at each transmitter is shown to be stationary and an efficient iterative algorithm is proposed to compute its steady-state distribution. Finally, we validate our analysis by numerical results and quantify the throughput gain compared with the best-effort delivery scheme.

Energy harvesting (EH) based communication has raised great research interests due to its wide applications and the feasibility of commercialization. In this paper, we consider wireless communications with EH constraints at the transmitter. First, for delay-tolerant traffic, we investigate the long-term average throughput maximization problem and analytically compare the throughput performance against that of a system supported by conventional power supplies. Second, for delay-sensitive traffic, we analyze the outage probability by studying its asymptotic behavior in the high energy arrival rate regime, where the new concept of energy diversity is formally introduced. Moreover, we show that the speed of outage probability approaching zero, termed energy diversity gain, varies under different power supply models.

This paper studies a Quantized Gossip-based Interactive Kalman Filtering (QGIKF) algorithm implemented in a wireless sensor network, where the sensors exchange their quantized states with neighbors via inter-sensor communications. We show that with the information loss due to quantization, the network can still achieve weak consensus, i.e., the estimation error variance sequence at a randomly selected sensor can converge weakly (in distribution) to a unique invariant measure. To prove the weak convergence, we first interpret the error variance sequence evolution as the interacting particle, then formulate the sequence as a Random Dynamical System (RDS), and finally prove that it is stochastically bounded.

In cognitive radio (CR) technology, the trend of sensing is no longer to only detect the presence of active primary users. A large number of applications demand for primary user behavior correlation in spatial, temporal, and frequency domains. To satisfy such requirements, we study the statistical relationship of primary users by introducing a Bayesian network (BN) based framework. How to learn such a BN structure is a long standing issue, not fully understood even in the statistical learning community. To solve such an issue in CR, this paper proposes a BN structure learning scheme which incurs significantly lower computational complexity compared with previous ones. Thus, with this scheme, cognitive users could efficiently understand the statistical pattern of primary networks.

This demo presents WiMAC, a general-purpose wireless testbed for researchers to quickly prototype a wide variety of real-time MAC protocols for wireless networks. As the interface between the link layer and the physical layer, MAC protocols are often tightly coupled with the underlying physical layer, and need to have extremely small latencies. Implementing a new MAC requires a long time. In fact, very few MACs have ever been implemented, even though dozens of new MAC protocols have been proposed. To enable quick prototyping, we employ the mechanism vs. policy separation to decompose the functionality in the MAC layer and the PHY layer. Built on the separation framework, WiMAC achieves the independence of the software from the hardware, offering a high degree of function reuse and design flexibility. Hence, our platform not only supports easy cross-layer design but also allows protocol changes on the fly. Following the 802.11-like reference design, we demonstrate that deploying a new MAC protocol is quick and simple on the proposed platform through the implementation of the CSMA/CA and CHAIN protocols.

In wireless communication, the amplitude and phase of the transmitted signal have certain patterns during one symbol duration, which introduces high correlation among the samples obtained by oversampling at the receiver. In this work, we aim to explore such correlation information for cognitive radios to enhance the performance of spectrum sensing. By jointly considering the signal modulation, multipath fading, and oversampling rate, we derive the distribution of the empirical autocorrelation function for the obtained samples, on which we propose two efficient spectrum-sensing algorithms, and then analyze their performance. Our theoretical results reveal that the proposed algorithms with oversampling perform strictly better than the conventional energy detection scheme, while requiring the same level of prior information. Finally, we show through simulations that the derived statistical characteristics approximate the true statistical distribution of the autocorrelation function well, and the proposed sensing algorithms significantly improve the sensing performance compared to several existing sensing schemes.

This paper investigates the resource allocation problem for the Gaussian multiple access channel (MAC) with conferencing links, where the two transmitters can talk to each other via wired rate-limited channels. Moreover, the two transmitters are powered by a shared energy harvester which captures energy from the environment. We consider both the non-causal (the energy arrival levels at future time slots are known before transmissions) and the causal (only the energy arrival levels of past and present slots are known) energy-harvesting (EH) models. For the non-causal case, we formulate a resource allocation problem over a finite horizon ofNtime slots to characterize the boundary of the maximum departure region. We then develop the optimal offline power and rate allocation scheme by exploiting the hidden convexity of this problem. Interestingly, it is shown that there exists a maximum transmission rate (named the capping rate) for one of the transmitters. For the causal case, we examine the performance of the greedy scheme, in which the energy is depleted within each slot. In particular, we measure the utility of this scheme against the optimal offline one by competitive analysis, where the competitive ratio of the online greedy scheme, i.e., the maximum ratio between the profits obtained by the offline and online schemes over arbitrary energy arrival profiles, is derived.

This paper studies the convergence of the estimation error process and the characterization of the corresponding invariant measure in distributed Kalman filtering for potentially unstable and large linear dynamic systems. A gossip network protocol termed modified gossip interactive Kalman filtering (M-GIKF) is proposed, where sensors exchange their filtered states (estimates and error covariances) and propagate their observations via intersensor communications of rate$\bar {\gamma }$;$\bar {\gamma }$is defined as the averaged number of intersensor message passages per signal evolution epoch. The filtered states are interpreted as stochastic particles swapped through local interaction. This paper shows that the conditional estimation error covariance sequence at each sensor under M-GIKF evolves as a random Riccati equation (RRE) with Markov modulated switching. By formulating the RRE as a random dynamical system, it is shown that the network achieves weak consensus, i.e., the conditional estimation error covariance at a randomly selected sensor converges weakly (in distribution) to a unique invariant measure. Further, it is proved that as$\bar {\gamma } \rightarrow \infty $this invariant measure satisfies the large deviation (LD) upper and lower bounds, implying that this measure converges exponentially fast (in probability) to the Dirac measure$\delta _{P^{*}}$, where$P^{*}$is the stable error covariance of the centralized (Kalman) filtering setup. The LD results answer a fundamental question on how to quantify the rate at which the distributed scheme approaches the centralized performance as the intersensor communication rate increases.

An explicit characterization of the capacity region of the general network coding problem is one of the best known open problems in information theory. A simple set of bounds that is often used in the literature to show that certain rate tuples are infeasible are based on the graph-theoretic notion of cut. The standard cut-set bounds, however, are known to be loose in general when there are multiple messages to be communicated in the network. This paper focuses on broadcast networks, for which the standard cut-set bounds are closely related to union as a specific set operation to combine different simple cuts of the network. A new set of explicit network coding bounds, which combine different simple cuts of the network via a variety of set operations (not just the union), are established via their connections to extremal inequalities for submodular functions. The tightness of these bounds are demonstrated via applications to combination networks.

We study the diversity-multiplexing tradeoff (DMT) for the slow fading interference channel with a relay (ICR). We derive four inner bounds on the DMT region: the first is based on the compress-and-forward (CF) relaying scheme, the second is based on the decode-and-forward (DF) relaying scheme, and the last two bounds are based on the half-duplex (HD) and full-duplex (FD) amplify-and-forward (AF) schemes. For the CF and DF schemes, we find conditions on the channel parameters and the multiplexing gains, under which the corresponding inner bound achieves the optimal DMT region. We also identify the cases in which the DMT region of the ICR corresponds to that of two parallel slow fading relay channels, implying that interference does not decrease the DMT for each pair, and that a single relay can be DMT-optimal for two pairs simultaneously. For the HD-AF scheme, we derive conditions on the channel coefficients under which the proposed scheme achieves the optimal DMT for the AF-based relay channel. Finally, we identify the conditions under which adding a relay strictly enlarges the DMT region relative to the interference channel without a relay.

In this paper, we present a methodology employing statistical analysis and stochastic geometry to study geometric routing schemes in wireless ad hoc networks. In particular, we analyze the network-layer performance of one such scheme, the random [ 1/ 2]disk routing scheme, which is a localized geometric routing scheme in which each node chooses the next relay randomly among the nodes within its transmission range and in the general direction of the destination. The techniques developed in this paper enable us to establish the asymptotic connectivity and the convergence results for the mean and variance of the routing path lengths generated by geometric routing schemes in random wireless networks. In particular, we approximate the progress of the routing path toward the destination by a Markov process and determine the sufficient conditions that ensure the asymptotic connectivity for both dense and large-scale ad hoc networks deploying the random [ 1/ 2]disk routing scheme. Furthermore, using this Markov characterization, we show that the expected length (hop count) of the path generated by the random [ 1/ 2]disk routing scheme normalized by the length of the path generated by the ideal direct-line routing, converges to 3π/4 asymptotically. Moreover, we show that the variance-to-mean ratio of the routing path length converges to 9π2/64-1 asymptotically. Through simulation, we show that the aforementioned asymptotic statistics are in fact quite accurate even for finite granularity and size of the network.

Power control in a digital handset is practically implemented in a discrete fashion, and usually, such a discrete power control (DPC) scheme is suboptimal. In this paper, we first show that in a Poison-distributed ad hoc network, if DPC is properly designed with a certain condition satisfied, it can strictly work better than no power control (i.e., users use the same constant power) in terms of average signal-to-interference ratio, outage probability, and spatial reuse. This motivates us to propose an N-layer DPC scheme in a wireless clustered ad hoc network, where transmitters and their intended receivers in circular clusters are characterized by a Poisson cluster process on the plane ℝ2. The cluster of each transmitter is tessellated into N-layer annuli with transmit power Piadopted if the intended receiver is located at the ith layer. Two performance metrics of transmission capacity (TC) and outage-free spatial reuse factor are redefined based on the N-layer DPC. The outage probability of each layer in a cluster is characterized and used to derive the optimal power scaling law Pi∈ Θ(ηi-(α/2)), with ηias the probability of selecting power Pi and α as the path loss exponent. Moreover, the specific design approaches to optimize Pi and N based on ηiare also discussed. Simulation results indicate that the proposed optimal N-layer DPC significantly outperforms other existing power control schemes in terms of TC and spatial reuse.

