Virtual reality (VR) streaming demands both high data rates and low latency, requiring advanced transmission strategies to enhance system performance in wireless networks. This paper proposes a hierarchical rate-splitting multiple access-based cell-free (HRS-CF) VR transmission strategy, which integrates field of view (FoV)-based user grouping, scalable video coding (SVC)-based message design, and message-centric base station selection. Moreover, to characterize the statistical data rate of VR services under a given delay constraint, we investigate the effective capacity (EC) of the VR video streaming under HRS-CF strategy. Furthermore, we jointly optimize the precoding matrix, rate splitting coefficients, and BS selection using the proximal policy optimization (PPO) algorithm, where a power threshold-based BS selection is introduced to reduce computational complexity. Simulation results show that the proposed HRS-CF strategy outperforms the conventional rate splitting and CF transmission schemes, achieving at least a 16% performance improvement.

Multimodal information plays an important role in the advanced Internet of Things (IoT) in the era of 6G, which provides reliable and comprehensive assistance for downstream tasks through further fusion and analysis via federated learning (FL). One of the primary challenges in FL is data heterogeneity, which may lead to domain shifts and sharply different local long-tailed category distribution across nodes. These issues hinder the large-scale deployment of FL in IoT applications equipped with multiple various multimodal sensors due to performance deterioration. In this paper, we propose a novel multimodal fusion framework to tackle the aforementioned coupled problems arising during the cooperative fusion of multimodal information without privacy exposure among decentralized nodes equipped with diverse sensors. Specifically, we introduce a flexible global logit alignment (GLA) method based on multi-view domains. This method enables the fusion of diverse multimodal information with the consideration of domain shifts caused by modality-based data heterogeneity. Furthermore, we propose a novel local angular margin (LAM) scheme, which dynamically adjusts decision boundaries for locally seen categories while preserving global decision boundaries for unseen categories. This effectively mitigates severe model divergence caused by significantly different category distributions. Extensive simulations demonstrate the superiority of the proposed framework, which exhibits significant merits in tackling model degeneration caused by data heterogeneity and enhancing modality-based generalization for heterogeneous scenarios.

The integrated sensing and communication (ISAC) technology has gained increasing attention in recent years due to its excellent performance of increasing the spectrum and hardware efficiencies. In this paper, we investigate the joint optimization of beam selection and digital beamforming for a millimeter-wave (mmWave) ISAC system to simultaneously improve the performance of communication and sensing. We propose a novel hybrid beamforming scheme based on deep learning by maximizing the sum of communication mutual information (CMI) and sensing mutual information (SMI) to enable multi-user multiple-input multiple-output (MU-MIMO) communication and multiple-input single-output (MISO) radar sensing. Specially, we propose a joint deep reinforcement learning and unfolding network (DRL-UN) to optimize the beam selection and digital beamforming matrices at the base station (BS) in an ISAC system. Simulation results demonstrate that the proposed hybrid beamforming scheme significantly outperforms the existing algorithms in terms of sensing and communication (S&C) sum-rate in a mmWave ISAC system.

Vehicular cooperative perception enhances the reliability and safety of autonomous driving systems by sharing perception information among vehicles. However, it often leads to issues of information redundancy and communication resource waste. To address the challenge of decreased communication efficiency due to frequent messaging, this paper proposes a joint selection method for collaborative agents and content. Firstly, we model the problem of jointly selecting collaborators and cooperative content as a parameterized action Markov decision process, where the action space is represented as a Multi-Discrete Action Space. Secondly, to improve perception performance and reduce communication resource consumption of vehicles, a deep reinforcement learning-based late fusion method is proposed to decouple the problem into two parts: cooperative agent selection managed by the Road Side Unit (RSU) and content selection handled by the vehicles. Finally, experimental results demonstrate that the proposed joint selection method with Dueling Deep Q-Networks (Dueling DQN) for cooperative perception achieves superior performance in improving perception confidence scores and reducing communication consumption.

Drawing upon a data-driven methodology, deep learning has emerged as an innovative approach for dynamic resource allocation in large-scale cellular networks. This paper proposes an optimization strategy relying on domain adversarial networks to reduce the number of poorly performing base stations (BSs). The approach dynamically allocates radio resources to address real-time mobile traffic needs. We calculate the interference coefficients among BSs and design a performance classifier that evaluates BS performance with respect to provided traffic-resource pairs as either poor or good. Most importantly, we use well-performing BSs as source domain data to reallocate the resources of poorly performing ones through the domain adversarial neural network. Experimental results demonstrate that the proposed domain adversarial resource allocation strategy effectively decreases the number of poorly performing BSs in the cellular network, which in turn outperforms other benchmark algorithms in terms of both the ratio of poor BSs and radio resource consumption.

