The advent of artificial intelligence (AI)-native wireless communication is fundamentally reshaping the design paradigm of next-generation (NextG) systems, where intelligent air interfaces are expected to operate adaptively and efficiently in highly dynamic environments. Conventional orthogonal frequency division multiplexing (OFDM) systems rely heavily on pilots and the cyclic prefix (CP), resulting in significant overhead and reduced spectral efficiency. To address these limitations, we propose an adaptive end-to-end (E2E) transceiver architecture tailored for pilot-free and CP-free wireless systems. The architecture combines AI-driven constellation shaping and a neural receiver through joint training. To enhance robustness against mismatched or time-varying channel conditions, we introduce a lightweight channel adapter (CA) module, which enables rapid adaptation with minimal computational overhead by updating only the CA parameters. Additionally, we present a framework that is scalable to multiple modulation orders within a unified model, significantly reducing model storage requirements. Moreover, to tackle the high peak-to-average power ratio (PAPR) inherent to OFDM, we incorporate constrained E2E training, achieving compliance with PAPR targets without additional transmission overhead. Extensive simulations demonstrate that the proposed framework delivers superior bit error rate (BER), throughput, and resilience across diverse channel scenarios, highlighting its potential for AI-native NextG.

In vehicle-to-everything (V2X) scenarios, the high dynamic characteristics of V2X environments impose significant challenges on communication channel estimation, where the emerging integrated sensing and communication technology could serve as a vital tool for achieving accurate channel estimation. This paper leverages radar-sensed angle information to assist in communication channel estimation and proposes a deep unfolding-based radar-assisted channel estimation network (Radar-CEnet). Specifically, for the radar module, to address the challenges posed by insufficient data in imperfect arrays, we employ a model-agnostic meta-learning with a convolutional neural network (MAML-CNN) approach to achieve high-precision direction-of-arrival (DOA) estimation. Then, the angle information obtained by the radar module, as prior knowledge, is used for channel estimation. Building on this, we design a novel soft-thresholding shrinkage function and propose the Radar-CEnet algorithm to efficiently estimate the sparse channel. Finally, we rigorously prove the convergence of the Radar-CEnet algorithm and demonstrate that it achieves a lower estimation error. Experimental results show that the proposed Radar-CEnet outperforms existing traditional methods and deep learning-based approaches in channel estimation performance. At an SNR of 20dB, the proposed Radar-CEnet method reduces the NMSE from –23.75dB to –27.15dB compared to the learning-based iterative soft-thresholding method, achieving an estimation accuracy improvement of approximately 54%.

Grant-free (GF) random access has emerged as a promising solution for massive machine-type communications (mMTC). However, the non-uniform distribution of user equipment (UE) and real-world limitations on base station (BS) placement lead to random access imbalance among BSs and heavy access collisions in some cells. To this end, an unsupervised learning-based location-aware random access (ULL-RA) scheme is proposed in this paper to select the accessing BSs and channels simultaneously. Specifically, ULL-RA adopts a deep learning model named ULL-RA-Net, comprising parameter-shared feedforward neural network (FFNN) layers that enable each active UE to select a BS and an access channel to maximize the achievable rate. The model is trained in an unsupervised manner to maximize a designed differentiable objective, mapping UE locations to channel access probabilities. Notably, we propose a rate-collision loss tailored to the model architecture, which combines a collision-free channel capacity term and a sparsity-inducing collision penalty term to reduce access collisions and enhance the achievable rate. Experimental results using the Deep-MIMO dataset indicate that ULL-RA outperforms conventional GF access in both the average achievable rate and the access success rate.

A novel near-field wideband integrated sensing and communication framework is proposed to address the prohibitively high pilot overhead challenge in extremely large-scale MIMO systems. Unlike conventional approaches that rely on exhaustive two-dimensional codebook search, a unified architecture leveraging true-time-delay-based rainbow beamforming with controllable distance-dependent beam squint is proposed to extend spatial coverage. Furthermore, the inter-antenna phase ambiguity is harnessed to introduce beam split phenomena, enabling simultaneous multi-angle and multi-distance sensing within a single pilot transmission. Based on this architecture, a two-stage low-complexity sensing protocol is carried out, where distance-ring identification via beam-split-enhanced rainbow beams is performed in the first stage using sub-array structures, followed by angle refinement in the second stage. To mitigate frequency-dependent beamwidth variations, a frequency-compensated joint reconstruction algorithm based on virtual grid mapping and sparse optimization is proposed. Additionally, an echo-aided velocity estimation method exploiting intra-symbol Doppler diversity across subcarriers is developed, eliminating the need for multiple pulse transmissions. Simulation results demonstrate that: 1) complete spatial coverage is achieved with only two OFDM symbols, representing over 98% overhead reduction compared to exhaustive search methods; 2) the proposed scheme achieves superior localization accuracy with root-mean-square errors below 0.001 in normalized angle domain and 0.01 in distance-ring domain at moderate SNR; 3) communication rates are improved by 7% to 15% compared to conventional near-field beam training approaches under identical pilot budgets.

The problem of uplink transmissions in massive connectivity is commonly dealt with using schemes for grant-free random access. When a large number of devices transmit almost synchronously, the receiver may not be able to resolve the collision. This could be addressed by assigning dedicated pilots to each user, leading to a contention-free random access (CFRA), which suffers from low scalability and efficiency. This paper explores contention-based random access (CBRA) schemes for asynchronous access in massive multiple-input multiple-output (MIMO) systems. The symmetry across the accessing users with the same pilots is broken by leveraging the delay information inherent to asynchronous systems and the angle information from massive MIMO to enhance activity detection (AD) and channel estimation (CE). The problem is formulated as a sparse recovery in the delay-angle domain. The challenge is that the recovery signal exhibits both row-sparse and cluster-sparse structure, with unknown cluster sizes and locations. We address this by a cluster-extended sparse Bayesian learning (CE-SBL) algorithm that introduces a new weighted prior to capture the signal structure and extends the expectation maximization (EM) algorithm for hyperparameter estimation. Simulation results demonstrate the superiority of the proposed method in joint AD and CE.

Semantic communication is emerging as a promising paradigm for the future of wireless communication, with recent progress in deep learning-based joint source-channel coding (JSCC) achieving notable success. However, the performance of wireless communication systems is often constrained by the variability and dynamics of channel conditions, which presents significant challenges for system robustness and adaptability. In real-world applications, such as in railway environments where multiple diverse scenarios are encountered, including rural, viaduct, tunnel and hilly terrain scenarios, the variability of the channel conditions can be particularly pronounced. Existing deep learning-based JSCC approaches typically train models in fixed channel models, limiting their ability to generalize across different channel scenarios. As a result, a model trained for one specific channel type is ineffective in others, necessitating the deployment of distinct models for different channel models. To address this issue, we propose a novel channel scenario adaptive deep joint source-channel coding (SA-DJSCC) method that integrates the channel type label into both the training and inference phases. By incorporating channel scenario information into the JSCC model, our approach allows a single model to effectively generalize across multiple channel models. Experimental results show that our method improves scenario adaptability and outperforms existing approaches across different channel models and signal-to-noise ratio conditions. This work represents a key advancement towards developing more resilient and efficient semantic communication systems capable of operating in dynamic wireless environments.

Although semantic communication (SC) has shown its potential in efficiently transmitting multimodal data such as texts, speeches and images, SC for videos has focused primarily on pixel-level reconstruction. However, these SC systems may be suboptimal for downstream intelligent tasks. Moreover, SC systems without pixel-level video reconstruction present advantages by achieving higher bandwidth efficiency and real-time performance of various intelligent tasks. The difficulty in such system design lies in the extraction of task-related compact semantic representations and their accurate delivery over noisy channels. In this paper, we propose an end-to-end SC system, named VideoQA-SC for video question answering (VideoQA) tasks. Our goal is to accomplish VideoQA tasks directly based on video semantics over noisy or fading wireless channels, bypassing the need for video reconstruction at the receiver. To this end, we develop a spatiotemporal semantic encoder for effective video semantic extraction, and a learning-based bandwidth-adaptive deep joint source-channel coding (DJSCC) scheme for efficient and robust video semantic transmission. Experiments demonstrate that VideoQA-SC outperforms traditional and advanced DJSCC-based SC systems that rely on video reconstruction at the receiver under a wide range of channel conditions and bandwidth constraints. In particular, when the signal-to-noise ratio is low, VideoQA-SC can improve the answer accuracy by 5.17% while saving almost 99.5% of the bandwidth at the same time, compared with the advanced DJSCC-based SC system. Our results show the great potential of SC system design for video applications.