In this paper we consider a cognitive radio network with access to N licensed primary frequency bands and their usage statistics, where the decentralized secondary users are subject to certain inter-network interference constraint. In particular, to limit the interference to the primary network, secondary users are equipped with spectrum sensors and are capable of sensing and accessing a limited number of channels at the same time due to hardware limitations. We consider both the error-free and erroneous spectrum sensing scenarios, and establish the jointly optimal random sensing and access scheme, which maximizes the secondary network expected sum throughput while honoring the primary interference constraint. We show that under certain conditions the optimal sensing and access scheme is independent of the primary frequency bandwidths and usage statistics; otherwise, they follow water-filling-like strategies. Moreover, we show that the performance of the secondary network depends on the ratio between the “opportunity-detection” probability and the “mis-detection” probability if the former is larger; otherwise, it depends on the ratio between the “false-alarm” probability and the “detection” probability. Finally, we demonstrate a binary behavior for the optimal access scheme at each channel, depending on whether the opportunity-detection probability or mis-detection probability is larger in that channel.

This paper considers an ad hoc network with multiple transmitter-receiver pairs, in which all transmitters are capable of harvesting renewable energy from the environment and compete for the same channel by random access. To quantify the roles of both the energy state information (ESI) and the channel state information (CSI), a distributed opportunistic scheduling (DOS) framework with a save-then-transmit scheme is proposed. First, in the channel probing stage, each transmitter probes the CSI via channel contention; next, in the data transmission stage, the successful transmitter decides to either give up the channel (if the expected reward calculated over the CSI and ESI is small) or hold and utilize the channel by optimally exploring the energy harvesting and data transmission tradeoff. With a constant energy arrival model, i.e., the energy harvesting rate keeps identical over the time of interest, the expected throughput maximization problem is formulated as an optimal stopping problem, whose solution is shown to exist and have a threshold-based structure, for both the homogeneous and heterogenous cases. Furthermore, we prove that there exists a steady-state distribution for the stored energy level at each transmitter, and propose an efficient iterative algorithm for its computation. Finally, we show via numerical results that the proposed scheme can achieve a potential 175% throughput gain compared with the method of best-effort delivery.

This paper provides a polyhedral description of the symmetrical latency capacity region of broadcast channels. The converse result is established via the recently proposed generalized cut-set bounds for broadcast networks. The achievability result is proved by considering a “pairwise exchange” scheme and successive encoding.

In beamforming, channel state information (CSI) is used to design the beamforming vector (or matrix). Since in a practical system the CSI always needs to be estimated, channel estimation (CE) errors are inevitable, which could severely affect the performance of a beamforming scheme. In this paper, we present a novel beamforming method for two-way relay (TWR) systems that is robust against CE errors. The proposed method obtains a sub-optimal solution for the associated non-convex robust optimization problem by solving a set of closed-form linear equations. Simulations show a considerable performance gain over the rank-one relaxation-based semidefinite programming (SDP) solutions, especially for the cases where the relaxed problem becomes infeasible. In addition, there is significant reduction in complexity, making this method very attractive for practical implementation.

We study the asymptotic performance of two multi-hop overlaid ad-hoc networks that utilize the same temporal, spectral, and spatial resources based on random access schemes. The primary network consists of Poisson distributed legacy users with density λ(p)and the secondary network consists of Poisson distributed cognitive radio users with density λ(s)= (λ(p))β(β > 0, β ≠ 1) that utilize the spectrum opportunistically. Both networks are decentralized and employ ALOHA medium access protocols where the secondary nodes are additionally equipped with range-limited perfect spectrum sensors to monitor and protect primary transmissions. We study the problem in two distinct regimes, namely β > 1 and 01. On the contrary, spectrum sensing turns out to be unnecessary when β <; 1 and employing spectrum sensors cannot improve the network performances.

This paper studies the optimal power allocation for outage probability minimization in point-to-point fading channels with the energy-harvesting constraints and channel distribution information (CDI) at the transmitter. Both the cases with non-causal and causal energy state information (ESI) are considered, which correspond to the energy-harvesting (EH) rates being known and unknown prior to the transmissions, respectively. For the non-causal ESI case, the average outage probability minimization problem over a finite horizon of N EH periods is shown to be non-convex for a large class of practical fading channels. However, the globally optimal "offline" power allocation is obtained by a forward search algorithm with at most N one-dimensional searches, and the optimal power profile is shown to be non-decreasing over time and have an interesting "save-then-transmit" structure. In particular, for the special case of N=1, our result revisits the classic outage capacity for fading channels with uniform power allocation. Moreover, for the case with causal ESI, we propose both the optimal and suboptimal "online" power allocation algorithms, by applying the technique of dynamic programming and exploring the structure of optimal offline solutions, respectively.

Spectrum sensing is an important building block to realize the cognitive radio concept. In order to combat fading in the wireless environment, cooperation among the sensing users is usually employed. In this paper, we develop a closed-form optimal local decision threshold for cooperative spectrum sensing in cognitive radio systems via large deviation analysis. The resultant strategy is independent of the total number of cooperating users. We show that it is not only asymptotically optimal when the number of sensing users approaches infinity, but also can achieve the maximum diversity. Numerical results are provided to verify our analysis.

This paper considers the use of energy harvesters for cooperative relaying in a large relay network, which consists of N energy-harvesting (EH) relays and one source-destination pair. In particular, a threshold-based “save-then-transmit” scheme is employed at the relays, where each relay transmits only when both the backward and forward link channel coefficients are above certain thresholds. We assume that the time scale of EH is much larger than that of communication blocks. For general channel fading models, we derive the asymptotic average throughput for the case with many relays, by using the amplify-and-forward (AF) relaying scheme. The throughput maximization is cast as a joint optimization problem over the transmission thresholds corresponding to all possible harvested energy rate states, which is shown to be non-convex in general. By applying a convexification technique via randomization, the original problem is transformed into a new formulation with a generalized threshold-based transmission scheme, which is shown to be efficiently solvable by bisection search, with the help of an offline look-up table only related to the channel statistics. Finally, with some numerical experiments, we demonstrate the performance gain of the proposed threshold-based transmission scheme against some suboptimal ones.

A discrete power control scheme based on the locations of receivers is proposed in a wireless clustered ad hoc network. Transmitters in the network form a homogeneous Poisson point process (PPP) on the plane ℝ2, and each of them has a random number of receivers that are distributed in a cluster with N tessellated layers. There are N discrete transmit powers used by each transmitter and which power is used depends on which layer the desired receiver is located in a cluster. The outage probabilities for receivers at different layers are derived. The optimal power control scheme of maximizing the redefined transmission capacity (TC) in our setting is found and it has the effect of balancing the spatial throughput achieved by receivers at all layers such that throughput fairness among receivers is achieved. Simulation results show that the proposed discrete power control significantly improves TC and outperforms other previous power control schemes in a Poisson-distributed wireless ad hoc network.

This paper considers joint estimation with multiple sensors powered by energy harvesters in wireless sensor networks. In particular, we focus on a network with K sensor nodes, which communicate with a fusion center via K orthogonal channels and power themselves by harvesting energy from the environment. Assuming a deterministic energy-harvesting model under which the harvested energy profile is known prior to transmission, the worst-case mean-square error (MSE) minimization problem over a finite horizon of T estimation periods is investigated. We consider the cases that the sensors have either infinite or finite battery capacity, and develop efficient iterative algorithms to compute the optimal power allocation strategy, with numerical results presented to validate our analysis.

This paper investigates the optimal resource allocation for the Gaussian multiple access channel (MAC) with conferencing links, where the two transmitters could talk to each other via some wired rate-limited channels. Moreover, the two transmitters are assumed to be powered by a shared energy harvester, and a deterministic energy-harvesting (EH) model is adopted by assuming that the energy arrival times and the corresponding harvested amounts are non-causally known prior to transmissions. We formulate a continuous-time power allocation problem to characterize the maximum departure region over a finite time horizon. By exploiting its convexity, this problem is simplified as a discrete-time problem and the optimal solution is obtained. In particular, it is shown that there exists a certain maximum possible transmission rate (the capping rate) at one of the transmitters. Finally, we compare the performance of the optimal offline algorithm against that of the online one.

We study the diversity-multiplexing tradeoff (DMT) for the slow fading interference channel with a relay (ICR). We first derive an outer bound on the DMT based on the cut-set bound. We then derive two inner bounds on the DMT: One is based on the compress-and-forward relaying scheme and the other is based on the decode-and-forward relaying scheme. We find conditions on the channel parameters and the multiplexing gains under which the proposed inner bounds achieve the optimal DMT. We also identify cases in which the DMT of the ICR is the same as two parallel fading relay channels, implying that interference does not decrease the DMT for each pair, and that a single relay can be DMT-optimal for two pairs simultaneously. Lastly, we identify conditions under which adding a relay strictly improves the DMT relative to the interference channel without a relay.

This paper investigates the optimal resource allocation for a Gaussian multiple access channel (MAC) with two transmitters powered by a shared energy harvester. A deterministic energy-harvesting (EH) model is adopted, which assumes that the energy arrival amounts and timing are non-causally known before transmissions. Besides, packets for both transmitters are assumed always ready before transmissions. We first formulate the resource allocation problem to characterize the maximum departure region over a finite time horizon as a convex optimization problem. The structural properties of the sum power profile is then studied by exploiting the convexity of the sum power function, which simplifies the optimization problem. Finally, the optimal resource allocation between the two transmitters, in which there exists a cut-off rate at the stronger transmitter, is obtained. We also demonstrate that under the same energy arrival profile, MAC with shared energy harvester achieves the same maximum departure region as its dual broadcast channel (BC).

