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.

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.

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.

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.

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.

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.

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.

Multiple Radio Access Technologies (Multi-RAT) make it possible to exploit the advantages of Heterogeneous networks (HetNets) resulting from a joint consideration of the networks as a whole. Users in HetNets can be served with a proper RAT to maximize the system-level utility. Especially, when user dynamics are considered, they can stay in a RAT or handover to another RAT with a transition probability depending on system configuration. By formulating these dynamics as a Markov chain model, the system-level utility is defined as a combinatorial object function. However, the combinatorial optimization is NP-hard, thus we can only use exhaustive search to obtain the optimum solution, which comes up with high computational complexity and is not practical. To this end, we use Markov approximation to obtain the approximate utility and transition probability. In addition, we propose a Count Down and Select (CDS) algorithm to implement the RAT selection. Numerical results validate the convergence of Markov approximation and the effectiveness of the CDS algorithm.

We consider the problem of cooperative spectrum sharing among primary users PUs and secondary users SUs in cognitive radio networks. In our system, each PU selects a proper set of SUs to serve as the cooperative relays for its transmission and in return, leases portion of channel access time to the selected SUs for their own transmission. PU decides how to select SUs and how much time it would lease to SUs, and the cooperative SUs decide their respective power levels in helping PU's transmission, which are proportional to their access times. We assume that both PUs and SUs are rational and selfish. In single-PU scenario, we formulate the problem as a noncooperative game and prove that it converges to a unique Stackelberg equilibrium. We also propose an iterative algorithm to achieve the unique equilibrium point. We then extend the proposed cooperative mechanism to a multiple-PU scenario and develop a heuristic algorithm to assign proper SUs to each PU considering both performance and fairness. The simulation results show that when the competition among SUs is fierce, the performance gap between our heuristic algorithm and the optimal one is smaller than 3%. Copyright © 2013 John Wiley & Sons, Ltd.

Offloading data from cellular networks to WiFi or femtocell networks is an efficient way to alleviate the network congestion caused by rapidly increasing demands for mobile data. To this end, an Internet Service Provider (ISP) is willing to deploy WiFi/femtocell networks. With the advent of such networks, it is necessary to analyze how the ISP sets data plans to improve its profit while mobile data offloading is supported. In this paper, we develop a contract-based scheme to deal with data plan setting problem in two-tier networks. The contract offered by the ISP is a set of data plans which provide different combinations of data volume and price. We classify consumers into different types according to their percentages of data traffic offloaded to WiFi/femtocell networks. Each consumer can choose its own data plan based on its type, which is private information. Under asymmetric information scenario, we provide the necessary and sufficient conditions for the feasibility of a contract and then we derive the optimal contract which maximizes the ISP's profit. Numerical results validate the effectiveness of our scheme and indicate that the ISP can improve its profit by raising the throughput of its WiFi/femtocell networks and/or lowering its energy cost.

We study the relay-based communication schemes for cooperative spectrum sharing among multiple primary users (PUs) and multiple secondary users (SUs) with incomplete information. Inspired by the matching theory, we model the network as a matching market. In this market, each PU proposes a certain proposal representing a combination of relay power and spectrum access time to attract the SUs, while each SU maximizes its utility by selecting the most suitable PU. We derive the sufficient and necessary conditions for a stable matching in which none of the PUs or SUs would like to change its decision. We further establish a distributed matching algorithm (DMA) and a DMA with utility increasing (DMA-UI) to achieve the equilibria in partially incomplete and incomplete information scenarios, respectively. Moreover, we provide detailed discussions on the implementation of the distributed algorithms in practical networks. Simulation results show that the losses of PUs' total utilities caused by incomplete information are diminished when the number of SUs increases. Specifically, the effects of the incomplete information are reduced as the competition among SUs (PUs) is more intensive than that among PUs (SUs).

Recently, many dynamic spectrum allocation schemes based on economics are proposed to improve spectrum utilization in cognitive radio networks (CRNs). However, existing mechanisms do not take into account the economic efficiency and the spatial reusability simultaneously, which leaves room to further enhance the spectrum efficiency. In this paper, we introduce the coalition double auction for efficient spectrum allocation in CRNs, where secondary users (SUs) are partitioned into several coalitions and the spectrum reusability can be executed within each coalition. The partition formation process is not only related to the interference condition between SUs, but also the expected economic goals. Therefore, we propose a fully-economic spatial spectrum allocation mechanism by incorporating the coalition formation approach with auction theory. With the proposed scheme, the primary operator acts as an auctioneer, who performs multiple virtual auctions to form a stable partition of SUs and conducts a final auction to decide the winning SUs. Moreover, we propose a possible operation rules for the primary operator to iteratively change the partition, and prove that the virtual auctions could converge in finite time. Comprehensive theoretical analysis and simulation results are presented to show that our scheme can satisfy the crucial economic robustness properties of double auction, and outperform existing mechanisms.

In this paper, we study on the problem of opportunistic channel access without prior information about channels. We model it as the multi-armed bandit (MAB) problem. There are N independent arms. The player can choose one arm to play each time and get a reward. Switch cost is taken into consideration when player switches arm. Switch cost includes reward loss and switch delay. The concept of regret is used to measure the performance of an access policy. We prove that the regret of Lai-Robbins policy with switch cost grows with time at logarithmic order as that without switch cost, though with a much higher leading constant. Then we propose a policy referred as reducing switch with advanced play (RSAP), whose regret is shown to grow with time at logarithmic order with a much smaller leading constant.

In cognitive radio networks, secondary users (SUs) are considered as selfish spectrum users, thus how to maximize the spectrum efficiency by these selfish users becomes an endless research topic. Recent studies mostly focus on the competition analysis between SUs using economic mechanism, such as game theory and auction, but the spectrum owner can hardly increase the spectrum efficiency directly when SUs apply the distributed manner. In this paper, we consider the slotted uplink scenario where several SUs have data transmitted to secondary access point (AP) under distributed random access manner. The AP decides how to divide its spectrum which maximizes the whole throughput, then SUs select the channels which would maximize their own profit. Our results show that SUs' channel selection process leads to a Nash Equilibrium, and the AP derives the proper number of channels based on the properties of NE. Moreover, we derive a rule for AP to decide which SUs should access the spectrum and lead to the increment in the whole throughput.

How to set spectrum sensing duration is an important issue in Cognitive Radio (CR) networks, which could greatly affect energy efficiency and system throughput. Over sensing would result in insufficient transmission period, while inadequate sensing would incur false alarm and miss detection. This paper studies how to choose an optimal sensing duration to strike a balance between energy consumption and system throughput. We focus on a cooperative sensing scenario, where several secondary users form a group to guarantee more accurate sensing results. By formulating the transmission cost in terms of the energy consumption of sensing process and transmission process, we propose a comprehensive utility function. The maximization of the utility function is obtained with the constraints of sufficient protect for primary users. The existence of the optimal sensing duration is proved accordingly. Numerical results show that secondary users can achieve almost the maximum throughput with significant energy saving when utilizing optimal sensing duration.