Smart railways integrate advanced information technologies into railway operating systems to improve efficiency and reliability. Although the development of the fifth generation (5G) has enhanced railway services, future smart railways require ultra-high speeds, ultra-low latency, ultra-high security, full coverage, and ultra-high positioning accuracy, which 5G cannot fully meet. Therefore, the sixth generation (6G) is envisioned to provide green and efficient all-day operations, strong information security, fully automatic driving, and low-cost intelligent maintenance. To achieve these requirements, we propose an integrated network architecture leveraging communications, computing, edge intelligence, and caching in railway systems. We have conducted in-depth investigations on key enabling technologies for reliable transmissions and wireless coverage. For high-speed mobile scenarios, we propose an artificial intelligence (AI)-enabled cross-domain channel modeling and orthogonal time–frequency space–time spread multiple access mechanism to alleviate the conflict between limited spectrum availability and massive user access. The roles of blockchain, edge intelligence, and privacy technologies in endogenously secure rail communications are also evaluated. We further explore the application of emerging paradigms such as integrated sensing and communications (SACs), AI-assisted Internet of Things (IoT), semantic communications (SCs), and digital twin (DT) networks for railway maintenance, monitoring, prediction, and accident warning. Finally, possible future research and development directions are discussed. © 2026 IEEE

CANDECOMP/PARAFAC (CP) decomposition is the mostly used model to formulate the received tensor signal in a massive MIMO system, as the receiver generally sums the components from different paths or users. To achieve accurate and low-latency channel estimation, good and fast CP decomposition (CPD) algorithms are desired. The CP alternating least squares (CPALS) is the workhorse algorithm for calculating the CPD. However, its performance depends on the initializations, and good starting values can lead to more efficient solutions. Existing initialization strategies are decoupled from the CPALS and are not necessarily favorable for solving the CPD. This paper proposes a deep-learning-aided CPALS (DL-CPALS) method that uses a deep neural network (DNN) to generate favorable initializations. The proposed DL-CPALS integrates the DNN and CPALS to a model-based deep learning paradigm, where it trains the DNN to generate an initialization that facilitates fast and accurate CPD. Moreover, benefiting from the CP low-rankness, the proposed method is trained using noisy data and does not require paired clean data. The proposed DL-CPALS is applied to millimeter wave MIMO-OFDM channel estimation. Experimental results demonstrate the significant improvements of the proposed method in terms of both speed and accuracy for CPD and channel estimation.

The increasing quality of service requirements and demand for satellite services necessitate higher data rates, spectral efficiency, and stability in satellite communication networks. Therefore, this paper investigates the non-orthogonal multiple access (NOMA) assisted satellite communication networks, where multiple users are uniformly distributed over a spherical hat according to the homogeneous Poisson point process (HPPP). Based on the characteristics of HPPP, we analyze the distance distributions of users. To evaluate the performance of the proposed networks, we first derive the closed-form expressions and approximated expressions of the outage probability (OP) for paired NOMA users. To obtain more insights into the proposed networks, the ergodic rate and diversity orders for paired NOMA users are also derived. Spectral efficiency is derived for NOMA and orthogonal multiple access (OMA) assisted satellite communication networks. Our analytical results demonstrate that the diversity order of the proposed networks is all one. Numerical results confirm that: 1) compared to OMA, the proposed NOMA-assisted satellite network demonstrates superior outage performance and spectral efficiency, especially in the higher power regimes; 2) fading factors have negligible effects on OP and spectral efficiency; and 3) under certain target rates for both near and far users, the outage performance of far users in NOMA-assisted satellite communication networks is superior to that of near users.

With the arrival of 6G, the Internet of Things (IoT) traffic is becoming more and more complex and diverse. To meet the diverse service requirements of IoT devices, massive machine-type communications (mMTC) becomes a typical scenario, and more recently, grant-free random access (GF-RA) presents a promising direction due to its low signaling overhead. However, existing GF-RA research primarily focuses on improving the accuracy of user detection and data recovery, without considering the heterogeneity of traffic. In this paper, we investigate a non-orthogonal GF-RA scenario where two distinct types of traffic coexist: event-triggered traffic with alarm devices (ADs), and status update traffic with monitor devices (MDs). The goal is to simultaneously achieve high detection success rates for ADs and high information timeliness for MDs. First, we analyze the age-based random access scheme and optimize the access parameters to minimize the average age of information (AoI) of MDs. Then, we design an age-based prior information aided autoencoder (A-PIAAE) to jointly detect active devices, together with learned pilots used in GF-RA to reduce interference between non-orthogonal pilots. In the decoder, an Age-based Learned Iterative Shrinkage Thresholding Algorithm (LISTA-AGE) utilizing the AoI of MDs as the prior information is proposed to enhance active user detection. Theoretical analysis is provided to demonstrate the proposed A-PIAAE has better convergence performance. Experiments demonstrate the advantage of the proposed method in reducing the average AoI of MDs and improving the successful detection rate of ADs.

This paper investigates the hybrid beamforming design for the reconfigurable intelligent surface (RIS)-aided full-duplex (FD) cell-free networks, where the access points (APs) connected to the central processing unit serve multiple users cooperatively. The weighted sum rate of uplink and downlink transmissions is maximized by jointly optimizing the digital and analog beamformers, the phase-shift coefficients of RISs and the uplink transmit power while satisfying the power budget constraints of the APs and users. For solving the problem, the objective function is reformulated using Lagrangian dual transform and fractional programming. Then, to tackle the variables coupling, the problem is divided into five subproblems that are solved iteratively via a proposed block coordinate descent (BCD)-based algorithm. To handle the unit-modulus constraint on analog beamformers and phase-shift coefficients, a monotonic fast proximal gradient (mFPG)-based method is proposed under the alternating direction method of multipliers (ADMM) framework. Numerical results demonstrate the effectiveness and efficiency of the proposed algorithm compared to three baseline algorithms. The impacts of the numbers of radio frequency chains and reflecting elements on the uplink and downlink transmissions are presented, respectively. The performance comparison between the FD and half-duplex schemes is provided.

This paper investigates the deep learning based approaches for simultaneous wireless information and power transfer (SWIPT). The quality-of-service (QoS) constrained sumrate maximization problems are, respectively, formulated for power-splitting (PS) receivers and time-switching (TS) receivers and solved by a unified graph neural network (GNN) based model termed SWIPT net (SWIPTNet). To improve the performance of SWIPTNet, we first propose a single-type output method to reduce the learning complexity and facilitate the satisfaction of QoS constraints, and then, utilize the Laplace transform to enhance input features with the structural information. Besides, we adopt the multi-head attention and layer connection to enhance feature extracting. Furthermore, we present the implementation of transfer learning to the SWIPTNet between PS and TS receivers. Ablation studies show the effectiveness of key components in the SWIPTNet. Numerical results also demonstrate the capability of SWIPTNet in achieving nearoptimal performance with millisecond-level inference speed which is much faster than the traditional optimization algorithms. We also show the effectiveness of transfer learning via fast convergence and expressive capability improvement.

Diffusion models have been extensively utilized in AI-generated content (AIGC) in recent years, thanks to the superior generation capabilities. Combining with semantic communications, diffusion models are used for tasks such as denoising, data reconstruction, and content generation. However, existing diffusion-based generative models do not consider the stringent bandwidth limitation, which limits its application in wireless communication. This paper introduces a diffusion-driven semantic communication framework with advanced VAE-based compression for bandwidth-constrained generative model. Our designed architecture utilizes the diffusion model, where the signal transmission process through the wireless channel acts as the forward process in diffusion. To reduce bandwidth requirements, we incorporate a downsampling module and a paired upsampling module based on a variational auto-encoder with reparameterization at the receiver to ensure that the recovered features conform to the Gaussian distribution. Furthermore, we derive the loss function for our proposed system and evaluate its performance through comprehensive experiments. Our experimental results demonstrate significant improvements in pixel-level metrics such as peak signal to noise ratio (PSNR) and semantic metrics like learned perceptual image patch similarity (LPIPS). These enhancements are more profound regarding the compression rates and SNR compared to deep joint source-channel coding (DJSCC).