This paper considers the use of energy harvesters, instead of conventional time-invariant energy sources, in wireless cooperative communication. For the purpose of exposition, we study the classic three-node Gaussian relay channel with decode-and-forward (DF) relaying, in which the source and relay nodes transmit with power drawn from energy-harvesting (EH) sources. Assuming a deterministic EH model under which the energy arrival time and the harvested amount are known prior to transmission, the throughput maximization problem over a finite horizon of N transmission blocks is investigated. In particular, two types of data traffic with different delay constraints are considered: delay-constrained (DC) traffic (for which only one-block decoding delay is allowed at the destination) and no-delay-constrained (NDC) traffic (for which arbitrary decoding delay up to N blocks is allowed). For the DC case, we show that the joint source and relay power allocation over time is necessary to achieve the maximum throughput, and propose an efficient algorithm to compute the optimal power profiles. For the NDC case, although the throughput maximization problem is non-convex, we prove the optimality of a separation principle for the source and relay power allocation problems, based upon which a two-stage power allocation algorithm is developed to obtain the optimal source and relay power profiles separately. Furthermore, we compare the DC and NDC cases, and obtain the sufficient and necessary conditions under which the NDC case performs strictly better than the DC case. It is shown that NDC transmission is able to exploit a new form of diversity arising from the independent source and relay energy availability over time in cooperative communication, termed "energy diversity", even with time-invariant channels.

In wireless networks, communication links may be subject to random fatal impacts: for example, sensor networks under sudden power losses or cognitive radio networks with unpredictable primary user spectrum occupancy. Under such circumstances, it is critical to quantify how fast and reliably the information can be collected over attacked links. For a single channel subject to random attacks, named as a dying channel, we model it as a block-fading (BF) channel with a finite and random channel length. For this channel, we first study the outage capacity and the outage probability when the data frame length is fixed and uniform power allocation is assumed. Furthermore, we discuss the optimization over the frame length and/or the power allocation over the constituting data blocks to minimize the outage probability. In addition, we extend the results from the single dying channel to the parallel multi-channel case where each sub-channel is a dying channel, and investigate the asymptotic behavior of the overall outage probability as the number of sub-channels goes to infinity with two different attack models: the independent-attack case and the m-dependent-attack case. It is shown that the asymptotic outage probability diminishes to zero for both cases as the number of sub-channels increases if the rate per unit cost is less than a certain threshold. The outage exponents are also studied to reveal how fast the outage probability improves with the number of sub-channels.

In this paper, the Gaussian diamond relay channel is considered, which consists of one source-destination pair and two relay nodes connected with rate-limited out-of-band conferencing links. In particular, we focus on the half-duplex alternative relaying strategy, in which the two relays operate alternatively over time. With different amounts of delay, two conferencing strategies are proposed, each of which can be implemented by either a general two-side conferencing scheme (for which both of the two conferencing links are used) or a special-case one-side conferencing scheme (for which only one of the two conferencing links is used). Based on the most general two-side conferencing scheme, we derive the achievable rates by using the decode-and-forward (DF) relaying scheme. By further exploiting the properties of the optimal solutions, the simpler one-side conferencing is shown to achieve the same rate as the two-side conferencing in term of the achievable rates under arbitrary channel conditions. Based on this result, the DF rate in closed-form is obtained, and the principle to use which one of the two conferencing links for one-side conferencing is also established. Moreover, the DF scheme is shown to be upper-bound-achieving under certain relay scheduling and conferencing strategy. Finally, numerical results are provided to validate our analysis.

In this paper, we consider a two-hop interference network, which consists of two source-destination pairs and two relay nodes connected with signal-to-noise ratio (SNR) limited out-of-band conferencing links. Assuming that the amplify-and-forward (AF) relaying scheme is adopted, this network is shown to be equivalent to a two-user interference channel (IC). By deploying two IC decoding schemes, i.e., single-user decoding and joint decoding, respectively, we characterize the achievable rate regions with a two-stage iterative optimization method: First, we fix the source power pair and maximize the sum rate over the relay combining vector; second, we fix the relay combining vector and optimize the source power pair. Specifically, for single-user decoding, we design a new routine to compute the optimal solution for the first subproblem, which is more efficient than the existing scheme; and for the second subproblem, we develop an iterative algorithm, with the closed-form solution for each iteration. Furthermore, it is revealed that the AF scheme with relay conferencing achieves the full degree-of-freedom (DoF), which outperforms the case without relay conferencing. Finally, simulation results show that relay conferencing can significantly improve the system performance under certain channel conditions.

In this paper, we study the state-dependent two-user Gaussian interference channel, where the Gaussian distributed state information is non-causally known at both transmitters but known to neither of the receivers. We apply the simultaneous encoding scheme and propose an active interference cancellation mechanism, which is a generalized dirty-paper coding technique, to partially eliminate the state effect at the receivers. The corresponding achievable rate region is then derived. We also propose several heuristic schemes for some special cases: the strong interference case, the mixed interference case, and the weak interference case. For the strong and mixed interference cases, numerical results are provided to show that active interference cancellation significantly enlarges the achievable rate region. For the weak interference case, flexible power splitting instead of active interference cancellation improves the performance significantly.

This paper revisits the optimal power allocation for outage minimization in the classic point-to-point fading channels with the channel distribution information (CDI) known at the transmitter. The channel state information (CSI) is assumed to be perfectly known at the receiver, but not available at the transmitter. In particular, we consider a finite horizon of N-block transmissions subject to an average power constraint at the transmitter. Although minimizing the time-averaged per-block outage probability over each N-block transmission is shown to be a non-convex problem for a large class of practical fading channels, we show that the globally optimal power allocation is obtainable by a simple one-dimensional search. It is shown that if the average transmit power is above a certain threshold determined by the distribution of the fading channel and the target transmission rate, the uniform power allocation is optimal; otherwise, an on-off power allocation is optimal. Moreover, a suboptimal low-complexity power allocation scheme is proposed, which is shown to be asymptotically optimal as N goes to infinity. Finally, numerical results are provided to validate our analysis.

In this paper, we study the three-node Gaussian relay channel with decode-and-forward (DF) relaying, in which the source and relay nodes transmit with power drawn from energy-harvesting sources. Assuming a deterministic energy-harvesting model under which the energy arrival time and the harvested amount are known prior to transmission, the throughput maximization problem over a finite horizon of N transmission blocks is investigated. We consider the nodelay-constrained (NDC) traffic case, for which the relay can store the decoded information from the source with arbitrary delay before forwarding it to the destination in each N-block transmission. Although the formulated problem is non-convex, we prove the optimality of a separation principle for the source and relay power allocation over time, based upon which a two-stage algorithm is developed to obtain the optimal source and relay power profiles separately.

The paper considers the problem of distributed estimation of an unknown deterministic scalar parameter (the target signal) in wireless sensor networks (WSNs), in which each sensor receives a single snapshot of the field. The observation or sensing mode is only partially known at the corresponding nodes, perhaps, due to their limited sensing capabilities or other unpredictable physical factors. Specifically, it is assumed that the observation process at a node switches stochastically between two modes, with mode one corresponding to the desired signal plus noise observation mode (a valid observation), and mode two corresponding to pure noise with no signal information (an invalid observation). With no prior information on the local sensing modes (valid or invalid), the paper introduces a learning-based distributed estimation procedure, the mixed detection-estimation (MDE) algorithm, based on closed-loop interactions between the iterative distributed mode learning and estimation. The online learning (or sensing mode detection) step re-assesses the validity of the local observations at each iteration, thus refining the ongoing estimation update process. The convergence of the MDE algorithm is established analytically. Simulation studies show that, in the high signal-to-noise ratio (SNR) regime, the MDE estimation error converges to that of an ideal (centralized) estimator with perfect information about the node sensing modes. This is in contrast with the estimation performance of a naive average consensus based distributed estimator (with no mode learning), whose estimation error blows up with an increasing SNR.

This paper presents the design of a robust beamforming scheme for a two-way relay network, composed of one multi-antenna relay and two single-antenna terminals, with the consideration of channel estimation errors. Given the assumption that the channel estimation error is within a certain range, we aim to minimize the transmit power at the multi-antenna relay and guarantee that the signal to interference and noise ratios (SINRs) at the two terminals are larger than a predefined value. Such a robust beamforming matrix design problem is formulated as a non-convex optimization problem, which is then converted into a semi-definite programming (SDP) problem by the S-procedure and rank one relaxation. The robust beamforming matrix is then derived from a principle eigenvector based rank-one reconstruction algorithm. We further propose a hybrid approach based on the best-effort principle to improve the outage probability performance, which is defined as the probability that one of two resulting terminal SINRs is less than the predefined value. Simulation results are presented to show that the robust design leads to better outage performance than the traditional non-robust approaches.

This paper considers the use of energy harvesters, instead of conventional time-invariant energy sources, in wireless cooperative communication. For the purpose of exposition, we study the classic three-node Gaussian relay channel with decode-and-forward (DF) relaying, in which the source and relay nodes transmit with power drawn from energy-harvesting sources. Assuming a deterministic energy-harvesting model under which the energy arrival time and the harvested energy amount are known prior to transmission, the throughput maximization problem over a finite horizon of N transmission blocks is investigated for the delay-constrained (DC) case (for which only one-block decoding delay is allowed at the destination). By exploiting the structures of the optimal source and relay power profiles, we show that the joint source and relay power allocation over time is necessary to achieve the maximum throughput, and propose an efficient forward two-dimensional search algorithm to compute the optimal power profiles.

This paper considers the use of energy harvesters in delay-constrained point-to-point wireless communications, where the source transmits with power drawn periodically from a device that harvests energy from the environment. In particular, the source is assumed to transmit over a block-fading channel with a constant transmission rate. It is also assumed that the channel state information (CSI) is unknown at the source but perfectly known at the destination, and the energy harvesting process is deterministic and known a priori at the source. The optimal power allocation is studied to minimize the receiver outage probability over a finite horizon of N energy-harvesting periods, each of which contains M communication blocks with independent channel fading coefficients. Although the outage minimization problem is shown to be non-convex, the optimal power allocation solution is obtained by the proposed forward search algorithm, which corresponds to an on-off transmission scheme. Moreover, a threshold-based sub-optimal low-complexity power allocation algorithm is proposed, which is shown to be asymptotically optimal as M goes to infinity. Finally, numerical results are provided to validate our analysis.