Target parameter estimation in high-speed scenarios is one of the main challenges in the integrated sensing and communication (ISAC) systems. In an ISAC system, the orthogonal time frequency space (OTFS) signal is able to successfully combat time-frequency-selective channels since the channel exhibits significant delay-Doppler (DD) sparsity characteristic. In this paper, we investigate the problem of parameter estimation of moving targets using OTFS modulation. We firstly derive signal model in the DD domain equivalent channel and recast the problem of parameter estimation into a compressed sensing (CS) problem. In order to improve the estimation performance, we then propose ADMM-Net by deep unfolding the iterations of the Alternating Direction Method of Multipliers (ADMM) algorithm into a deep learning network. Experimental results demonstrate that the proposed ADMM-Net algorithm outperforms the other methods in terms of estimation accuracy and running time for OTFS-based parameter estimation.

In the 6G environment, addressing the challenges of data loss and off-grid issues during target parameter estimation poses a significant challenge for the Integrated Sensing and Communication (ISAC) system. In the ISAC framework, a commonly used method for parameter estimation is compressive sensing. However, compressive sensing may encounter off-grid issues in continuous parameter estimation. In contrast, the atomic norm proves effective in addressing off-grid problems, making it more suitable for continuous parameter estimation. We explore the application of the atomic norm in ISAC and further derive an ISAC model based on OFDM (Orthogonal Frequency Division Multiplexing) utilizing the atomic norm under conditions of incomplete data. To ensure improved convergence speed and accuracy of our algorithm, we employ the Alternating Direction Method of Multipliers (ADMM) for iterative implementation. Experimental results demonstrate that our proposed AN algorithm accurately estimates target parameters in the presence of data loss, exhibiting higher precision and robustness compared to traditional methods.

Decentralized federated learning (DFL) is a novel distributed machine-learning paradigm where participants collaborate to train machine-learning models without the assistance of the central server. The decentralized framework can effectively overcome the communication bottleneck and single-point-of-failure issues encountered in federated learning (FL). However, most existing DFL methods may ignore the communication resource constraints of the system. This may result in these methods unsuitable for many practical scenarios because the given resource constraints cannot be guaranteed. In this article, we propose a novel DFL, called DFL with adaptive compression ratio (AdapCom-DFL), that can adaptively adjust the compression ratio of transmission data to keep the communication latency within the constraint. Furthermore, we propose a communication network topology pruning approach to reduce communication overhead by pruning poor links with low data rates while ensuring the convergence. Additionally, a power allocation approach is presented to improve the performance by reallocating the power of communication links while complying with the communication energy constraint. Extensive simulation results demonstrate that the proposed AdapCom-DFL with network pruning and power allocation approach achieves better performance and requires less bandwidth under the given resource constraints compared with some existing approaches.

Federated learning (FL), as a privacy-enhancing distributed learning paradigm, has recently attracted much attention in wireless systems. By providing communication and computation services, the base station (BS) helps participants collaboratively train a shared model without transmitting raw data. Concurrently, with the advent of integrated sensing and communication (ISAC) and the growing demand for sensing services, it is envisioned that BS will simultaneously serve sensing services, as well as communication and computation services, e.g., FL, in future 6G wireless networks. To this end, we provide a novel integrated sensing, communication and computation (ISCC) system, called Fed-ISCC, where BS conducts sensing and FL in the same time-frequency resource, and the over-the-air computation (AirComp) is adopted to enable fast model aggregation. To mitigate the interference between sensing and FL during uplink transmission, we propose a receive beamforming approach. Subsequently, we analyze the convergence of FL in the Fed-ISCC system, which reveals that the convergence of FL is hindered by device selection error and transmission error caused by sensing interference, channel fading and receiver noise. Based on this analysis, we formulate an optimization problem that considers the optimization of transceiver beamforming vectors and device selection strategy, with the goal of minimizing transmission and device selection errors while ensuring the sensing requirement. To address this problem, we propose a joint optimization algorithm that decouples it into two main problems and then solves them iteratively. Simulation results demonstrate that our proposed algorithm is superior to other comparison schemes and nearly attains the performance of ideal FL.

Federated learning (FL) has been extensively studied as a means of ensuring data privacy while cooperatively training a global model across decentralized devices. Among various FL approaches, asynchronous federated learning (AFL) has distinct advantages in overcoming the straggler problem via server-side aggregation as soon as it receives a local model. However, AFL still faces several challenges in large-scale real-world applications, such as stale model problems and modality heterogeneity across geographically distributed and industrial devices with different functions. In this article, we propose a multimodal fusion framework for AFL to address the aforementioned problems. Specifically, a novel multilinear block fusion model is designed to fuse various multimodal information, which serves as an enhancement for perceiving and transmitting the important modality and block during local training. An adaptive aggregation strategy is further developed to fully utilize heterogeneous data by allowing the global model to favor the received local model based on both freshness and the importance of the local data. Extensive simulations with different data distributions demonstrate the superiority of the proposed framework in heterogeneity scenarios, which exhibits significant merits in the improvement of modality-based generalization without sacrificing convergence speed and communication consumption.