This paper proposes a novel deep learning enabled beamforming design for ultra-dense wireless networks by integrating prior knowledge and graph neural network (GNN), termed model-based GNN. An energy efficiency (EE) maximization problem is first subject to the power budget and quality of service (QoS) requirements, and then reformulated based on the minimum mean square error scheme and the hybrid zero-forcing and maximum ratio transmission scheme. The model-based GNN is designed to realize the mapping from channel state information to beamforming vectors to address the reformulated problems. Particularly, the multi-head attention mechanism and the residual connection are adopted to enhance the feature extracting, and a scheme selection module is designed to improve the adaptability to channel conditions. The unsupervised learning is adopted, and a various-input training strategy is proposed to enhance the stability of the model-based GNN. Numerical results demonstrate that the proposed model-based GNN can realize a millisecond-level inference with limited performance loss, the scalability to different numbers of users and the adaptability to various channel conditions and QoS requirements in ultra-dense wireless networks.

In this paper, we propose two deep joint source and channel coding (DJSCC) structures with attention modules for the multi-input multi-output (MIMO) channel, including a serial structure and a parallel structure. With singular value decomposition (SVD)-based precoding scheme, the MIMO channel can be decomposed into various sub-channels, and the feature outputs will experience sub-channels with different channel qualities. In the serial structure, one single network is used at both the transmitter and the receiver to jointly process data streams of all MIMO subchannels, while data steams of different MIMO sub-channels are processed independently via multiple sub-networks in the parallel structure. The attention modules in both serial and parallel architectures enable the system to adapt to varying channel qualities and adjust the quantity of information outputs with the channel qualities. Experimental results demonstrate the proposed DJSCC structures have improved image transmission performance, and reveal the phenomenon via non-parameter entropy estimation that the learned DJSCC transceivers tend to transmit more information over better sub-channels.

Scalability is a major concern in implementing deep learning (DL) based methods in wireless communication systems. Given various channel reconstruction tasks, applying one DL model for one specific task is costly in both model training and model storage. In this paper, we propose a novel unsupervised deep plug-and-play prior method for three channel reconstruction tasks in the downlink of massive multiple-input multiple-output (MIMO) systems, including channel estimation, antenna extrapolation and channel state information (CSI) feedback. The proposed method corresponding to these three channel reconstruction tasks employs a common DL model, which greatly reduces the overhead of model training and storage. Unlike general multi-task learning, the DL model of the proposed method does not require further fine-tuning for specific channel reconstruction tasks. Extensive experiments are conducted on the DeepMIMO dataset to demonstrate the convergence, performance, and storage overhead of the proposed method for the three channel reconstruction tasks.

Grant-free non-orthogonal multiple access has been regarded as a viable approach to accommodate access for a massive number of machine-type devices with small data packets. The sporadic activation of the devices creates a multiuser setup where it is suitable to use compressed sensing in order to detect the active devices and decode their data. We consider asynchronous access of machine-type devices that send data packets of different frame sizes, leading todata length diversity. We address the composite problem of activity detection, channel estimation, and data recovery by posing it as a structured sparse recovery, having three-level sparsity caused by sporadic activity, symbol delay, and data length diversity. We approach the problem through approximate message passing with a backward propagation algorithm (AMP-BP), tailored to exploit the sparsity, and in particular the data length diversity. Moreover, we unfold the proposed AMP-BP into a network, termed learned AMP-BP (LAMP-BP), which enhances detection performance. The results show that the proposed LAMP-BP outperforms existing methods in activity detection and data recovery accuracy.

To reduce multiuser interference and maximize the spectrum efficiency in orthogonal frequency division duplexing massive multiple-input multiple-output (MIMO) systems, the downlink channel state information (CSI) estimated at the user equipment (UE) is required at the base station (BS). This paper presents a novel method for massive MIMO CSI feedback via a one-sided one-for-all deep learning framework. The CSI is compressed via linear projections at the UE, and is recovered at the BS using deep learning (DL) with plug-and-play priors (PPP). Instead of using handcrafted regularizers for the wireless channel responses, the proposed approach, namely CSI-PPPNet, exploits a DL based denoisor in place of the proximal operator of the prior in an alternating optimization scheme. In this way, a DL model trained once for denoising can be repurposed for CSI recovery tasks with arbitrary compression ratio. The one-sided one-for-all framework reduces model storage space, relieves the burden of joint model training and model delivery, and could be applied at UEs with limited device memories and computation power. Extensive experiments over the open indoor and urban macro scenarios show the effectiveness and advantages of the proposed method.

Massive machine-type communication (mMTC) is an important scenario for 5G and future 6G networks, as it can provide massive connectivity for internet of things (IoT) devices. However, the large number of supported devices raises challenges to random access with limited spectrum resources. In this paper, we propose a dictionary learning based method for active user detection (AUD) in massive MIMO systems, which leverages the potential spatial channel characteristics of users. Our approach separates users into clusters and reuses the same pilot pool among different clusters, which greatly saves the pilot resource. To resolve collisions caused by the reuse of pilots, we propose a cluster-specific dictionary to differentiate multiple active users of different clusters. Numerical experiments demonstrate the improved performance of the proposed AUD algorithm in comparison to the existing methods.

Due to the inherent high-mobility features in the Vehicles-to-Everything (V2X) scenarios, accurate channel estimation is essential to ensure the quality of communication services. Recently, multiple-input multiple-output (MIMO) radar has shown the potential to aid channel estimation. In this paper, we consider the MIMO radar aided V2X communication systems and propose a prior information aided deep learning method for channel estimation. Specifically, we use the MIMO radar to measure the angle information of moving vehicles. Based on the estimated angles, we obtain the non-zero position information of sparse angle-frequency channel. Then, by formulating the channel estimation as solving a group row sparse recovery problem, we propose a new shrinkage function and derive a priori assisted deep unfolding method. Experimental results show that the proposed method achieves the highest channel estimation accuracy compared with existing compressive sensing algorithms and deep-learning-based baseline methods.

Powered by deep learning, semantic communication is an intelligent communication paradigm, aiming to transmit useful information in the semantic domain. In most existing work, robust semantic features can be learned against wireless channel degradation, and directly transmitted in an analog fashion. However, analog semantic communication raises various challenges to the existing system from hardware/protocol to encryption issues. In this paper, we propose a novel non-linear quantization module to efficiently quantify semantic features. A sparse scaling vector is further incorporated to reduce the dimension of transmitted semantic features. Experimental results demonstrate that the proposed nonlinear quantization achieves better performance than the linear quantization method, and the performance of the digital system achieves better performance.

The increased throughput brought by MIMO technology relies on the knowledge of channel state information (CSI) acquired in the base station (BS). To make the CSI feedback overhead affordable for the evolution of MIMO technology (e.g., massive MIMO and ultra-massive MIMO), deep learning (DL) is introduced to deal with the CSI compression task. In traditional communication systems, the compressed CSI bits is treated equally and expected to be transmitted accurately over the noisy channel. While the errors occur due to the limited bandwidth or low signal-to-noise ratios (SNRs), the reconstruction performance of the CSI degrades drastically. As a branch of semantic communications, deep joint source-channel coding (DJSCC) scheme performs better than the separate source-channel coding (SSCC) scheme—the cornerstone of traditional communication systems—in the limited bandwidth and low SNRs. In this paper, we propose a DJSCC based framework for the CSI feedback task. In particular, the proposed method can simultaneously learn from the CSI source and the wireless channel. Instead of truncating CSI via Fourier transform in the delay domain in existing methods, we apply non-linear transform networks to compress the CSI. Furthermore, we adopt an SNR adaption mechanism to deal with wireless channel variations. The extensive experiments demonstrate the validity, adaptability, and generality of the proposed framework.