Consider a wireless network that has two tiers with different priorities: a primary tier vs. a secondary tier, which is an emerging network scenario with the advancement of cognitive radio technologies. The primary tier consists of randomly distributed legacy nodes of density n, which have an absolute priority to access the spectrum. The secondary tier consists of randomly distributed cognitive nodes of density m=nβwith β≥ 2, which can only access the spectrum opportunistically to limit the interference to the primary tier. Based on the assumption that the secondary tier is allowed to route the packets for the primary tier, we investigate the throughput and delay scaling laws of the two tiers in the following two scenarios: (i) the primary and secondary nodes are all static; (ii) the primary nodes are static while the secondary nodes are mobile. With the proposed protocols for the two tiers, we show that the primary tier can achieve a per-node throughput scaling of λp(n)=Θ(1/log n) in the above two scenarios. In the associated delay analysis for the first scenario, we show that the primary tier can achieve a delay scaling of Dp(n)=Θ(√(nβlog n λp(n))) with λp(n)=O(1/log n). In the second scenario, with two mobility models considered for the secondary nodes: an i.i.d. mobility model and a random walk model, we show that the primary tier can achieve delay scaling laws of Θ(1) and Θ(1/S), respectively, where S is the random walk step size. The throughput and delay scaling laws for the secondary tier are also established, which are the same as those for a stand-alone network.

In this correspondence, we consider a half-duplex large relay network, consisting of one source-destination pair and N relay nodes, each of which is connected with a subset of the other relays via signal-to-noise ratio (SNR)-limited out-of-band conferencing links. The asymptotic achievable rates of two basic relaying schemes with the "p-portion" conferencing strategy are studied: For the decode-and-forward (DF) scheme, we prove that the DF rate scales as \mathcal{O} ( log (N) ); for the amplify-and-forward (AF) scheme, we prove that it asymptotically achieves the capacity upper bound in some interesting scenarios as N goes to infinity.

We consider the half-duplex diamond relay channel, which consists of one source-destination pair and two relay nodes connected with two-way rate-limited out-of-band conferencing links. Three basic coding schemes are studied: For the decode-and-forward (DF) scheme, we obtain an achievable rate by letting the source send a common message and two private messages; for the compress-and-forward (CF) scheme, we exploit the conferencing links to help with the compression of the received signals, or to exchange messages intended for the second hop to introduce different levels of cooperations; for the amplify-and-forward (AF) scheme, we study the optimal combining strategy between the received signals from the source and the conferencing link. Moreover, we show that these schemes could achieve the capacity upper bound under certain conditions. Finally, we evaluate various achievable rates for the Gaussian case with numerical results.

In this paper, we study a practical coding scheme for the dying binary erasure channel (DBEC), which is a binary erasure channel (BEC) subject to a random fatal failure. We consider the rateless codes and optimize the degree distribution to maximize the average recovery probability. In particular, we first study the upper bound of the average recovery probability, based on which we define the objective function as the gap between the upper bound and the average recovery probability achieved by a particular degree distribution. We then seek the optimal degree distribution by minimizing the objective function. A simple and heuristic approach is also proposed to provide a suboptimal but good degree distribution. Simulation results are presented to show the significant performance gain over the conventional LT codes.

In this paper, we consider a two-hop interference network, which consists of two source-destination pairs and two relay nodes connected with signal-to-noise ratio (SNR) limited out-of-band conferencing links. Assuming that the amplify-and-forward (AF) relaying scheme is adopted, this network is shown to be equivalent to a two-user interference channel (IC). By deploying two IC decoding schemes, i.e., single-user decoding and joint decoding, respectively, we characterize the achievable rate regions with a two-stage iterative optimization method. The associated convergence issue is also studied. Furthermore, we compare the rates in the high SNR regime. Finally, simulation results show that relay conferencing can significantly improve the system performance under certain channel conditions.

Algorithms are studied for distributed least-squares (DLS) estimation of a scalar target signal in sensor networks. Due to the observation locality and the limited sensing ability, the individual sensor estimates are far from being reliable. To obtain a more reliable estimate of the target signal, the sensors could collaborate by iteratively exchanging messages with their neighbors, to refine their local estimates over time. Such an iterative DLS algorithm is investigated in this paper with and without the consideration of node failures. In particular, without sensor node failures it is shown that every instantiation of the DLS algorithm converges, i.e., consensus is reached among the sensors, with the limiting agreement value being the centralized least-squares estimate. With node failures during the iterative exchange process, the convergence of the DLS algorithm is still guaranteed; however, an error exists between the limiting agreement value and the centralized least-squares estimate. In order to reduce this error, a modified DLS scheme, the M-DLS, is provided. The M-DLS algorithm involves an additional weight compensation step, in which a sensor performs a one-time weight compensation procedure whenever it detects the failure of a neighbor. Through analytical arguments and simulations, it is shown that the M-DLS algorithm leads to a smaller error than the DLS algorithm, where the magnitude of the improvement dependents on the network topology.

The paper studies the convergence properties of the estimation error processes in distributed Kalman filtering for potentially unstable linear dynamical systems. In particular, it is shown that, in a weakly connected communication network, there exist (randomized) gossip based information dissemination schemes leading to a stochastically bounded estimation error at each sensor for any non-zero rate γ̄ of inter-sensor communication (the rate γ̄ is defined to be the average number of inter-sensor communications per signal evolution epoch). A gossip-based information exchange protocol, the M-GIKF, is presented, in which sensors exchange estimates and aggregate observations at a rate γ̄ > 0, leading to desired convergence properties. Under the assumption of global (centralized) detectability of the signal/observation model (necessary for a centralized estimator having access to all sensor observations at all times to yield bounded estimation error), it is shown that the distributed M-GIKF leads to a stochastically bounded estimation error at each sensor. The conditional estimation error covariance sequence at each sensor is shown to evolve as a random Riccati equation (RRE) with Markov modulated switching. The RRE is analyzed through a random dynamical system (RDS) formulation, and the asymptotic estimation error at each sensor is characterized in terms of an associated invariant measure µγ̄ of the RDS.

We study a special class of the cognitive radio channel in which the receiver of the cognitive pair does not suffer interference from the primary user. Previously developed general encoding schemes for this channel are complex as they attempt to cope with arbitrary channel conditions, which leads to rate regions that are difficult to evaluate. The focus of our work is to derive simple rate regions that are easily computable, thereby providing more insights into achievable rates and good coding strategies under different channel conditions. We first present several explicit achievable regions for the general discrete memoryless case. We also present an improved outer bound on the capacity region for the case of high interference. We then extend these regions to Gaussian channels. With a simple outer bound we establish a new capacity region in the high-interference regime. Lastly, we provide numerical comparisons between the derived achievable rate regions and the outer bounds.

In this paper, we study the state-dependent two-user interference channel, where the state information is non-causally known at both transmitters but unknown to either of the receivers. We propose two coding schemes for the discrete memoryless case: simultaneous encoding for the sub-messages in the first one and superposition encoding in the second one, both with rate splitting and Gel'fand-Pinsker coding. The corresponding achievable rate regions are established.

This paper considers distributed detection over a noisy network, in which each connected sensor pair can communicate over an additive noise channel. With non-identically distributed generic sensor observations, a mixed time scale recursive algorithm for binary hypothesis testing over such networks is proposed. Under some mild assumptions on network connectivity and global detectability (the positivity of the global or centralized Kullback-Liebler divergence), this algorithm yields asymptotically zero probabilities of Type-I and Type-II errors (henceforth referred to as probabilities of error). When sensor observations are identically distributed, a simplified single time scale version of the proposed algorithm is shown to achieve asymptotically zero probabilities of error. Convergence rate guarantees in terms of asymptotic normality of certain scaled decision variables are provided for this simplified procedure. As an example, a practical Gaussian sensor network is considered, for which the error decay exponents are explicitly characterized in terms of the network and noise parameters.

In this paper, we study the two-user state-dependent Gaussian interference channel, where the Gaussian distributed state information is non-causally known at both transmitters, but unknown to either of the receivers. We focus on the simultaneous encoding scheme and propose an active interference cancelation mechanism, which is a generalized dirty-paper coding technique, to partially eliminate the state effect at the receivers. The corresponding achievable rate region is derived and the strong interference case is studied, where we propose several heuristic schemes and provide numerical results to show that active interference cancelation significantly enlarges the achievable rate region.

To understand the network response to large-scale physical attacks, we investigate the asymptotic capacity of a half-duplex fading relay network with random node failures when the number of relays N gets infinitely large. In this paper, a simplified independent attack model is assumed where each relay node fails with a certain probability. The noncoherent relaying scheme is considered, which corresponds to the case of zero forward-link channel state information (CSI) at the relays. Accordingly, the whole relay network can be shown equivalent to a Rayleigh fading channel, where we derive the ε-outage capacity upper bound according to the multiple access (MAC) cut-set, and the ε-outage achievable rates for both the amplify-and-forward (AF) and decode-and-forward (DF) strategies. Furthermore, we show that the DF strategy is asymptotically optimal as the outage probability ε goes to zero, with the AF strategy strictly suboptimal over all signal to noise ratio (SNR) regimes. Regarding the rate loss due to random attacks, the AF strategy suffers a less portion of rate loss than the DF strategy in the high SNR regime, while the DF strategy demonstrates more robust performance in the low SNR regime.

In this paper, we consider the full-duplex and half-duplex Gaussian relay channels where the noises at the relay and destination are arbitrarily correlated. We first derive the capacity upper bound and the achievable rates with three existing schemes: Decode-and-Forward (DF), Compress-and-Forward (CF), and Amplify-and-Forward (AF). We present two capacity results under specific noise correlation coefficients, one being achieved by DF and the other being achieved by direct link transmission (or a special case of CF). The channel for the former capacity result is equivalent to the traditional Gaussian degraded relay channel and the latter corresponds to the Gaussian reversely-degraded relay channel. For CF and AF schemes, we show that their achievable rates are strictly decreasing functions of the correlation coefficient when the correlation coefficient is negative. Moreover, when the noise correlation coefficient is positive, the CF achievable rate may also outperform the independent-noise case if the noise correlation coefficient is within a certain range. Through numerical comparisons under different channel settings, we observe that although DF completely disregards the noise correlation while the other two can potentially exploit such extra information, none of the three relay schemes always outperforms the others over different correlation coefficients. Moreover, the exploitation of noise correlation by CF and AF accrues more benefit when the source-relay link is weak. This paper also considers the optimal power allocation problem under the correlated-noise channel setting. With individual power constraints at the relay and the source, it is shown that the relay should use all its available power to maximize the achievable rates under any correlation coefficient. With a total power constraint across the source and the relay, the achievable rates are proved to be concave functions over the power allocation factor for AF and CF under full-duplex mode, where the closed-form power allocation strategy is derived.