We consider the problem of hyperspectral image (HSI) reconstruction, which aims to recover 3D hyperspectral data from 2D compressive HSI measurements acquired by a coded aperture snapshot spectral imaging (CASSI) system. Existing deep learning methods have achieved acceptable results in HSI reconstruction. However, these methods did not consider the imaging system degradation pattern. In this article, based on observing the initialized HSIs obtained by shifting and splitting the measurements, we propose a dynamic Fourier network based on degradation learning, called the degradation-aware dynamic Fourier-based network (DADF-Net). We estimate the degradation feature maps from the degraded hyperspectral images to realize the linear transformation and dynamic processing of the features. In particular, we use the Fourier transform to extract the HSI non-local features. Extensive experimental results show that the proposed model outperforms state-of-the-art algorithms on simulation and real-world HSI datasets.

Relying on a data-driven methodology, deep learning has emerged as a new approach for dynamic resource allocation in large-scale cellular networks. This paper proposes a knowledge-assisted domain adversarial network to reduce the number of poorly performing base stations (BSs) by dynamically allocating radio resources to meet real-time mobile traffic needs. Firstly, we calculate theoretical inter-cell interference and BS capacity using Voronoi tessellation and stochastic geometry, which are then incorporated into a neural network as key parameters. Secondly, following the practical assessment, a performance classifier evaluates BS performance based on given traffic-resource pairs as either poor or good. Most importantly, we use well-performing BSs as source domain data to reallocate the resources of poorly performing ones through the domain adversarial neural network. Our experimental results demonstrate that the proposed knowledge-assisted domain adversarial resource allocation (KDARA) strategy effectively decreases the number of poorly performing BSs in the cellular network, and in turn, outperforms other benchmark algorithms in terms of both the ratio of poor BSs and radio resource consumption.

Mobile edge computing (MEC) and field-of-view (FoV) prediction are two key techniques to enable the wireless virtual reality (VR) service. On this basis, we investigate a practical issue, that is, how to efficiently achieve the user's pre-set quality-of-experience (QoE) requirement on both video quality and delay tolerance. A constrained reward-steering algorithm based on reinforcement learning is proposed in this work to solve this multi-objective optimization problem, which finds the optimal policy approaching the user's targeted QoE. Meanwhile, both an instantaneous service delay constraint and a long-term energy constraint are satisfied by the Lagrangian-based method. Simulation results demonstrate that the proposed algorithm outperforms conventional reinforcement learning relying on weights, i.e. achieving an average reward vector much closer to the user's targeted QoE, and meeting both constraints.

Images collected in low-light environments usually suffer from multiple, non-uniform distributed distortions, including local dark, dim light, backlit and so on. In this paper, we propose a Stage-Transformer-Guided Network (STGNet) that effectively handles region-specific distributions and enhance diverse low-light images. Specifically, our STGNet adopts a multi-stage way to progressively learn hierarchical features that benefit the robustness of our model. At each stage, we design an efficient transformer with horizontal and vertical attentions that jointly capture degradation distributions with different magnitudes and orientations. We also introduce learnable degradation queries to adaptively select task-specific features of degradations for enhancement. In addition, we design a histogram loss for enhancement and combine it with other loss functions, in order to exploit both global contrast and local details during network training. Benefiting from the above contributions, our STGNet achieves the state-of-the-art performances on both synthetic and real-world datasets.

Hyperspectral images (HSIs) contain rich spatial and spectral information. A double dispersers coded aperture snapshot spectral imaging (DD-CASSI) system takes advantage of compressive sensing (CS) theory to map 3D HSI data into a single 2D measurement. One of key components of DD-CASSI is to reconstruct high quality hyperspectral image from measurement. Traditional model-based methods use mathematical optimization to reconstruct hyperspectral images according to prior knowledge. Current deep learning based methods achieve pleasant results. But fully learned deep learning methods lack interpretability, and model-based deep learning methods cannot achieve pleasant performance. In this paper, we propose a novel HSI reconstruction framework named Refinement Boosted and Attention Guided Tensor FISTA(Fast Iterative Shrinkage-Thresholding Algorithm)-Net (ReAttFISTA-Net), which combines model-based deep learning and fully learned deep learning reconstruction strategies. In this framework, we introduces Attention Guided Fusion Mechanism which enhances spatial-spectral information, refinement sub-network and auxiliary loss terms to improve the reconstruction performance. Extensive experimental results show that the proposed reconstruction algorithm outperforms the state-of-the-art algorithms on both simulation and real-world datasets.