Wireless random access (RA) faces huge challenges under the explosive growth of the Internet of Things devices to be connected to the base station. This leads to inevitable RA collisions. In this paper, we consider the RA in massive multiple-input multiple-output (MIMO) systems, and propose an activity detection and channel estimation algorithm based on dictionary learning. By exploiting the sporadic feature of massive connected devices that a small fraction of them being active, we exploit compressive sensing for simultaneous channel estimation and activity detection. More importantly, the proposed algorithm utilizes dictionary learning to abstract the sparse characteristics of the spatial channel in the massive MIMO system. A dictionary is learned by using historical channel information of each cell, and thus is appropriate for the specific cell. Simulation results demonstrate the superiority of the proposed method compared with existing methods.

In the next-generation wireless communication systems, learning-based dynamic spectrum access strategy at the medium access control layer and physical layer shows its powerful capability of achieving optimal resources allocation, and it has become a hot research topic for the harmonious coexistence of heterogeneous wireless networks. In this paper, we propose a multiple access control method to achieve high network throughput by combining deep reinforcement learning and memory module. In specific, we introduce the bidirectional gated recurrent unit (Bi-GRU) in deep Q-learning (DQL) to utilize the information of varying environment observation at each time-step. Furthermore, we apply the method in a freeway scenario with real-world datasets, where the DQL node contends the same wireless channel with other nodes. Evaluated results demonstrate that the proposed approach learns an optimal policy without using complex mechanism or prior. Moreover, we consider realistic cases involving saturated or unsaturated uplink traffic flows of nodes on a freeway segment, and the on-line training strategies of the DQL node near the roadside facilities. The experimental results show that the proposed scheme leads to the highest throughput in all cases compared with the competing approaches.

Benefiting from the advantage of joint source-channel coding in the finite block length regime and the advancement of AI technologies, deep joint source channel coding (DJSCC) has been extensively investigated for various sources (e.g., text source, image source, and video source) and achieved remarkable performance in limited bandwidth and low signal-to-noise ratios (SNRs). However, the performance gain brought by DJSCC methods is all observed via simulations in literature, where synchronization, channel estimation and the power amplifier are assumed to be perfect. In this paper, we design a software-defined radio (SDR) based platform to validate the DJSCC method for image transmission in real wireless scenarios.

In massive machine-type communications (mMTC), the conflict between millions of potential access devices and limited channel freedom leads to a sharp decrease in spectrum efficiency. The nature of sporadic activity in mMTC provides a solution to enhance spectrum efficiency by employing compressive sensing (CS) to perform multiuser detection (MUD). However, CS-MUD suffers from high computation complexity and fails to meet the strict latency requirement in some critical applications. To address this problem, in this paper, we propose a novel deep learning (DL) based framework for grant-free non-orthogonal multiple access (GF-NOMA), where we utilize the information distilled from the initial data recovery phase to further enhance channel estimation, which in turn improves data recovery performance. Besides, we design an interpretable and structured Model-driven Prior Information Aided Network (M-PIAN) and provide theoretical analysis that demonstrates the proposed M-PIAN can converge faster and support more users. Experiments show that the proposed method outperforms existing CS algorithms and DL methods in both computation complexity and reconstruction accuracy.

Nowadays, nonnegative matrix factorization (NMF)-based methods have been widely applied to blind spectral unmixing. Introducing proper regularizers to NMF is crucial for mathematically constraining the solutions and physically exploiting spectral and spatial properties of images. Generally, properly handcrafted regularizers and solving the associated complex optimization problem are nontrivial tasks. In our work, we propose an NMF-based unmixing framework which jointly uses a learned regularizer from data and a handcrafted regularizer. To be specific, we plug learned priors of abundances where the associated subproblem can be addressed using various image denoisers, and we consider an$\ell _{2,1}$-norm as an example to illustrate the way of integrating handcrafted regularizers. The proposed framework is flexible and extendable. Both synthetic data and real airborne data are conducted to confirm the effectiveness of our method.

We consider the problem of increasing the resolution of a hyperspectral image (HSI) with the aid of a high-resolution RGB image of the same scene. The current state-of-the-art algorithms for this task are based on convolutional neural networks (CNNs) and generally assume that the relation between the RGB image and the HSIs remains constant during training and testing. In particular, their performance quickly degrades if we use different color spaces, e.g., CIEXYZ or CIERGB during these stages. In this paper, we propose a method that addresses this problem. Specifically, our method requires no RGB images during training, but still can leverage an RGB image during testing to improve the performance of super-resolution. Furthermore, the method works even if the relation between the RGB and HSI images, captured by the camera spectral response (CSR), is not known precisely. Our experiments demonstrate that the proposed method not only outperforms state-of-the-art methods for joint RGB-HSI super-resolution, but also works for various types of color images.

Recent research on joint source channel coding (JSCC) for wireless communications has achieved great success owing to the employment of deep learning (DL). However, the existing work on DL based JSCC usually trains the designed network to operate under a specific signal-to-noise ratio (SNR) regime, without taking into account that the SNR level during the deployment stage may differ from that during the training stage. A number of networks are required to cover the scenario with a broad range of SNRs, which is computational inefficiency (in the training stage) and requires large storage. To overcome these drawbacks our paper proposes a novel method called Attention DL based JSCC (ADJSCC) that can successfully operate with different SNR levels during transmission. This design is inspired by the resource assignment strategy in traditional JSCC, which dynamically adjusts the compression ratio in source coding and the channel coding rate according to the channel SNR. This is achieved by resorting to attention mechanisms because these are able to allocate computing resources to more critical tasks. Instead of applying the resource allocation strategy in traditional JSCC, the ADJSCC uses the channel-wise soft attention to scaling features according to SNR conditions. We compare the ADJSCC method with the state-of-the-art DL based JSCC method through extensive experiments to demonstrate its adaptability, robustness and versatility. Compared with the existing methods, the proposed method takes less storage and is more robust in the presence of channel mismatch.

Spectral unmixing is a widely used technique in hyperspectral image processing and analysis. It aims to separate mixed pixels into the component materials and their corresponding abundances. Early solutions to spectral unmixing are performed independently on each pixel. Nowadays, investigating proper priors into the unmixing problem has been popular as it can significantly enhance the unmixing performance. However, it is nontrivial to handcraft a powerful regularizer, and complex regularizers may introduce extra difficulties in solving optimization problems in which they are involved. To address this issue, we present a plug-and-play (PnP) priors framework for hyperspectral unmixing. More specifically, we use the alternating direction method of multipliers (ADMM) to decompose the optimization problem into two iterative subproblems. One is a regular optimization problem depending on the forward model, and the other is a proximity operator related to the prior model and can be regarded as an image denoising problem. Our framework is flexible and extendable which allows a wide range of denoisers to replace prior models and avoids handcrafting regularizers. Experiments conducted on both synthetic data and real airborne data illustrate the superiority of the proposed strategy compared with other state-of-the-art hyperspectral unmixing methods.

Video surveillance is gaining increasing popularity to assist in railway intrusion detection in recent years. However, efficient and accurate intrusion detection remains a challenging issue due to: (a) limited sample number: only small sample size (or portion) of intrusive video frames is available; (b) high inter-scene dissimilarity: various railway track area scenes are captured by cameras installed in different landforms; (c) high intra-scene similarity: the video frames captured by an individual camera share a same background. In this paper, an efficient few-shot learning solution is developed to address the above issues. In particular, an enhanced model-agnostic meta-learner is trained using both the original video frames and segmented masks of track area extracted from the video. Moreover, theoretical analysis and engineering solutions are provided to cope with the highly similar video frames in the meta-model training phase. The proposed method is tested on realistic railway video dataset. Numerical results show that the enhanced meta-learner successfully adapts unseen scene with only few newly collected video frame samples, and its intrusion detection accuracy outperforms that of the standard randomly initialised supervised learning.