This paper studies the downlink beamforming for a multi-cell system, where multiple base stations (BSs) each with multiple antennas cooperatively design their respective transmit beamforming vectors to optimize the overall system performance. It is assumed that all mobile stations (MSs) are equipped with a single antenna each, and there is one active MS in each cell at one time. Accordingly, the system of interest can be modeled by a multiple-input single-output (MISO) Gaussian interference channel (IC), termed as MISO-IC, with interference treated as additive Gaussian noise. We are interested in designing a multi-cell cooperative downlink beamforming scheme to achieve different rate-tuples for active MSs on the Pareto boundary of the achievable rate region for the MISO-IC, which is in general a non-convex problem due to the coupled signal structure. By exploring the relationship between the MISO-IC and the cognitive radio (CR) MISO channel, we show that each Pareto-boundary rate-tuple of the MISO-IC can be achieved in a decentralized manner when each of the MSs attains its own channel capacity subject to a certain set of interference-power constraints (also known as interference-temperature constraints in the CR system) at the other MS receivers. Furthermore, we show that this result leads to a decentralized algorithm for implementing the multi-cell cooperative downlink beamforming, where all different pairs of BSs independently search for their mutually desirable interference-temperature constraints, under which their respective beamforming vectors are optimized to maximize the individual transmit rates. It is shown that this algorithm guarantees to improve the rates for a given pair of BSs at each iteration with those for the other BSs unaffected, and converges when there are no further incentives for all the BSs to adjust their mutual interference-temperature constraints.

In this paper, we consider a half-duplex Gaussian relay channel where the noises at the relay and destination are arbitrarily correlated. For this generalized relay channel, we first evaluate the cut-set bound as well as the achievable rates with three existing relay schemes: Decode-and-Forward (DF), Compress-and-Forward (CF), and Amplify-and-Forward (AF), with performance comparison under various channel settings. We observe that although DF completely disregards the noise correlation while the other two could exploit such extra information, none of the three relay schemes always outperforms the others over different correlation coefficients; however, the exploitation of noise correlation by CF and AF leads to more significant benefit when the source-relay channel is weak. It is further shown that a negative noise correlation is always helpful for AF. We also establish two capacity-achieving results under two special noise correlation coefficients, with one being achieved by DF and the other being achieved by direct link transmission (or a special case of CF), which correspond to the capacity results for the traditional Gaussian degraded relay channel and the Gaussian reversely-degraded one.

This paper studies the achievable rates of the multi-antenna or multiple-input multiple-output (MIMO) secrecy channel with multiple single-/multi-antenna eavesdroppers. By assuming Gaussian input, the maximum achievable secrecy rate is obtained with the optimal transmit covariance matrix that maximizes the minimum difference between the channel mutual information of the secrecy user and those of the eavesdroppers. The maximum secrecy rate computation can thus be formulated as a non-convex max-min problem, which cannot be solved efficiently by existing methods. To handle this difficulty, this paper explores a new relationship between the secrecy channel and the recently developed cognitive radio (CR) channel, in which the secondary user transmits over the same spectrum simultaneously with multiple primary users, subject to the received interference power constraints at the primary users, or the so-called "interference temperature (IT)" constraints. By constructing an auxiliary multi-antenna CR channel that has the same channel responses as the secrecy channel, this paper shows that the optimal transmit covariance to achieve the maximum secrecy rate is the same as that to achieve the CR spectrum sharing capacity with properly selected IT constraints. Thereby, finding the optimal complex transmit covariance matrix for the secrecy channel becomes equivalent to searching over a set of real IT constraints in the auxiliary CR channel. Based on this relationship, efficient algorithms are proposed to solve the non-convex secrecy rate maximization problem by transforming it into a sequence of convex CR spectrum sharing capacity computation problems, under various setups of the secrecy channel.

We study the scaling laws for the throughputs and delays of two coexisting wireless networks that operate in the same geographic region. The primary network consists of Poisson distributed legacy users of densityn, and the secondary network consists of Poisson distributed cognitive users of densitym, withm>n. The primary users have a higher priority to access the spectrum without particular considerations for the secondary users, while the secondary users have to act conservatively in order to limit the interference to the primary users. With a practical assumption that the secondary users only know the locations of the primary transmitters (not the primary receivers), we first show that both networks can achieve the same throughput scaling law as what Gupta and Kumar (IEEE Trans. Inf. Theory,vol. 46, no. 2, pp. 388-404, Mar. 2000) established for a standalone wireless network if proper transmission schemes are deployed, where a certain throughput is achievable for each individual secondary user (i.e., zero outage) with high probability. By using a fluid model, we also show that both networks can achieve the same delay-throughput tradeoff as the optimal one established by El Gamal (IEEE Trans. Inf. Theory, vol. 52, no. 6, pp. 2568-2592, Jun. 2006) for a standalone wireless network.

Consider a wireless network of two tiers with different priorities: a primary tier and a secondary tier, which is an emerging network scenario with the advancement of cognitive radio technologies. The primary tier is constructed over static nodes of density n, which are randomly distributed and have an absolute priority to access the spectrum. The secondary tier contains mobile nodes of density m = nßwith ß ¿ 2, which can only access the spectrum opportunistically to limit the interference to the primary tier. By allowing the secondary tier to relay the packets for the primary tier, we show that the achievable per-node throughput scaling for the primary tier can be improved to ¿p(n) = ¿(1/log n). In the associated delay analysis, two mobility models are considered for the secondary nodes: an i.i.d. mobility model and a random walk model. We show that the primary tier can achieve delay scaling laws of ¿(1) and ¿(1/S) with the two mobility models, respectively, where S is the random walk step size. Furthermore, we show that the primary tier can achieve a delay-throughput tradeoff of Dp(n) = O (n¿p(n)) with ¿p(n) = O(1/log n) for the random walk model. The throughput and delay scaling laws for the secondary tier are also established, which are the same as those for a stand-alone mobile network.

In this paper, we investigate the asymptotic e-outage capacity of a half-duplex large fading relay network, which consists of one source node, one destination node, and N relay nodes. The relay nodes are assumed to be randomly deployed in a given area and under fatal independent random attacks with probability p. With a total power constraint on all the nodes, we examine the e-outage rate of the amplify-and-forward (AF) strategy when N tends to infinity, assuming no channel state information at the relays. We further quantify the gap between the e-outage rate and the e-outage cut-set bound, which is determined by the attack probability p, the source vs. sum power allocation factor ¿, and the topology of the networks. Moreover, we examine the effect of random attacks on the eoutage rate, and calculate the relative losses in low and high SNR regimes, respectively. Finally, for general SNR, we show that it is a quasiconcave problem to determine the optimal power allocation between the source and the relays, and we could obtain the optimal ¿ efficiently.

As an enabling functionality of overlay cognitive radio networks, spectrum sensing needs to reliably detect licensed signal in the band of interest. To achieve reliable sensing, we propose a linear fusion scheme for distributed spectrum sensing to combine the sensing results from multiple spatially distributed cognitive radios. The optimal linear fusion design is formulated into a nonconvex optimization problem. We show that the optimal solution of such a nonconvex problem can be solved via semi-definite programming reformulation.

We study the transmission capacities of two coexisting wireless networks (a primary network vs. a secondary network) that operate in the same geographic region and share the same spectrum. We define transmission capacity as the product among the density of transmissions, the transmission rate, and the successful transmission probability (1 minus the outage probability). The primary (PR) network has a higher priority to access the spectrum without particular considerations for the secondary (SR) network, where the SR network limits its interference to the PR network by carefully controlling the density of its transmitters. Assuming that the nodes are distributed according to Poisson point processes and the two networks use different transmission ranges, we quantify the transmission capacities for both of these two networks and discuss their tradeoff based on asymptotic analysis. Our results show that if the PR network permits a small increase of its outage probability, the sum transmission capacity of the two networks (i.e., the overall spectrum efficiency per unit area) will be boosted significantly over that of a single network.

In this paper, we investigate the optimal transmission schemes for dying channels, which were introduced in (M. Zeng et al., 2008). The dying channels are resulted in wireless networks subject to random fatal impacts, e.g., sensor networks under sudden physical attacks or cognitive radio networks with unpredictable primary user occupancy. Due to the non-ergodic and delay- limited nature of a dying channel, the outage capacity is adopted as the performance metric. Firstly, we show that the optimal power allocation profile is non-increasing when fading gains are independently and identically distributed (i.i.d.). Secondly, when the fading gains over the blocks are the same, we prove that the optimal number of blocks over which a codeword should be spanned is K = 1. At last, we consider the case where uniform power allocation is utilized and fading gains are i.i.d. In this case, we derive the upper and lower bounds for the outage probability. Moreover, for the high signal-to-noise ratio (SNR) case with Rayleigh fading , we derive analytical results on the optimal number of coding blocks K. For the low SNR case, we show that repetition transmissions are approximately optimal.

This paper studies the wirelesstwo-wayrelaychannel(TWRC), where two source nodes, S1 and S2, exchange information through an assisting relay node, R. It is assumed that R receives the sum signal from S1 and S2 in one time-slot, and then amplifies and forwards the received signal to both S1 and S2 in the next time-slot. By applying the principle ofanaloguenetworkcoding(ANC), each of S1 and S2 cancels the so-called "self-interference" in the received signal from R and then decodes the desired message. Assuming that S1 and S2 are each equipped with a single antenna and R with multi-antennas, this paper analyzes thecapacityregionof an ANC-based TWRC with linear processing (beamforming) at R. The capacity region contains all the achievable bidirectional rate-pairs of S1 and S2 under the given transmit power constraints at S1, S2, and R. We present the optimal relay beamforming structure as well as an efficient algorithm to compute the optimal beamforming matrix based on convex optimization techniques.