Federated learning (FL) provides a novel framework to collaboratively train a shared model in a distribution fashion by virtue of a central server. However, FL is inappropriate for a serverless scenario and also suffers from some major drawbacks in Industrial Internet of Things (IIoT) networks, such as unresilience to network failures and communication bottleneck effect. In this article, we propose a novel decentralized federated learning (DFL) approach for IIoT devices to achieve model consensus by exchanging model parameters only with their neighbors rather than a central server. We firstly formulate the problem of model consensus in DFL as a fastest mixing Markov chain problem and then optimize the consensus matrix to improve the convergence rate. Meanwhile, a practical medium access control protocol with time slotted channel hopping is taken into account to implement the proposed approach. Furthermore, we also propose an accumulated update compression method to alleviate communication cost. Finally, extensive simulation results demonstrate that the proposed approach improves accuracy and reduces communication cost especially under the nonindependent identically distribution data distribution.

It is important to estimate the global network traffic data from partial traffic measurements for many network management tasks, including status monitoring and fault detection. However, existing estimation approaches cannot well handle the topological correlations hidden in network traffic and suffer from limited imputation performance. This paper proposes a deep learning approach for network traffic imputation, which well exploits the topological structure of network traffic. We first model the network traffic as a novel graph-tensor and derive a theoretical recovery guarantee. Then we develop an iterative graph-tensor completion algorithm and propose a graph neural network for network traffic imputation by unfolding the iterative algorithm. The proposed graph neural network well captures the topological correlations of network traffic and achieves accurate imputation. Extensive experiments on real-world datasets show that the proposed graph neural network achieves about one-half lower relative square error while at least ten times faster imputation speed than the existing methods.

In order to mitigate thedouble near-far effect, we focus on a mobile wireless-powered communication network (WPCN), where sensor nodes harvest energy from the radio frequency (RF) signal of the mobile energy access point (EAP), and transmit data to the mobile data access point (DAP) by using the harvested energy. Only the sensor nodes with energy larger than a threshold, which is mainly determined by the energy consumption of one transmission, have opportunities to transmit data. Due to the mobility of the EAP and DAP, the distance between the EAP and DAP changes over time. When the DAP moves into the operation region of the EAP, the EAP and DAP could not work simultaneously due to the severe interference, and an energy harvesting probability is employed to denote the probability that the EAP works in this scenario. The purpose of this paper is to identify the optimal transmission policy, i.e., the optimal pairing of the energy consumption of one transmission and the energy harvesting probability, that maximizes the throughput of the WPCN under an energy causality constraint. By analyzing the energy causality constraint, we show that the WPCN could be divided into an energy-sufficient state and an energy-limited state by the pairing of the energy consumption of one transmission and the energy harvesting probability. Since the energy consumption of one transmission and the energy harvesting probability are jointly intertwined with the energy causality constraint, making the joint optimization problem intractable, we divide the throughput maximization problem into two layers. In the inner problem, we investigate the optimal energy consumption of one transmission with a given energy harvesting probability. In the outer problem, we derive the optimal energy harvesting probability based on the obtained optimal energy consumption of one transmission. According to the aforementioned investigations, we propose a two-layer algorithm to obtain the specific solution. Numerical results are conducted to validate the theoretical results and the efficiency of the proposed two-layer algorithm.

Network latency is an important metric for network performance evaluation. However, device faults inevitably occur during the data collection process, resulting in abnormal data or even missing data. It is desirable to accurately estimate the network latencies and detect the abnormal data. Existing approaches can not provide reliable performance due to the limitation in exploiting the spatial-temporal correlations of latency data. In this paper, we propose a novel Robust Spatial-Temporal Graph-Tensor Recovery (RSTGTR) algorithm which simultaneously recovers the missing data and detects the anomalies in network latencies. Firstly, we model the network latency data as a novel graph-tensor for exploring the topological structure of networks. Secondly, we develop spatial-temporal constraints and propose a graph-tensor recovery algorithm (RSTGTR). Finally, we conduct extensive experiments to evaluate the performance of the proposed algorithm by using a real-world latency dataset. Experimental results show the proposed algorithm outperforms existing methods in terms of latency estimation and anomaly detection.

Federated learning, as an effective machine learning paradigm, can collaboratively training an efficient global model by exchanging the network parameters between edge nodes and the cloud server without sacrificing data privacy. Unfortunately, the obtained global model cannot generalize to newly collected unlabeled data since the unlabeled data collected by different edge devices are diverse. Furthermore, the distributions of collected labeled data and unlabeled data are also different for edge devices. In this paper, we propose a method named Incremental Unsupervised Adversarial Domain Adaptation (IUADA) for federated learning, which aims to reduce the domain shift between the labeled data and unlabeled data in the edge nodes and enhance the performance of the personalized target domain models based on the local unlabeled data. Finally, we evaluate the performance of the proposed method by using three real-world datasets. Extensive experimental results demonstrate that the proposed method is efficient to solve the problem of domain shift and achieves a better performance for unlabeled data for federated learning.