To support massive connectivity for sporadically active devices is a challenging task, as the randomness of the channel and the large number of users lead to enormous increase of communication overhead. Different to the existing methods that differentiate users in resources including time, frequency and code, we propose a new joint activity detection and channel estimation framework for massive multiple-input multiple-output (MIMO) systems, where angular domain information of active users is exploited to enhance activity detection and channel estimation. By exploiting the sporadic activity of users and the angular spread of the wireless signals, the activity detection and channel estimation is formulated as a compressive sensing problem with multiple measurement vectors, which has a simultaneously row-sparse and clustered sparse structure. The sizes and positions of the nonzero clusters are arbitrary, which brings new challenges for algorithm derivation. To this end, we develop new algorithms based on sparse Bayesian learning, where novel hyper-priors are proposed to capture the structural signal characteristics, and appropriate approximations are employed to facilitate algorithm derivations. Numerical experiments demonstrate the improved activity detection and channel estimation performance of the proposed approach in comparison to the existing methods.

Massive machine-type communications (mMTC) are poised to provide ubiquitous connectivity for billions of Internet-of-Things (IoT) devices. However, the required low-latency massive access necessitates a paradigm shift in the design of random access schemes, which invokes a need of efficient joint activity and data detection (JADD) algorithms. By exploiting the feature of sporadic traffic in massive access, a beacon-aided slotted grant-free massive access solution is proposed. Specifically, we spread the uplink access signals in multiple subcarriers with pre-equalization processing and formulate the JADD as a multiple measurement vectors (MMV) compressive sensing problem. Moreover, to leverage the structured sparsity of uplink massive access signals among multiple time slots, we develop two computationally efficient detection algorithms, which are termed as orthogonal approximate message passing (OAMP)-MMV algorithm with simplified structure learning (SSL) and accurate structure learning (ASL). To achieve accurate detection, the expectation maximization algorithm is exploited for learning the sparsity ratio and the noise variance. To further improve the detection performance, channel coding is applied and successive interference cancellation (SIC)-based OAMP-MMV-SSL and OAMP-MMV-ASL algorithms are developed, where the likelihood ratio obtained in the soft-decision can be exploited for refining the activity identification. Finally, the state evolution of the proposed OAMP-MMV-SSL and OAMP-MMV-ASL algorithms is derived to predict the performance theoretically. Simulation results verify that the proposed solutions outperform various state-of-the-art baseline schemes, enabling low-latency random access and high-reliable massive IoT connectivity with overloading.

In massive machine-type communications (mMTC), the conflict between millions of potential access devices and limited channel freedom leads to a sharp decrease in spectral efficiency. The sparse nature of mMTC provides a solution by using compressive sensing (CS) to perform multiuser detection (MUD) but suffers conflict between the high computation complexity and low latency requirements. In this paper, we propose a novel Dual-network for joint channel estimation and data recovery. The proposed Dual-Net utilizes the sparse consistency between the channel vector and data matrix of all users. Experimental results show that the proposed Dual-Net outperforms existing CS algorithms and general neural networks in computation complexity and accuracy, which means reduced access delay and more supported devices.

Linear inverse problems appear in many applications, where different algorithms are typically employed to solve each inverse problem. Nowadays, the rapid development of deep learning (DL) provides a fresh perspective for solving the linear inverse problem: a number of well-designed network architectures results in state-of-the-art performance in many applications. In this overview paper, we present the combination of the DL and the Plug-and-Play priors (PPP) framework, showcasing how it allows solving various inverse problems by leveraging the impressive capabilities of existing DL based denoising algorithms. Open challenges and potential future directions along this line of research are also discussed.

Hyperspectral (HS) images contain detailed spectral information that has proven crucial in applications like remote sensing, surveillance, and astronomy. However, because of hardware limitations of HS cameras, the captured images have low spatial resolution. To improve them, the low-resolution hyperspectral images are fused with conventional high-resolution RGB images via a technique known as fusion based HS image super-resolution. Currently, the best performance in this task is achieved by deep learning (DL) methods. Such methods, however, cannot guarantee that the input measurements are satisfied in the recovered image, since the learned parameters by the network are applied to every test image. Conversely, model-based algorithms can typically guarantee such measurement consistency. Inspired by these observations, we propose a framework that integrates learning and model based methods. Experimental results show that our method produces images of superior spatial and spectral resolution compared to the current leading methods, whether model-or DL-based.

Sparse signal recovery problems from noisy linear measurements appear in many areas of wireless communications. In recent years, deep learning (DL) based approaches have attracted interests of researchers to solve the sparse linear inverse problem by unfolding iterative algorithms as neural networks. Typically, research concerning DL assume a fixed number of network layers. However, it ignores a key character in traditional iterative algorithms, where the number of iterations required for convergence changes with varying sparsity levels. By investigating on the projected gradient descent, we unveil the drawbacks of the existing DL methods with fixed depth. Then we propose an end-to-end trainable DL architecture, which involves an extra halting score at each layer. Therefore, the proposed method learns how many layers to execute to emit an output, and the network depth is dynamically adjusted for each task in the inference phase. We conduct experiments using both synthetic data and applications including random access in massive MTC and massive MIMO channel estimation, and the results demonstrate the improved efficiency for the proposed approach.

The variable span linear filters (VSLFs) constitute a unified framework of conventional subspace and linear filtering techniques for noise reduction. The construction of VSLFs, however, relies on the generalized eigendecomposition (GEVD) methods, which are computationally expensive. This in turn stymies the employment of such filters in practical online processing problems. To address this issue, we first propose in this paper a fixed-point iteration technique to extract the generalized eigenvectors. It is based on maximizing the pre-whitened generalized Rayleigh quotient (GRQ). We then integrate this technique with online statistic estimation to construct VSLFs. Our proposed method is computationally efficient and can also harness parallel architectures. To show its effectiveness, we consider a speech enhancement application and compare the results with those of several existing methods.

The B5G is expected to support multiple massive machine-type communication (mMTC) services. However, limited resources impede the access of a large number of machine-type devices, and existing access frameworks are not efficient for transmitting small data packets. In this paper, we propose a compressive sensing (CS) approach for contention based random access scheme to support massive connections (≥ 106devices) within a certain time-frequency resource. Different from the four-step contention based access scheme in LTE, we adopt a one-step access scheme to save the energy and spectrum resources. It breaks the bottleneck of existing CS multiuser detection methods which have poor scalability (e.g., assuming ≤ 104devices in relevant literatures) due to the high computational complexity of CS. The improved performance of the newly proposed method is observed in our experimental results.

Exploiting the latent low-rankness of tensors is crucial in tensor denoising. Classically, many methods use the Tucker model to find the low-rank structure of a tensor. Recently, the tensor train (TT) model has drawn wide attention owing to its powerful representation ability, and well-balanced matricization scheme for a tensor, and it has been successfully applied to various problems in signal processing, and machine learning applications. In this letter, we propose a tensor denoising method using the TT singular value decomposition, and information criteria, where we leverage the minimum description length to automatically estimate the TT rank. Furthermore, we establish the relationship between Tucker decomposition, and TT decomposition. In specific, the low Tucker rank of a tensor is the sufficient but unnecessary condition to the low TT rank. It unveils in theory the potential advantages of the TT model in characterizing the latent low-rankness of tensor. Denoising experiments on both synthetic data, and real HSI dataset demonstrate its superiority against Tucker-based methods.

Caching is a promising technique that can effectively reduce peak traffic by pushing popular content items proactively to the users during off- peak hours. Two key performance metrics for proactive caching are the energy efficiency and hit ratio. In this paper, we are interested in balancing the hit ratio and energy efficiency in social-aware proactive caching. More specifically, we present a social behavior driven pushing and caching policy that is capable of maximizing the hit ratio while satisfying the power constraint. For two notable schemes, namely uncoded caching and coded caching, we formulate two optimization problems that give the optimal tradeoff between the energy efficiency and caching hit ratio. Furthermore, the optimal solutions present the energy efficient joint pushing and buffer update polices. Our simulation results show that the social-aware pushing and update policies may bring a significant hit ratio gain when the buffer size is far less than the number of items.

As one of the three major scenarios of fifth-generation (5G) communication system, massive machine type communications (mMTC) raises more challenges for development of new radio access technology. Unlike the human-to-human communications, the mMTC typically exhibit features such as massive number of devices, small-sized packets and sporadic transmission, which requires novel solutions to satisfy these features. In this paper, we propose a novel grant-free compressive sensing based solution with system level design to jointly conduct the active devices detection, channel estimation and data recovery without knowing the number of active devices. Specifically, in proposed solution, we integrate the sparsity information of channel coefficients and data symbols and exploit the constellation information of modulation and the data diversity of different devices to enhance the performance of receiver. Simulation results demonstrate that the proposed solutions have improve the performance of active device detection, channel estimation, data recovery and the throughput of the system.