Consider a static wireless network that has two tiers with different priorities: a primary tier vs. a secondary tier. The primary tier consists of randomly distributed legacy nodes of density n, which have an absolute priority to access the spectrum. The secondary tier consists of randomly distributed cognitive nodes of density m = nbetawith beta ges 2, which can only access the spectrum opportunistically to limit the interference to the primary tier. By allowing the secondary tier to route the packets for the primary tier, we show that the primary tier can achieve a throughput scaling of lambdap(n) = Theta(1/log n) per node and a delay-throughput tradeoff of Dp(n) = Theta (radic(nbetalog nlambdap(n))) for lambdap(n) = O (1/log n), while the secondary tier still achieves the same optimal delay-throughput tradeoff as a stand-alone network.

We investigate the Gaussian relay channel where the additive noises at the relay and destination are correlated. We obtain achievable rates for the compress-and-forward and decode-and-forward relaying strategies, and compare them to each other and to the capacity upper bound.We show that neither scheme is uniformly best over all channel gains and correlation coefficients. We also derive specific relationships between the channel gains and noise correlations for which one of these schemes is capacity-achieving, thereby increasing the class of relay channels for which capacity is known.

We study the transmission capacities of two coexisting wireless networks (a primary network vs. a secondary network) that operate in the same geographic region and share the same spectrum. The primary (PR) network has a higher priority to access the spectrum without particular considerations for the secondary (SR) network, where the SR network limits its interference to the PR network by carefully controlling its node density. Considering general power-law wireless channels with path-loss exponent α ≫ 2 and small-scale Rayleigh fading, based on the stochastic geometry theory, we derive the transmission capacities for both of the two networks and quantify their tradeoff via asymptotic analysis. Our results show that if the PR network permits a small increase of its outage probability, the sum transmission capacity of the two networks (i.e., the overall spectrum efficiency per unit area) will be boosted significantly over that of a single network, which generalizes our previous result in [1] over a special case of deterministic power-law channel with α = 4.

We study the delay-throughput tradeoffs for two coexisting wireless networks that operate in the same geographic region. The primary network consists of Poisson distributed legacy users of density n, and the secondary network consists of Poisson distributed cognitive users of density m, with m > n. The primary users have a higher priority to access the spectrum without particular considerations for the secondary users, while the secondary users have to act conservatively in order to limit the interference to the primary users. With a practical assumption that the secondary users only know the locations of the primary transmitters (not the primary receivers), based on our previous work in, we show that both networks can achieve the same delay-throughput tradeoff as the optimal one established by El Gamal et al. for a stand-alone wireless network.

This paper studies the wireless two-way relay channel (TWRC), where two source nodes, S1 and S2, exchange information through an assisting relay node, R. It is assumed that R receives the sum signal from S1 and S2 in one timeslot, and then amplifies and forwards the received signal to both S1 and S2 in the next time-slot. By applying the principle of analogue network coding (ANC), each of S1 and S2 cancels the so-called "self-interference" in the received signal from R and then decodes the desired message. Assuming that S1 and S2 are each equipped with a single antenna and R with multi-antennas, this paper analyzes the capacity region of the ANC-based TWRC with linear processing (beamforming) at R. The capacity region contains all the achievable bidirectional rate-pairs of S1 and S2 under the given transmit power constraints at S1, S2, and R. We present the optimal relay beamforming structure as well as an efficient algorithm to compute the optimal beamforming matrix based on convex optimization techniques. Low-complexity suboptimal relay beamforming schemes are also presented, and their achievable rates are compared against the capacity with the optimal scheme.

This paper studies the information-theoretic limits of a secondary or cognitive radio (CR) network under spectrum sharing with an existing primary radio network. In particular, the fading cognitive multiple-access channel (C-MAC) is first studied, where multiple secondary users transmit to the secondary base station (BS) under both individual transmit-power constraints and a set of interference-power constraints each applied at one of the primary receivers. This paper considers the long-term (LT) or the short-term (ST) transmit-power constraint over the fading states at each secondary transmitter, combined with the LT or ST interference-power constraint at each primary receiver. In each case, the optimal power allocation scheme is derived for the secondary users to achieve the ergodic sum capacity of the fading C-MAC, as well as the conditions for the optimality of the dynamic time-division multiple-access (D-TDMA) scheme in the secondary network. The fading cognitive broadcast channel (C-BC) that models the downlink transmission in the secondary network is then studied under the LT/ST transmit-power constraint at the secondary BS jointly with the LT/ST interference-power constraint at each of the primary receivers. It is shown that D-TDMA is indeed optimal for achieving the ergodic sum capacity of the fading C-BC for all combinations of transmit-power and interference-power constraints.

We study the scaling laws for the throughputs of two coexisting wireless networks that operate in the same geographic region. The primary network consists of Poisson distributed legacy users of density n, and the secondary network consists of Poisson distributed cognitive users of density m, with m > n. The primary users have a higher priority to access the spectrum without particular considerations for the secondary users, while the secondary users have to act conservatively in order to limit interference to the primary users. With a practical assumption that the secondary users only know the locations of the primary transmitters, we show that both networks can achieve the same throughput scaling law as a stand-alone wireless network if proper transmission schemes are deployed, where a finite throughput is achievable for each individual secondary user (i.e., zero outage) with high probability.

In this paper, we investigate a new type of channels named as "dying" channels, which are resulted in wireless networks subject to random fatal impacts, e.g., sensor networks under sudden physical attacks or cognitive radio networks with unpredictable primary user occupancy. Under such circumstances, it is critical to quantify how fast and reliable information can be collected over "dying" links. In this paper, we focus on a simple point-to-point communication setting and model the "dying" channel by the traditional if-block block-fading (BF) model subject to a fatal attack that may happen randomly in any of the K blocks. The resultant channel is non-ergodic and delay-limited in nature, and thus its information-theoretic limit can be measured by adopting the conventional outage capacity concept. An outage event in a "dying" channel could be caused by two mechanisms: fading over finite K blocks; and random attack within K blocks. In this paper, we present the general problem formulation to determine the outage capacity of the "dying" channel, as well as the corresponding optimal transmit power allocation over the K blocks, based upon the known probability distributions of both the fading channel and the attack time. The optimal power allocation turns out to be in general nonuniform with a decreasing profile over transmission blocks. In addition, there exists an optimal number of blocks over which the codeword is spanned. Some other interesting observations are also made pertinent to the optimal transmission over a "dying" channel.

Spectrum sensing is an essential functionality that enables cognitive radios to detect spectral holes and opportunistically use under-utilized frequency bands without causing harmful interference to primary networks. Since individual cognitive radios might not be able to reliably detect weak primary signals due to channel fading/shadowing, this paper proposes a cooperative wideband spectrum sensing scheme, referred to as spatial-spectral joint detection, which is based on a linear combination of the local statistics from spatially distributed multiple cognitive radios. The cooperative sensing problem is formulated into an optimization problem, for which suboptimal but efficient solutions can be obtained through mathematical transformation under practical conditions.

In this paper, the power and rate control schemes for a single cognitive radio (CR) link are considered, which operate over multiple frequency bands in the presence of licensed primary radios (PRs). Specifically, with considerations of delay in spectrum sensing, an optimal algorithm based on dynamic programming (DP) is proposed to maximize the average rate of the CR link over a finite time horizon under the constraints on the total energy budget and the CR-to-PR interference. In the proposed algorithm, the PR occupancy of the channels is modeled as a discrete-time Markov chain (DTMC). Based on such a model, the optimal power and rate control strategy is derived, which is a function of the delayed spectrum sensing output and the instantaneous battery level. Simulation results show that the proposed algorithm leads to significant performance improvement over heuristic algorithms.

Spectrum sensing is an essential enabling functionality for cognitive radio networks to detect spectrum holes and opportunistically use the under-utilized frequency bands without causing harmful interference to legacy networks. This paper introduces a novel wideband spectrum sensing technique, called multiband joint detection, which jointly detects the signal energy levels over multiple frequency bands rather than consider one band at a time. The proposed strategy is efficient in improving the dynamic spectrum utilization and reducing interference to the primary users. The spectrum sensing problem is formulated as a class of optimization problems in interference limited cognitive radio networks. By exploiting the hidden convexity in the seemingly non-convex problem formulations, optimal solutions for multiband joint detection are obtained under practical conditions. Simulation results show that the proposed spectrum sensing schemes can considerably improve the system performance. This paper establishes important principles for the design of wideband spectrum sensing algorithms in cognitive radio networks.

In this paper, we present a novel joint power/channel allocation scheme that uses a distributed pricing strategy to improve the network's performance. According to this scheme, the spectrum allocation problem is modeled as a non-cooperative game. A price-based iterative water-filling (PIWF) algorithm is proposed, which allows users to converge to the Nash Equilibrium (NE). This PIWF algorithm can be implemented distributively, with CRs repeatedly negotiating their best transmission powers and spectrum. We propose a protocol that implements our price- based resource allocation algorithm. The proposed MAC protocol allows multiple CR pairs to first contend through an admission phase, and then to iteratively negotiate their transmission powers and spectrum via control-packet exchanges. Subsequently, CRs proceed concurrently with their data transmissions. Simulations are used to study the performance of our protocol and demonstrate its effectiveness in improving the overall network throughput and reducing the average transmission power.

Subcarrier, power, and rate allocation schemes for the orthogonal frequency division multiple access (OFDMA) uplink are considered. A low-complexity algorithm with fairness consideration is proposed to maximize the sum rate under individual rate and transmit power constraints. The complexity reduction and fairness are achieved by performing subcarrier allocation in two steps: Initial subcarrier allocation that considers per-user fairness; and residual subcarrier allocation that further increases the sum rate. The proposed algorithm achieves superior outage performance and has linear complexity in the number of users and subcarriers, while incurring negligible sum-rate degradation when the average path-loss is the same among users.

Cooperative communication with single relay selection is a simple but effective communication scheme for energy-constrained networks. In this paper, we propose a novel selective single-relay cooperative scheme, combining selective-relay cooperative communication with physical-layer power control. Based on the MAC-layer RTS-CTS signaling, a set of potential relays compute individually the required transmission power to participate in the cooperative communication, and compete within a window of fixed length. The "best" relay is selected in a distributed fashion with minimum signaling overhead. We derive power-control solutions corresponding to two policies on relay selection: one is to minimize the energy consumption per data packet, and the other is to maximize the network lifetime. Our numerical and simulation results confirm that the proposed scheme achieves significant energy savings and prolongs the network lifetime considerably.