In intelligent transportation systems (ITS), incomplete traffic data due to sensor malfunctions and communication faults, seriously restricts the related applications of ITS. Recovering missing data from incomplete traffic data becomes an important issue for ITS. Existing works on traffic data imputation cannot achieve satisfactory accuracy due to inefficiently exploiting the underlying topological structure of the traffic data. In this paper, we model the topology of the road network as a graph and introduce graph Fourier transform (GFT) to process the traffic data. Then we utilize an algebraic framework termed as graph-tensor singular value decompositions (GT-SVD) to extract the hidden spatial information of traffic data. Furthermore, we propose a novel graph spectral regularized tensor completion algorithm based on GT-SVD and construct temporal regularized constraints to improve the recovery accuracy. The extensive experimental results on real traffic datasets demonstrate that the proposed algorithm outperforms the state-of-the-art methods under different missing patterns.

Traffic information can be used for real-time traffic management and long-term transportation planning to increase traffic efficiency and safety. However, data containing both missing and deviating values, can seriously affect the accuracy of traffic information, even leading to incorrect results in traffic data analysis. In this paper, we propose a novel tensor-based data recovery method named spatial-temporal tensor robust principal component analysis (ST-TRPCA) to recover traffic data from corrupted and incomplete observations. Specifically, we not only fully account for the spatial-temporal properties of traffic data to increase the data recovery accuracy, but also utilize tensor factorization and its low-dimensional representation to improve computational efficiency. The extensive experimental results performed on real-world traffic dataset under various scenarios show that ST-TRPCA outperforms other state-of-the-art methods in both missing data recovery and anomaly detection, especially when the traffic data are severely corrupted.

Federated learning, as a novel machine learning paradigm, aims to collaboratively train a global model while keeping the training data on local devices, which protects data privacy and security of distributed devices. However, the model cannot generalize to new devices because of domain shift caused by the statistical difference between the labeled data and unlabeled data collected by different devices in heterogeneous internet of things networks. In this paper, we propose a method named Unsupervised Federated Adversarial Domain Adaptation with Controller Modules (UFADACM), which aims to reduce the distribution difference between source nodes with labeled data and target nodes with unlabeled data, and reduce the parameter cost and communication overhead while achieving a comparable performance. We also conduct extensive experiments to demonstrate the effectiveness of the proposed method.

Intelligent reflecting surfaces (IRSs) have been proposed in recent years as a promising technology to enhance the quality of transmissions in high-frequency spectrum. Currently, the research on the performance of large networks with IRSs is still in its infancy. Different from the commonly-used stochastic geometry model for the study of traditional networks, where only transmitters and receivers are modeled as point processes, in an IRS network, the blockages and reflectors also need to be accounted for. In this paper, we study a bipolar network with a line segment object model, and derive the probability that an IRS can successfully reflect a signal from a transmitter to a receiver, as well as the distribution of the distance traveled by the reflected signal. With these analytic results, the signal to interference ratio (SIR) and the achievable rate are obtained in closed-form expressions. From the analysis, we can observe that IRSs have a great potential to enhance the network performance, as they are able to boost the signal power, while preventing the inter-cell interference from rising rapidly. More importantly, we find that even with a limited number of IRSs, the network can still achieve a higher achievable rate than a conventional one without IRSs.

Deep learning has recently garnered significant interest in many applications especially for big data analytics in the edge computing environment. Federated learning, as a novel machine learning technique, aims to build a shared learning model from training data on distributed edge nodes to protect data privacy. However, the model update in federated learning requires parameter exchanges among edge nodes, which is rather bandwidth-consuming. This article proposes a novel distributed hierarchical tensor deep computation model by condensing the model parameters from a high-dimensional tensor space into a set of low-dimensional subspaces to reduce the bandwidth consumption and storage requirement for federated learning. Moreover, an updating approach with a hierarchical tensor back-propagation algorithm is developed by directly computing the gradients of low-dimensional parameters to reduce the memory requirement of training for edge nodes and improve training efficiency. Finally, extensive simulations on classical datasets with different local data distributions are presented for the performance evaluation. The results demonstrate that the proposed model relieves the burden of communication bandwidth and reduces energy consumption at edge nodes for federated learning.

Accurate short-time traffic flow prediction has gained gradually increasing importance for traffic plan and management with the deployment of intelligent transportation systems (ITSs). However, the existing approaches for short-term traffic flow prediction are unable to efficiently capture the complex nonlinearity of traffic flow, which provide unsatisfactory prediction accuracy. In this paper, we propose a deep learning based model which uses hybrid and multiple-layer architectures to automatically extract inherent features of traffic flow data. Firstly, built on the convolutional neural network (CNN) and the long short-term memory (LSTM) network, we develop an attention-based Conv-LSTM module to extract the spatial and short-term temporal features. The attention mechanism is properly designed to distinguish the importance of flow sequences at different times by automatically assigning different weights. Secondly, to further explore long-term temporal features, we propose a bidirectional LSTM (Bi-LSTM) module to extract daily and weekly periodic features so as to capture variance tendency of the traffic flow from both previous and posterior directions. Finally, extensive experimental results are presented to show that the proposed model combining the attention Conv-LSTM and Bi-LSTM achieves better prediction performance compared with other existing approaches.