As a 3-order tensor, a multi-spectral image (MSI) has dozens of spectral bands, which can deliver more faithful representation for real scenes. However, MSIs are often corrupted by noise in the sensing process, which deteriorates the performance of higher-level classification and recognition tasks. In this paper, we propose a novel tensor dictionaries learning method for MSI denoising, where two shared dictionaries are learned from MSI groups of similar blocks in the spatial domain and the spectral domain, respectively. In addition, we enforce a low rank structure for the representations of MSI groups under the learned dictionaries, which captures the latent structure in MSIs. Our experiments demonstrate that the proposed method achieves the best performance in comparison with the state-of-the-art methods.

The surge of demand of the Internet of Things (IoTs) poses more requirements for cellular communication systems, such as higher levels of reliability, latency and supported number of devices. A common scenario of IoT is massive machinetype communications (mMTC). The characteristics of mMTC, i.e. the massive number of sensors, small packets and sporadic transmission, require novel methods to detect the active devices, estimate their channel state information and recovery their data with high accuracy and low latency. In this paper, we propose a new compressive sensing based method to jointly conduct active device detection, channel estimation and data recovery. As a high level description of the proposed method, we improve the performance of active device detection and channel estimation by using side information brought from the data recovery, and improve the performance of data recovery in successive interference cancellation by using the side information brought from the active device detection. Simulation results demonstrate that the proposed method has improved data recovery performance.

Massive machine-type communication (MTC) with sporadically transmitted small packets and low data rate requires new designs on the PHY and MAC layer with light transmission overhead. Compressive sensing based multiuser detection (CS-MUD) is designed to detect active users through random access with low overhead by exploiting sparsity, i.e., the nature of sporadic transmissions in MTC. However, the high computational complexity of conventional sparse reconstruction algorithms prohibits the implementation of CS-MUD in real communication systems. To overcome this drawback, in this paper, we propose a fast Deep learning based approach for CS-MUD in massive MTC systems. In particular, a novel block restrictive activation nonlinear unit, is proposed to capture the block sparse structure in wide-band wireless communication systems (or multi-antenna systems). Our simulation results show that the proposed approach outperforms various existing algorithms for CS-MUD and allows for ten-fold decrease of the computing time.

Scene recognition remains one of the most challenging problems in image understanding. With the help of fully connected layers (FCL) and rectified linear units (ReLu), deep networks can extract the moderately sparse and discriminative feature representation required for scene recognition. However, few methods consider exploiting a sparsity model for learning the feature representation in order to provide enhanced discriminative capability. In this paper, we replace the conventional FCL and ReLu with a new dictionary learning layer, that is composed of a finite number of recurrent units to simultaneously enhance the sparse representation and discriminative abilities of features via the determination of optimal dictionaries. In addition, with the help of the structure of the dictionary, we propose a new label discriminative regressor to boost the discrimination ability. We also propose new constraints to prevent overfitting by incorporating the advantage of the Mahalanobis and Euclidean distances to balance the recognition accuracy and generalization performance. Our proposed approach is evaluated using various scene datasets and shows superior performance to many state-of-the-art approaches.

Various dictionary learning methods have gained tremendous success for signal classification. However, traditional dictionary learning methods for classification assume there is no outlier in the training data, which may not be the case in practical applications. In this paper, we propose a new discriminative dictionary learning framework for classification, which simultaneously learns a discriminative dictionary and detects outliers in the data. We formulate the dictionary learning framework into an optimization problem with designed regularizers to promote both the discrimination and outlier-detection capability. An efficient and effective iterative algorithm based on the alternating direction method of multipliers (ADMM) is provided to solve the proposed optimization problem. We demonstrate the superior performance of the proposed approach in comparison with state-of-the-art methods on some image classification tasks.

An effective representation model plays an important role in the visual tracking, as it relates to how the most meaningful information are recognized and understood in the dictionary space. However, it is difficult to know the structure and the weights of tracking objects in advance. In addition, how to balance the adaption and robustness in tracking algorithms remains a nontrivial problem. In this paper, we propose a robust visual tracker based on adaptive structure-enhanced regularizations, and achieve a sequential Monte Carlo searching via simplified particle filters. Specifically, multiple atomic norms are incorporated in the cost function in the target dictionary space, and their weights are updated adaptively during the detection step between each frame. Sparse and low-rank structures as well as other atomic norms enhance the robustness by capturing various features meanwhile ruling out outliers, and the velocity of moving objects are considered accordingly in the probabilistic distribution of particles. Moreover, the algorithm has been accelerated by adopting prefilters as classifiers for target particles using pixel variances in colours and intensities, which ensures a real-time tracking in practice. On challenging tracking datasets, the proposed approach show advantages in tracking fast-moving objects and favorable performance against other 10 state-of-the-art visual trackers.

Low rank tensor completion, which solves a linear inverse problem with the principle of parsimony, is a powerful technique used in many application domains in computer vision and pattern recognition. As a surrogate function of the matrix rank that is non-convex and discontinuous, the nuclear norm is often used instead to derive efficient algorithms for recovering missing information in matrices and higher order tensors. However, the nuclear norm is a loose approximation of the matrix rank, and what is more, the tensor nuclear norm is not guaranteed to be the tightest convex envelope of a multilinear rank. Alternative algorithms either require specifying/tuning several parameters (e.g., the tensor rank), and/or have a performance far from reaching the theoretical limit where the number of observed elements equals the degree of freedom in the unknown low-rank tensor. In this paper, we propose a pseudo-Bayesian approach, where a Bayesian-inspired cost function is adjusted using appropriate approximations that lead to desirable attributes including concavity and symmetry. Although deviating from the original Bayesian model, the resulting non-convex cost function is proved to have the ability to recover the true tensor with a low multilinear rank. A computational efficient algorithm is derived to solve the resulting non-convex optimization problem. We demonstrate the superior performance of the proposed algorithm in comparison with state-of-the-art alternatives by conducting extensive experiments on both synthetic data and several visual data recovery tasks.

Many simultaneous sparse estimation approaches focus on joint estimation of multiple sparse vectors with a common support from given linear observations, which is however too strict in some real applications. In this letter, instead of forcing a common support, a general joint-sparse model is considered where sparse vectors just have partially shared supports. This letter provides a Bayesian approach that extends the sparse Bayesian learning for such a joint-sparse model. The proposed Bayesian framework is composed of a number of parametric models that correspond to distinct patterns of partially shared supports, and an efficient deterministic Bayesian inference algorithm is developed. The proposed method is characterized as a tuning parameter-free approach, which can effectively infer the underlying sparse structure and also the noise level. Experimental results show the superiority of the proposed approach.

This paper focuses on the problem of causally reconstructing compressive sensing (CS) captured video. The state-of-the-art causal approaches usually assume that the signal support is static or changing sufficiently slowly over time, where magnetic resonance imaging is widely used as a motivating example. However, such an assumption is too restrictive for many other video applications, where the signal support changes rapidly. In this paper, we propose a framework that combines motion estimation (ME), the Kalman filter (KF), and CS to adapt the reconstruction process to motions in the video so that the slowly changing assumption on the signal support is relaxed and consequently is more suitable for video reconstruction. Explicit and implicit ME are designed to provide motion-aware predictions, upon which a modified KF procedure is applied. Furthermore, three CS algorithms with embedded ME and KF are developed, and theoretical analyses are conducted via reconstruction error upper bounds to characterize the various factors that affect reconstruction accuracy. Extensive simulations utilizing actual videos are carried out, and the superiority of our methods is demonstrated.

This paper focuses on the reconstruction of a tensor captured using Compressive Sensing (CS). Instead of processing the signals via vectorization as is done in conventional CS, in tensor CS high dimensional signals are kept in their original formats, which benefits hardware implementation and eases memory requirements. In addition, more structures exist in a tensor along its various dimensions than in its vectorized format. Utilizing these various structures, this paper proposes a general reconstruction approach for tensor CS. Employing the proximity operator of a nonconvex norm function, a special case for a tensor with low rank and sparse structures is elaborated, which is shown to outperform the state-of-art tensor CS reconstruction methods when applied to magnetic resonance imaging and hyper-spectral imaging.