Power and rate control schemes for a single cognitive radio (CR) channel are considered in the presence of licensed primary radios (PRs). A dynamic programming (DP) based algorithm is proposed to maximize the long-term average rate for the CR link under constraints on the total energy budget and the CR-to-PR disturbance. In the proposed algorithm, the behavior of PRs is modeled as a two-state Markov chain. Based on such a model, the optimal power and rate control strategy for each time slot is derived, which is a function of the energy level at the beginning of current time slot and the previous behavior of PRs. Simulation results show that the proposed algorithm can lead to a significant performance improvement in term of the long-term average rate while keeping the probability of CR-to- PR disturbance below a given level.

Spectrum sensing is a key enabling functionality in cognitive radio (CR) networks, where the CRs act as secondary users that opportunistically access free frequency bands. Due to the effects of channel fading, individual CRs may not be able to reliably detect the existence of a primary radio, who is a licensed user for the particular band. In this paper, we present optimal cooperation strategies for spectrum sensing to combat the effects of destructive channels and malfunctioning devices. Our approach conducts spectrum sensing based on the linear combination of local test statistics from individual secondary users. We propose two optimization schemes to control the combining weights, and compare their performance. Our first approach is to optimize the probability distribution function of the global test statistics at the fusion center. For the second scheme, we maximize the global detection sensitivity under constraints on the false alarm probability. Simulation results illustrate the significant cooperative gain achieved by the proposed strategies.

We develop a q-expectation maximization (q-EM) simulated annealing method for parameter estimation. The g-EM algorithm is a one-parameter generalization of normal expectation maximization (EM) algorithm based on Tsallis entropy. By incorporating the simulated annealing method, we propose the q-deterministic annealing expectation maximization (q-DAEM) algorithm. Given the inherent connection between a physical annealing process and statistical mechanics, we show that the proposed algorithm actually minimizes a counterpart of the free energy in statistical mechanics by controlling an effective temperature. Simulations of mixed Gaussian parameter estimation show that the proposed method is much less initialization-dependent than the standard EM algorithm and converges dramatically faster than the DAEM algorithm.

The distributed detection problem is considered with wireless sensors sending local decisions to a fusion center over slow-fading orthogonal multiple access channels. The J-divergence between the distributions under different hypotheses is used as the performance criterion. Given the slow-fading channels between the sensors and the fusion center, two new concepts called detection outage and detection diversity are introduced to quantify the long-term system performance, where the detection outage probability is defined as the probability that the instantaneous J-divergence is smaller than a certain threshold. Based on the detection outage probability, detection diversity order is defined as the slope of the outage probability curve (versus system power consumption) when things are plotted in logarithm domain. It is shown that even with a simple uniform transmission strategy, full detection diversity can be achieved on the order of K with K the total number of sensors in the network. We further show that if the transmission power is optimized across sensors, adaptive power gain can be achieved in addition to full diversity gain.

The general joint design of the physical, MAC, and routing layers to minimize network energy consumption is complex and hard to solve. Heuristics to compute approximate solutions and high-complexity algorithms to compute exact solutions have been previously proposed. In this paper, we focus on synchronous small-scale networks with interference-free link scheduling and practical MQAM link transmission schemes. We show that the cross-layer optimization problems can be closely approximated by convex optimization problems that can be efficiently solved. There are two main contributions of this paper. First of all, we minimize the total network energy that includes both transmission and circuit energy consumptions, where we explore the tradeoff between the two energy elements. Specifically, we use interference-free TDMA as the medium access control scheme. We optimize the routing flow, TDMA slot assignment, and MQAM modulation rate and power on each link. The results demonstrate that the minimum energy transmission scheme is a combination of multihop and single-hop transmissions for general networks; including circuit energy favors transmission schemes with fewer hops. Secondly, based on the solved optimal transmission scheme, we quantify the best trade-off curve between delay and energy consumption, where we derive a scheduling algorithm to minimize the worst-case packet delay.

Energy-constrained multihop wireless links are considered, where the total power consumption is minimized under given requirements on the end-to-end bit error rate (BER). As multihop transmissions are known to be able to save transmission energy in a wireless environment, we study the optimal power scheduling schemes over intermediate hops when the source- relays-destination link can be modeled as cascaded binary symmetric channels. The problem is formulated with an end- to-end BER constraint, and the resulting power consumption is compared with that of the individual link requirement strategy where each hop assigns power under a per-link BER constraint. Results show that the proposed joint power scheduling strategy can achieve a maximum power reduction factor of M in an M-hop route.

Adaptive subcarrier, bit, and power allocation schemes for orthogonal frequency division multiple access (OFDMA) systems are considered. A low-complexity algorithm is proposed to minimize the overall transmit power under certain fairness constraints: Data rate and bit-error-rate (BER) requirements for each user. In the proposed algorithm, subcarrier allocation is first executed, which follows two steps: An initial subcarrier allocation to satisfy the individual data rate requirement; then a residual subcarrier allocation to further reduce the overall transmit power. After that, a single-user optimal bit and power allocation is performed for each user based on the water-filling strategy. The proposed heuristic subcarrier allocation algorithm has linear complexity in the number of users and subcarriers, as opposed to the combinatorial nature of the optimal solutions. More importantly, the proposed resource allocation scheme explores frequency diversity in a more efficient way compared with reference schemes of similar complexity, which makes it more preferable in highly frequency-selective fading channels.

A multi-channel parallel transmission protocol is proposed for the medium access control in cognitive radio networks (CRNs). This protocol contains two key elements: multi-channel assignment and multi-channel contention. For an incoming flow-based connection request, the minimum number of parallel channels are assigned to satisfy the rate and interference mask constraints. For the contention of the assigned channels, our protocol provides an extension of the single-channel RTS- CTS-DATA-ACK handshaking of the IEEE 802.11 scheme. The proposed MAC coherently integrates optimization results into a practical implementation. Through numerical examples, we verify that our protocol provides lower connection blocking probability and higher system throughput for CRNs than its single-channel counterpart.

We consider the distributed estimation of an unknown vector signal in a bandwidth constrained sensor network with a fusion center (FC). Due to power and bandwidth limitations, each sensor compresses its data in order to minimize the amount of information that needs to be communicated to the FC. In this context, we design a linear decentralized estimation scheme (DES), where each sensor linearly encodes its observations before the transmission to the FC, which performs a minimum mean squared error (MMSE) estimation for the unknown vector signal based on the received messages. When the channels between sensors and the FC are orthogonal, it has been shown previously that the complexity of designing the optimal encoding matrices is NP-hard in general. In this paper, we study the optimal design of linear DES for the case of non- orthogonal multiple access channel (MAC) under both bandwidth and power constraints. We show that when the MAC between sensors and the FC is noiseless, the resulting problem has a closed-form solution, while in the noisy MAC case, the problem can be efficiently solved by semi-definite programming (SDP).

We investigate distributed estimation based on measurements from multiple wireless sensors. For the same target, different sensors have different observations, which are modeled by additive observation noises of different variances. The observations are transmitted using (analog) amplify-and-forward transmissions from the sensors over non-ideal wireless channels to a fusion center, where they are combined to generate an estimate of the observed target. Our goal is to minimize total end-to-end distortion under certain power constraints, assuming the Best Linear Unbiased Estimator (BLUE) is used. We analyze the system outage performance, and show an achievable diversity gain of order K, which is the number of sensors. We also show that by turning off bad sensors, i.e., sensors with bad channels, we achieve adaptive power gain without losing diversity gain, where the adaptive power gain is similar to the array gain achieved in Multiple Input Single Output (MISO) systems when channel conditions are known to the transmitter.

We consider energy-constrained multihop wireless links where the power consumption must be minimized under given requirements on the end-to-end bit error probability. As multihop is known to be able to save transmission energy in a wireless environment, we study the power scheduling schemes over intermediate hops where each source-relay-destination link is modeled as a cascaded binary symmetric channel. The optimal power scheduling problem is formulated with the end-to-end bit error probability constraints. Results show that with the proposed joint power scheduling strategy, a maximum power reduction factor of M is possible in a M-hop route compared with the individual link requirement strategy where each hop assigns power individually to meet certain bit error probability at each intermediate link.

Transmission power control has often been used in wireless packet networks to improve the channel reuse and reduce energy consumption. It has been mainly applied to single-input single-output (SISO) systems, where each node is equipped with a single antenna. In this paper, we propose a power-controlled channel access protocol for MIMO-capable wireless LANs with two antennas per node. Our protocol, called E-BASIC, is based on a modification of the classic CSMA/CA access scheme, where we dynamically adjust the "transmission mode" and the transmission power on a per-packet basis so as to minimize the total energy consumption. By "transmission mode" we mean one of the four possible transmit/receive antenna configurations: 1 x 1 (SISO), 2 x 1 (MISO), 1 x 2 (SIMO), and 2 x 2 (MIMO). Our energy model accounts for both the transmission and the circuit powers. While the MIMO mode requires less RF transmission energy than the other three modes, it also incurs the highest circuit energy consumption. Depending on the transmitter-receiver distance, any of the four modes can be the "optimal" one in terms of minimizing the total energy consumption. We study the performance of E-BASIC in ad hoc and infrastructure-based scenarios and compare it with two channel access protocols (802.11 and BASIC) that use a fixed transmission mode (SISO or MIMO) all the time. Our simulations show that E-BASIC often consumes much less total energy than the reference protocols. We further incorporate E-BASIC into the design of a power-aware routing (PAR) scheme that uses a variant of Dijkstra's algorithm to select the most energy-efficient end-to-end path between nodes.