In recent years, with the growing prevalence of personal devices, network latency of devices has drawn much attention due to its significant influence on user experience. Thus network latency estimation is considered to be an important index for network performance evaluation. However, the existing works on network latency estimation are unable to achieve satisfactory estimation accuracy due to the adoption of the conventional matrix or tensor model. In this paper, we construct a novel tensor model based on tensor-SVD for network latency data to make full use of its potential latent factors. Besides, we also propose a graph-laplacian regularization tensor completion algorithm (GLRTC), which mines the underlying spatial information by introducing graph-laplacian regularization constraints to improve the recovery performance. Finally, we conduct extensive simulations on the real-world latency dataset and demonstrate the effectiveness of the proposed algorithm. Comparing with the existing approaches, the proposed algorithm achieves significant improvement in terms of recovery accuracy.

Aiming at the problem that mobile terminal (MT) harvests less energy from ambient radio frequency (RF) sources, the resource allocation strategy in mobile edge computing (MEC) system with hybrid energy harvesting is investigated in this paper. By deploying multiple magnetic induction energy quick charging stations (MI-CSs) within the coverage area of the base station, the MT can supplement extra energy at a nearby MI-CS when the energy harvested from ambient RF sources is about to be exhausted. The MT offloads computing task to edge server by leveraging MEC technology. The resource allocation problem is formulated as an optimization problem. The objective is to minimize the total energy consumption of MTs under the constraints of computing capability range of MT, maximal computing resource of edge server, computing delay of task, and battery energy of MT. The suboptimal solution is obtained by adopting the quantum-behaved particle swarm optimization (QPSO) algorithm. Simulation results show that the QPSO algorithm has less energy consumption compared with the standard particle swarm optimization algorithm and the fixed computing resource allocation method.

In recent years, end-to-end network latency estimation has attracted much attention because of its significance for network performance evaluation. Given the widespread use of personal devices, latency estimation from partially observed samples becomes more complicated due to unstable communication conditions, while measuring the latencies between all nodes in a large-scale network is infeasible and costly. Hence, reducing the measurement cost becomes critical for the latency estimation of personal device network. In this paper, we propose an adaptive sampling scheme based on leverage scores to reduce the measurement cost while achieving high estimation accuracy. Furthermore, we provide theoretical analysis to characterize the performance bounds of the proposed scheme in terms of sampling complexity and estimation error. Finally, we demonstrate the efficiency of the proposed scheme by conducting extensive simulations on both synthetic and real datasets. The results show that the proposed scheme is able to not only improve the estimation accuracy of network latency but also reduce the sample budget compared to the state-of-the-art approaches.

In recent years Wi-Fi fingerprinting has attracted much attention in indoor localization because of the availability of high-quality signal and pervasive deployment of wireless LANs. For fingerprint-based localization, however, offline site survey is usually time-consuming and labor-intensive. Therefore, reducing the burden of offline site survey becomes an important issue for fingerprint-based indoor localization. In this paper, using a low-tubal-rank tensor to model Wi-Fi fingerprints of all reference points (RPs), we propose an adaptive sampling scheme via approximate volume sampling to improve reconstruction accuracy of radio map with reduced expenditure. We propose a rank-increasing strategy to effectively estimate the rank of the underlying fingerprint tensor to alleviate the computation burden for tensor completion. We provide a theoretical foundation to analyze the proposed scheme and derive the performance bounds in terms of sample complexity and reconstruction error. We prove that the proposed scheme can achieve a relative error guarantee. Finally, we validate the effectiveness of the proposed scheme through extensive simulations using both synthetic and real datasets. The simulation results demonstrate that the proposed scheme is able to not only reduce reconstruction error and improve localization accuracy but also reduce running time compared to the state-of-the-art schemes.

Accurate estimation of the traffic state can help to address the issue of urban traffic congestion, providing guiding advices for people's travel and traffic regulation. In this paper, we propose a novel short-term traffic flow prediction algorithm based on an adaptive multi-kernel support vector machine (AMSVM) with spatial-temporal correlation, which is named as AMSVM-STC. First, we explore both the nonlinearity and randomness of the traffic flow, and hybridize Gaussian kernel and polynomial kernel to constitute the AMSVM. Second, we optimize the parameters of AMSVM with the adaptive particle swarm optimization algorithm, and propose a novel method to make the hybrid kernel's weight adjust adaptively according to the change tendency of real-time traffic flow. Third, we incorporate the spatial-temporal correlation information with AMSVM to predict the short-term traffic flow. We evaluate our algorithm by doing thorough experiment on real data sets. The results demonstrate that our algorithm can do a timely and adaptive prediction even in the rush hour when the traffic conditions change rapidly. At the same time, the proposed AMSVM-STC outperforms the existing methods.