Compressed sensing (CS) is a signal acquisition paradigm that utilises the finding that a small number of linear projections of a sparse signal have enough information for stable recovery. This paper develops a Bayesian CS algorithm to simultaneously recover multiple signals that follow the Type-3 joint sparse model [1], [2], where signals share a non-sparse common component and have distinct sparse innovation components. By employing the expectation-maximization (EM) algorithm, the proposed algorithm iteratively updates the estimates of the common component and innovation components. In particular, we find that the update rule for the non-sparse common component in the proposed algorithm, differs from all the other methods in the literature, and we provides an interpretation that gives a valuable insight into why the proposed algorithm is successful in estimating the non-sparse common component. The superior performance of the proposed algorithm is demonstrated by numerical simulation results.

In this paper, we propose a novel channel impulse response (CIR) clustering algorithm using a sparsity-based method, which exploits the feature of CIR that power of multipath component (MPC) is exponentially decreasing with increasing delay. We first use a sparsity-based optimization to recover CIRs, which can be well solved by using reweighted ℓ1minimization. Then a heuristic approach is provided to identify clusters in the recovered CIRs, which leads to improved clustering accuracy in comparison to identifying clusters directly in the raw CIRs. The proposed algorithm incorporates the physical behaviors of MPCs into the clustering framework and enables applications with no prior knowledge of the clusters, such as number and initial locations of clusters. The results in this paper can be used to parameterize the CIR model of radio channels.

Compressive sensing (CS), as a new sensing/sampling paradigm, facilitates signal acquisition by reducing the number of samples required for reconstruction of the original signal, and thus appears to be a promising technique for applications where the sampling cost is high, e.g., the Nyquist rate exceeds the current capabilities of analog-to-digital converters (ADCs). Conventional CS, although effective for dealing with one signal, only leverages the intrasignal correlation for reconstruction. This paper develops a decentralized Bayesian reconstruction algorithm for networked sensing systems to jointly reconstruct multiple signals based on the distributed compressive sensing (DCS) model that exploits both intra- and intersignal correlations. The proposed approach is able to address-networked sensing system applications with privacy concerns and/or for a fusion-center-free scenario, where centralized approaches fail. Simulation results demonstrate that the proposed decentralized approaches have good recovery performance and converge reasonably quickly.

Distributed compressive sensing (DCS) concerns the reconstruction of multiple sensor signals with reduced numbers of measurements, which exploits both intra- and inter-signal correlations. In this paper, we propose a novel Bayesian DCS algorithm based on variational Bayesian inference. The proposed algorithm decouples the common component, that characterizes inter-signal correlation, from innovation components, that represent intra-signal correlation. Such an operation results in a computational complexity of reconstruction which is linear with the number of signals. The superior performance of the algorithm, in terms of the computing time and reconstruction quality, is demonstrated by numerical simulations in comparison with other existing reconstruction methods.

This article focuses on the problem of reconstructing dynamic sparse signals from a series of noisy compressive sensing measurements using a Kalman Filter (KF). This problem arises in many applications, e.g., Magnetic Resonance Imaging (MRI), Wireless Sensor Networks (WSN) and video reconstruction. The conventional KF does not consider the sparsity structure presented in most practical signals and it is therefore inaccurate when being applied to sparse signal recovery. To deal with this issue, we derive a novel KF procedure which takes the sparsity model into consideration. Furthermore, an algorithm, namely Sparsity-fused KF, is proposed based upon it. The method of iterative soft thresholding is utilized to refine our sparsity model. The superiority of our method is demonstrated by synthetic data and the practical data gathered by a WSN.

Subspace segmentation is the process of clustering a set of data points that are assumed to lie on the union of multiple linear or affine subspaces, and is increasingly being recognized as a fundamental tool for data analysis in high dimensional settings. Arguably one of the most successful approaches is based on the observation that the sparsest representation of a given point with respect to a dictionary formed by the others involves nonzero coefficients associated with points originating in the same subspace. Such sparse representations are computed independently for each data point via \ell_1-norm minimization and then combined into an affinity matrix for use by a final spectral clustering step. The downside of this procedure is two-fold. First, unlike canonical compressive sensing scenarios with ideally-randomized dictionaries, the data-dependent dictionaries here are unavoidably highly structured, disrupting many of the favorable properties of the \ell_1 norm. Secondly, by treating each data point independently, we ignore useful relationships between points that can be leveraged for jointly computing such sparse representations. Consequently, we motivate a multi-task learning-based framework for learning coupled sparse representations leading to a segmentation pipeline that is both robust against correlation structure and tailored to generate an optimal affinity matrix. Theoretical analysis and empirical tests are provided to support these claims.

Conventional dictionary learning frameworks attempt to find a set of atoms that promote both signal representation and signal sparsity fora class of signals. In distributed compressive sensing (DCS), in addition to intra-signal correlation, inter-signal correlation is also exploited in the joint signal reconstruction, which goes beyond the aim of the conventional dictionary learning framework. In this letter, we propose a new dictionary learning framework in order to improve signal reconstruction performance in DCS applications. By capitalizing on the sparse common component and innovations (SCCI) model, which captures both intra- and inter-signal correlation, the proposed method iteratively finds a dictionary design that promotes various goals: i) signal representation; ii) intra-signal correlation; and iii) inter-signal correlation. Simulation results showthat our dictionary design leads to an improved DCS reconstruction performance in comparison to other designs.

In this paper, we propose an active node selection framework for compressive sleeping wireless sensor networks (WSNs) in order to improve the signal acquisition performance and network lifetime. The node selection can be seen as a specialized sensing matrix design problem where the sensing matrix consists of selected rows of an identity matrix. By capitalizing on a genie-aided reconstruction procedure, we formulate the active node selection problem into an optimization problem, which is then approximated by a constrained convex relaxation plus a rounding scheme. The proposed approach also exploits the partially known signal support, which can be obtained from the previous signal reconstruction. Simulation results show that our proposed active node selection approach leads to an improved reconstruction performance and network lifetime in comparison to various node selection schemes for compressive sleeping WSNs.

We propose a novel dynamic reweighted ℓ2(DRℓ2) algorithm in the regime of dynamic compressive sensing. Our analysis shows that aiming to solve a Type II optimization problem, DRℓ2is effectively minimizing a `convex-concave' penalty in the coefficients that transitions from a convex region to a concave function using knowledge of past estimations. DRℓ2thus provides superior reconstruction performance compared with state-of-the-art dynamic CS algorithms.

In this paper, we propose a Generalized Kalman Filtered Compressive Sensing (Generalized-KFCS) framework to reconstruct a video sequence, which relaxes the assumption of a slowly changing sparsity pattern in Kalman Filtered Compressive Sensing [1, 2, 3, 4]. In the proposed framework, we employ motion estimation to achieve the estimation of the state transition matrix for the Kalman filter, and then reconstruct the video sequence via the Kalman filter in conjunction with compressive sensing. In addition, we propose a novel method to directly apply motion estimation to compressively sensed samples without reconstructing the video sequence. Simulation results demonstrate the superiority of our algorithm for practical video reconstruction.

Vehicle tracking information is necessary to enable safety communication systems and intelligent transportation systems. Compression technologies with high efficiency and low complexity provide a promising approach to address the transmission and computing problems in vehicle tracking applications. Especially, vehicular trajectory with sparse errors that happened in the measurement sensing process poses a great challenge on traditional compression algorithms. In this paper, we analyze and design a compressive sensing (CS) based erroneous trajectory compression and recovery algorithm for vehicle tracking scenario. Moreover, some theoretical bounds for the proposed recovery optimization problem are analyzed and proved. The CS-based method proposed in this paper could not only achieve a fairly high compression rate and recovery accuracy, but fit the bandwidth mismatch between the road side unit (RSU) and on board unit (OBU). In another aspect, the Kalman filtering (KF) technology is applied for further optimizing the system performance, e.g. mean square error (MSE). Extensive simulations with real vehicular trajectories are carried out, which shows that CS-based compression algorithm achieves relatively high compression performance compared to some state-of-the-art trajectory compression algorithms.