We consider the joint optimal design of the physical, medium access control (MAC), and routing layers to maximize the lifetime of energy-constrained wireless sensor networks. The problem of computing lifetime-optimal routing flow, link schedule, and link transmission powers for all active time slots is formulated as a non-linear optimization problem. We first restrict the link schedules to the class of interference-free time division multiple access (TDMA) schedules. In this special case, we formulate the optimization problem as a mixed integerconvex program, which can be solved using standard techniques. Moreover, when the slots lengths are variable, the optimization problem is convex and can be solved efficiently and exactly using interior point methods. For general non-orthogonal link schedules, we propose an iterative algorithm that alternates between adaptive link scheduling and computation of optimal link rates and transmission powers for a fixed link schedule. The performance of this algorithm is compared to other design approaches for several network topologies. The results illustrate the advantages of load balancing, multihop routing, frequency reuse, and interference mitigation in increasing the lifetime of energy-constrained networks. We also briefly discuss computational approaches to extend this algorithm to large networks

We consider sensor networks where energy is a limited resource so that energy consumption must be minimized while satisfying given throughput requirements. Moreover, energy consumption must take into account both the transmission energy and the circuit processing energy for short-range communications. In this context, we analyze energy-efficient joint routing and link scheduling to achieve the optimal tradeoff between energy and delay. For networks composed of multiple clusters of nodes, we propose and analyze the cooperative multiple-input multiple-output (MIMO) approach where multiple sensor nodes in the same cluster cooperate in signal transmission and/or reception. We show that local information exchange within the cluster is not necessary for node cooperation based on Alamouti diversity codes if the transmissions are properly scheduled. We further show that the routing optimization problem based on cooperative MIMO can be solved by designing an equivalent single-input single-output (SISO) system, where each cluster is treated as a super node. For both SISO-based and MIMO-based cases, we derive the best energy-delay tradeoff curves and show that the cooperative MIMO approach dramatically improves the energy-delay performance.

Sensor networks in which energy is a limited resource so that energy consumption must be minimized for the intended application are considered. In this context, an energy-efficient method for the joint estimation of an unknown analog source under a given distortion constraint is proposed. The approach is purely analog, in which each sensor simply amplifies and forwards the noise-corrupted analog observation to the fusion center for joint estimation. The total transmission power across all the sensor nodes is minimized while satisfying a distortion requirement on the joint estimate. The energy efficiency of this analog approach is compared with previously proposed digital approaches with and without coding. It is shown in our simulation that the analog approach is more energy-efficient than the digital system without coding, and in some cases outperforms the digital system with optimal coding.

We consider sensor networks where energy is a limited resource so that energy consumption must be minimized while satisfying given throughput requirements. Moreover, energy consumption must take into account both the transmission energy and the circuit processing energy for short-range communications. We emphasize that the energy efficiency must be supported across all layers of the protocol stack through a cross-layer design. In this context, we analyze energy-efficient joint routing, scheduling, and link adaptation strategies that maximize the network lifetime. We propose variable-length TDMA schemes where the slot length is optimally assigned according to the routing requirement while minimizing the energy consumption across the network. We show that the optimization problems can be transformed into or approximated by convex problems that can be efficiently solved using known techniques. The results show that multihop routing schemes are more energy-efficient when only transmission energy is considered, but single-hop transmissions may be more efficient when the circuit processing energy is considered.

We consider a wireless sensor network where the nodes have limited energy. We first analyze the delay performance of a transmission scheme based on time division multiple access (TDMA). We propose a simple link scheduling algorithm to find the minimum-delay schedule given the slot lengths for all the links. We then combine these results with our previous work on energy-optimal cross-layer design to minimize the delay in transferring a fixed number of bits from the source nodes to the sink, in an energy-constrained manner. We also study the tradeoff between the total energy consumption and delay. Pareto optimal energy-delay curves are computed by solving a series of convex optimization problems where each objective function is a weighted sum of the delay and the total energy consumption. The computation is done for networks with and without link adaptation capabilities.

We consider the joint optimal design of physical, medium access control (MAC), and routing layers to maximize the lifetime of energy-constrained wireless sensor networks. The problem of computing a lifetime-optimal routing flow, link schedule, and link transmission powers is formulated as a non-linear optimization problem. We first restrict the link schedules to the class of interference-free time division multiple access (TDMA) schedules. In this special case we formulate the optimization problem as a mixed integer-convex program, which can be solved using standard techniques. For general non-orthogonal link schedules, we propose an iterative algorithm that alternates between adaptive link scheduling and computation of optimal link rates and transmission powers for a fixed link schedule. The performance of this algorithm is compared to other design approaches for several network topologies. The results illustrate the advantages of load balancing, multihop routing, frequency reuse, and interference mitigation in increasing the lifetime of energy-constrained networks. We also describe a partially distributed algorithm to compute optimal rates and transmission powers for a given link schedule.

Wireless systems where the nodes operate on batteries so that energy consumption must be minimized while satisfying given throughput and delay requirements are considered. In this context, the best modulation strategy to minimize the total energy consumption required to send a given number of bits is analyzed. The total energy consumption includes both the transmission energy and the circuit energy consumption. For uncoded systems, by optimizing the transmission time and the modulation parameters, it is shown that up to 80% energy savings is achievable over nonoptimized systems. For coded systems, it is shown that the benefit of coding varies with the transmission distance and the underlying modulation schemes.

Blind signal detection in multiuser code division multiple access (CDMA) system is particularly attractive when only the desired user signature is known to a given receiver. A problem common to several existing blind multiuser CDMA detectors is that the detection performance is very sensitive to the signature waveform mismatch (SWM) which may be caused by channel distortion. In this paper, we consider the design of a blind multiuser CDMA detector that is robust to the SWM. We present a convex formulation for this problem by using the second-order cone (SOC) programming. The resulting SOC problem can be solved efficiently using the recently developed interior point methods. Computer simulations indicate that the performance of our new robust blind multiuser detector is superior to those of many existing methods.

We consider radio applications in sensor networks where energy is a limited resource so that energy consumption must be minimized while satisfying given delay and throughput requirements. In this context, we analyze energy-efficient data collection strategies where we minimize the total energy consumption necessary for collecting a certain amount of data from multiple sensors. The total energy consumption includes both the transmission energy and the circuit energy consumption. We propose a variable-length TDMA scheme where the slot length is adaptively assigned according to the number of bits in the transmitting queues and the distance between the transmitting nodes and the collecting node. The underlying goal is to finish the collection of information bits from the multiple sensor nodes before a deadline T with minimum energy cost. We show that the problem can be efficiently solved by convex relaxation methods, and in some special cases simple analytical solutions can be derived.

We consider the problem of optimal power scheduling for the decentralized estimation of a noise-corrupted signal in an inhomogeneous sensor network. Sensor observations are first quantized into discrete messages, then transmitted to the fusion center where a final estimate is generated. Based on the sensor noise levels and channel gains from sensors to the fusion center, optimal quantization levels and transmit power levels at the local sensors can be chosen to minimize the total transmitting power, while ensuring a given mean squared error (MSE) performance. The proposed optimal power scheduling scheme suggests that the sensors with bad channels or poor observation qualities should decrease their quantization resolutions or simply become inactive in order to conserve power. For the remaining active sensors, their optimal quantization and transmit power levels are determined jointly by individual channel gains, local observation noise variance, and the targeted MSE performance. Numerical examples show that up to 60% energy savings is possible when compared with the uniform quantization strategy.

We consider radio applications in sensor networks, where the nodes operate on batteries so that energy consumption must be minimized, while satisfying given throughput and delay requirements. In this context, we analyze the best modulation and transmission strategy to minimize the total energy consumption required to send a given number of bits. The total energy consumption includes both the transmission energy and the circuit energy consumption. We first consider multi-input-multi-output (MIMO) systems based on Alamouti diversity schemes, which have good spectral efficiency but also more circuitry that consumes energy. We then extend our energy-efficiency analysis of MIMO systems to individual single-antenna nodes that cooperate to form multiple-antenna transmitters or receivers. By transmitting and/or receiving information jointly, we show that tremendous energy saving is possible for transmission distances larger than a given threshold, even when we take into account the local energy cost necessary for joint information transmission and reception. We also show that over some distance ranges, cooperative MIMO transmission and reception can simultaneously achieve both energy savings and delay reduction.

We consider radio applications where the nodes operate on batteries so that energy consumption must be minimized while satisfying given throughput and delay requirements. In this context, we analyze the best modulation strategy to minimize the total energy consumption required to send a given number of bits when error-control codes are used. The total energy consumption includes both the transmission energy and the circuit energy consumption. We show that for both MQAM and MFSK the total energy consumption may be reduced significantly if the transmission time T/sub on/ is optimized to reduce the sum of transmission energy and circuit energy consumption. Our optimization considers both delay and peak-power constraints. Numerical examples are given, where we exhibit up to 90% energy savings over modulation strategies that minimize the transmission energy alone. We also show that the benefit of coding varies with the transmission distance and the underlying modulation schemes.

We consider radio applications where the nodes operate on batteries so that energy consumption must be minimized while satisfying given throughput and delay requirements. In this context, we analyze the best modulation strategy to minimize the total energy consumption required to send a given number of bits. The total energy consumption includes both the transmission energy and the circuit energy consumption. We show that for both MQAM and MFSK the transmission energy decreases with the BT/sub on/ product while the circuit energy consumption increases with T/sub on/, where B is the modulation bandwidth and T/sub on/ is the transmission time. Thus, in short-range applications where the circuit energy consumption is nonnegligible compared with the transmission energy, the total energy consumption is minimized by using the maximum system bandwidth along with an optimized transmission time T/sub on/. We derive this optimal T/sub on/ for MQAM and MFSK modulation in both AWGN channels and Rayleigh fading channels. Our optimization considers both delay and peak-power constraints. Numerical examples are given, where we exhibit up to 68% energy savings over modulation strategies that minimize the transmission energy alone.

A common problem with the existing blind multiuser CDMA detectors is that their performance is very sensitive to the signature waveform mismatch (SWM) caused by channel distortion. We consider the problem of designing a blind multiuser CDMA detector which is robust to the SWM. We present a convex formulation for this problem by using the second order cone (SOC) programming. We also propose the use of recently developed interior point methods to efficiently solve the resulting SOC problem. Computer simulations indicate that the performance of our new robust blind multiuser detector is superior.