The recent advances of compressive sensing (CS) have witnessed a great potential of efficient compressive data gathering (CDG) in wireless sensor network systems (WSNSs). However, most existing work on CDG mainly focuses on multihop relaying strategies to improve the performance of data gathering. In this paper, we propose a mobile CDG scheme including a random walk-based algorithm and a kernel-based method for sparsifying sensory data from irregular deployments. The proposed scheme allows a mobile collector to harvest data by sequentially visiting a number of nodes along a random path. More importantly, toward building the gap between CS and machine learning theories, we explore a theoretical foundation for understanding the feasibility of the proposed scheme. We prove that the CS matrices, constructed from the proposed random walk algorithm combined with a kernel-based sparsity basis, satisfy the restricted isometry property. Particularly, we also show that m = O(klog(n/k)) measurements collected by a mobile collector are sufficient to recover a k-sparse signal and t = O(klog(n/k)) steps are required to collect these measurements in a network with n nodes. Finally, we also present extensive numerical results to validate the effectiveness of the proposed scheme by evaluating the performance in terms of energy consumption and the impact of packet losses. The numerical results demonstrate that the proposed scheme is able to not only significantly reduce communication cost but also combat unreliable wireless links under various packet losses compared to the stateof-the-art schemes, which provides an efficient alternative to data relaying approaches for CDG in WSNS.

Participatory sensing is a rising paradigm which utilizes mobile phones to collect data and build application on the cloud. But there are many problems to be resolved, poor quality of received information caused by task executors has been one of them. So incentive mechanism is essential for attracting users to participate in and submit high-quality data. Inspired by contract theory, we model participatory sensing as a contractual relationship and devote to design reasonable rewards for relevant results to maximize the benefit of task publisher. Under complete information scenario where task executors' efforts can be observed and incomplete information scenario where task executors' efforts can not be observed, we take advantage of maximization problem to infer the optional contract reward for task executors. In addition, based on the utility of task publisher, we propose optimal effort and optimal effort discriminant inequality (OEDI). Furthermore, we discuss the influence of noise, cost and boundary on optimal effort from aspects of theory and reality. Finally, we evaluate our contract-based approach by thorough simulations to show its effectiveness and accuracy.

Accurate prediction of the traffic state can help to solve the problem of urban traffic congestion, providing guiding advices for people's travel and traffic regulation. In this paper, we propose a novel short-term traffic flow prediction algorithm, which is based on Multi-kernel Support Vector Machine (MSVM) and Adaptive Particle Swarm Optimization (APSO). Firstly, we explore both the nonlinear and randomness characteristic of traffic flow, and hybridize Gaussian kernel and polynomial kernel to constitute the MSVM. Secondly, we optimize the parameters of MSVM with a novel APSO algorithm by considering both the historical and real-time traffic data. We evaluate our algorithm by doing thorough experiment on a large real dataset. The results show that our algorithm can do a timely and adaptive prediction even in the rush hour when the traffic conditions change rapidly. At the same time, the prediction results are more accurate compared to four baseline methods.

Abstract Task allocation is an important issue in wireless sensor networks (WSNs), and the existing traditional solutions to this problem in high‐performance computing cannot be directly implemented in WSNs because of limitations such as resource availability and shared communication medium. In this paper, we address the task allocation problem for a heterogeneous WSN, and a trust dynamic task allocation algorithm is proposed. Firstly, to ensure the nodes in the same coalition are mutually closer in distance, a discrete particle swarm optimization (PSO) is designed to generate a structure of the parallel coalitions. Secondly, in order to minimize the execution time of the tasks, save the energy cost of the nodes and balance the load of the network, we design task strategies and payoff functions by invoking the game theory in WSNs and propose a PSO with the redesigned fitness function to find the Nash equilibrium point for the purpose of improving the effectiveness of scheduling and the reliability of the network. In this step, the sink node will play the role of trust manager, and it will allocate tasks based on the Nash equilibrium point, which is a trust solution to make sure all tasks can be finished. Finally, the extensive experiments are conducted to compare our algorithm with two other algorithms. The experimental results show the feasibility and effectiveness of our algorithm, which can obtain a good balance between local solution and global exploration and achieve superior energy efficiency and network reliability within a short period. Copyright © 2014 John Wiley & Sons, Ltd.

In this paper, we study the problem of data gathering with compressive sensing (CS) in wireless sensor networks (WSNs). Unlike the conventional approaches, which require uniform sampling in the traditional CS theory, we propose a random walk algorithm for data gathering in WSNs. However, such an approach will conform to path constraints in networks and result in the non-uniform selection of measurements. It is still unknown whether such a non-uniform method can be used for CS to recover sparse signals in WSNs. In this paper, from the perspectives of CS theory and graph theory, we provide mathematical foundations to allow random measurements to be collected in a random walk based manner. We find that the random matrix constructed from our random walk algorithm can satisfy the expansion property of expander graphs. The theoretical analysis shows that a k-sparse signal can be recovered using `1 minimization decoding algorithm when it takes m = O(k log(n=k)) independent random walks with the length of each walk t = O(n=k) in a random geometric network with n nodes. We also carry out simulations to demonstrate the effectiveness of the proposed scheme. Simulation results show that our proposed scheme can significantly reduce communication cost compared to the conventional schemes using dense random projections and sparse random projections, indicating that our scheme can be a more practical alternative for data gathering applications in WSNs.

Compressive sensing (CS) provides a new paradigm for efficient data gathering in wireless sensor networks (WSNs). In this paper, with the assumption that sensor data is sparse we apply the theory of CS to data gathering for a WSN where n nodes are randomly deployed. We investigate the fundamental limitation of data gathering with CS for both single-sink and multi-sink random networks under protocol interference model, in terms of capacity and delay. For the single-sink case, we present a simple scheme for data gathering with CS and derive the bounds of the data gathering capacity. We show that the proposed scheme can achieve the capacity Θ(\frac{nW}{M}) and the delay Θ(M\sqrtfrac{nlog n}), where W is the data rate on each link and M is the number of random projections required for reconstructing a snapshot. The results show that the proposed scheme can achieve a capacity gain of Θ (\frac{n}{M}) over the baseline transmission scheme and the delay can also be reduced by a factor of Θ(\fracsqrt{n\log n}{M}). For the multi-sink case, we consider the scenario where n_d sinks are present in the network and each sink collects one random projection from n_s randomly selected source nodes. We construct a simple architecture for multi-session data gathering with CS. We show that the per-session capacity of data gathering with CS is Θ(\frac{n\sqrt{n}W}{M n_d \sqrt{n_s \log n}}) and the per-session delay is Θ(M\sqrtfrac{{n}{log n}}). Finally, we validate our theoretical results for the scaling laws of the capacity in both single-sink and multi-sink networks through simulations.

In this paper, we study data gathering with compressive sensing from the perspective of in-network computation in random networks, in which n nodes are uniformly and independently deployed in a unit square area. We formulate the problem of data gathering to compute multiround random linear function. We study the performance of in-network computation with compressive sensing in terms of energy consumption and latency in centralized and distributed fashions. For the centralized approach, we propose a tree-based protocol for computing multiround random linear function. The complexity of computation shows that the proposed protocol can save energy and reduce latency by a factor of Θ(√n= log n) for data gathering comparing with the traditional approach, respectively. For the distributed approach, we propose a gossip-based approach and study the performance of energy and latency through theoretical analysis. We show that our approach needs fewer transmissions than the scheme using randomized gossip.

Compressive sensing (CS) provides a new paradigm for efficient data gathering in wireless sensor networks (WSNs). The theory of CS allows to reconstruct all sensor data of the network, while only collecting a small number of measurements at a sink. In this paper, we consider a scenario where a sink collects spatially correlated sensor data from n sensor nodes randomly deployed in a region. We investigate the fundamental limitation of data gathering with CS in such a scenario, in terms of capacity and delay. We construct a scheduling and routing scheme based on CS for data gathering in WSNs. We show that the proposed scheme can achieve a per-node transport capacity of Θ(1/ log n) under physical interference model. Furthermore, we also study the delay performance of the proposed scheme and show that the delay for collecting a snapshot with CS is Θ(√n log n). In particular, our results demonstrate that the proposed scheme can achieve a capacity gain of Θ(n/log n) over the case without CS and the delay can also be reduced by a factor of Θ(√n/log n).

Compressed sensing is an emerging theory which provides a new framework for sampling and compressing a sparse signal simultaneously at a reduced sampling rate. Besides this, compressed sensing also provides a new approach for the task of detection. Detection from compressive measurements without reconstructing the signals remains as a challenging problem. In this paper, we investigate the performance of compressive detection and propose a sequential compressive detection scheme to reduce the number of measurements for target detection in wireless sensor networks. We derive the sequential compressive decision rules and analyze its detection performance in terms of the number of measurements. Simulations show that sequential compressive detection can save about 50 percents of the average number of measurements under a given detection performance requirement compared with that of compressive detection.

MIMO-OFDM is a promising technique for the broadband wireless communication system. In this paper, we propose a novel scheme that integrates multiple description coding (MDC), error resilient video coding, and unequal error protection strategy with various space time coding codes for robust video transmission over MIMO-OFDM system. The proposed MDC coder generates multiple bitstreams of equal importance which are very suitable for multiple antennas system. Furthermore, according to the contribution to the reconstructed video quality, we apply unequal error protection strategy using BLAST and STBC space time codes for each video bitstream. Experimental results have demonstrated that the proposed scheme can achieve desired tradeoff between the reconstructed video quality and the transmission efficiency