This paper advocates the use of the emerging distributed compressive sensing (DCS) paradigm in order to deploy energy harvesting (EH) wireless sensor networks (WSN) with practical network lifetime and data gathering rates that are substantially higher than the state-of-the-art. In particular, we argue that there are two fundamental mechanisms in an EH WSN: i) the energy diversity associated with the EH process that entails that the harvested energy can vary from sensor node to sensor node, and ii) the sensing diversity associated with the DCS process that entails that the energy consumption can also vary across the sensor nodes without compromising data recovery. We also argue that such mechanisms offer the means to match closely the energy demand to the energy supply in order to unlock the possibility for energy-neutral WSNs that leverage EH capability. A number of analytic and simulation results are presented in order to illustrate the potential of the approach.

In this paper, we propose a novel Compressive Sensing (CS)-enhanced spectrum sensing approach for Cognitive Radio (CR) systems. The new framework enables cyclic feature detection with a significantly reduced sampling rate. We associate the new framework with a novel model-based greedy reconstruction algorithm: interdependent matching pursuit (IMP). For IMP, the hidden block sparsity owing to the symmetry present in the cyclic spectrum is exploited which effectively reduces the degree of freedom of problem. Compared with conventional CS with independent support selection, a remarkable spectrum reconstruction improvement is achieved by IMP.

In this letter, we propose a new method for the joint design of both the projections and the sparsifying dictionary in order to improve signal reconstruction performance in compressive sensing (CS) applications. By capitalizing on the optimized projection matrix design in , which admits a closed-form expression as a function of any overcomplete dictionary, the proposed method does not need to involve directly the projection matrix. The projection matrix of our joint design can be directly derived based on the learned dictionary. Simulation results show that our joint design framework, which is constituted based on a set of training image patches, leads to an improved reconstruction performance in comparison to other recent approaches.

Sparse signals can be sensed with a reduced number of random projections and then reconstructed if compressive sensing (CS) is employed. Traditionally, the projection matrix has been chosen as a random Gaussian matrix, but improved reconstruction performance can be obtained by optimizing the projection matrix. In this paper, we are interested in projection matrix designs for sensing sparse signals in overcomplete dictionaries. In particular, we put forth a closed form design that stems from the formulation of an optimization problem, which bypasses the complexity of iterative design approaches.

Compressive sensing (CS), a new sampling paradigm, has recently found several applications in wireless sensor networks (WSNs). In this paper, we investigate the design of novel sensing matrices which lead to good expected-case performance - a typical performance indicator in practice - rather than the conventional worst-case performance that is usually employed when assessing CS applications. In particular, we show that tight frames perform much better than the common CS Gaussian matrices in terms of the reconstruction average mean squared error (MSE). We also showcase the benefits of tight frames in two WSN applications, which involve: i) robustness to data sample losses; and ii) reduction of the communication cost.

This letter considers the design of sensing matrices with good expected-case performance for compressive sensing applications. By capitalizing on the mean squared error (MSE) of the oracle estimator, whose performance has been shown to act as a benchmark to the performance of standard sparse recovery algorithms, we demonstrate that a unit-norm tight frame is the closest design-in the Frobenius norm sense-to the solution of a convex relaxation of the optimization problem that relates to the minimization of the MSE of the oracle estimator with respect to the sensing matrix. Simulation results reveal that the MSE performance of a unit-norm tight frame based sensing matrix surpasses that of other standard sensing matrix designs in various scenarios, which include sparse recovery with basis pursuit denoise (BPDN), the Dantzig selector and orthogonal matching pursuit (OMP). This also has important practical implications because a unit-norm tight frame based sensing matrix can be designed very efficiently.

Compressive sensing leverages the compressibility of natural signals to trade off the convenience of data acquisition against computational complexity of data reconstruction. Thus, CS appears to be an excellent technique for data acquisition and reconstruction in a wireless sensor network (WSN) which typically employs a smart fusion center (FC) with a high computational capability and several dumb front-end sensors having limited energy storage. This paper presents a novel signal reconstruction method based on CS principles for applications in WSNs. The proposed method exploits both the intra-sensor and inter-sensor correlation to reduce the number of samples required for reconstruction of the original signals. The novelty of the method relates to the use of the Frechet mean of the signals as an estimate of their sparse representations in some basis. This crude estimate of the sparse representation is then utilized in an enhanced data recovering convex algorithm, i.e., the penalized ℓ1minimization, and an enhanced data recovering greedy algorithm, i.e., the precognition matching pursuit (PMP). The superior reconstruction quality of the proposed method is demonstrated by using data gathered by a WSN located in the Intel Berkeley Research lab.

In this paper, we propose a penalized ℓ1minimization algorithm for reconstructing a time-varying signal based on compressive sensing (CS) principles. The time-varying signal can be seen as a sequence of slow-changing frames. In the proposed algorithm, all frames of the sequence are sampled at an equal rate, which makes the encoder simpler than frame-categorized methods. We introduce a specialized Fréchet mean of the target frame and several adjacent frames as the penalty vector to make the algorithm close to ℓ0minimization. We prove that the specialized Fréchet mean is a good approximation of the target frame for a sequence of slow time-varying signals. Experimental results demonstrates the superior reconstruction quality of the proposed algorithm.

This paper presents a novel signal reconstruction method based on the distributed compressive sensing (DCS) framework for application to wireless sensor networks (WSN). The proposed method exploits both the intra-sensor correlation and the inter-sensor correlation to reduce the number of samples required for recovering the original signals. An innovative feature of our method is using the Fr' echet mean of the signals to discover the common support of their sparse representations in some basis. Then a new greedy algorithm, called precognition matching pursuit (PMP), is proposed to further reduce the number of required samples with the knowledge of the common support. The superior reconstruction quality of the proposed method is demonstrated by both computer-generated signals and real data gathered by a WSN located in the Intel Berkeley Research lab.

Relay stations (RSs) are usually used to enhance the signal strength of the mobile stations (MSs) close to the cell boundary. However, the introduction of RSs to the cellular networks increases the interference to the MS served by the base station (BS) under spatial multiplexing mode. In this paper, we propose a method which can cancel the interference through a novel use of dirty paper coding (DPC). Then, under a practical spatial multiplexing frame structure, we evaluate the cellularrelay system performance under three scheduling algorithms, Round Robin (RR) algorithm, Maximum Signal to Interference and Noise Ratio (MaxSINR) algorithm and Proportional Fair (PF) algorithm. Simulation and analytical results show that the spectral efficiency (SE) of cellular-relay system can be dramatically increased by 30% by the proposed interference cancellation method, and RR algorithm achieves the highest gain, while the gain of MaxSINR algorithm is limited gain in our scheme.

In this paper, we consider a multiuser relay network under amplify and forward (AF) scheme, where each mobile station (MS) is supported by a unique relay station (RS). While the RSs forwards the scaled signal to the MSs, the interference caused by the multiuser nature of the system propagates via the RSs, which significantly degrades the performance gain of the relay technique. We propose two beamforming methods, zero forcing (ZF) method and maximizing signal to leakage ratio (MSLR) method, to suppress the multiuser interference in the relay networks. ZF method can effectively cancel multiuser interference for all MSs and RSs, while it requires a relatively large number of transmit antennas equipped at the base station (BS). The other method, MSLR, aims to maximize the received signal at each MS and RS, while minimize the energy leaking to other MSs and RSs. The advantage of MSLR method is that it does not impose a condition on the relation between the number of transmit antennas equipped at the BS and the receive antennas at the RSs and MSs. Simulations show that MSLR method has significant performance gains over ZF method.

In 802.16e OFDMA systems, noise and co-channel interference on pilots degrade the accuracy of channel state information (CSI) obtained from channel estimation (CE) and system performance is deteriorated. To evaluate the system performance loss due to imperfect CSI, CE model should be used to incorporate noise and interference's impact on CSI into system level simulator (SLS). In this paper, a CE model, which considers 802.16e OFDMA system's pilot pattern and subcarrier randomization, are proposed. The signal model of 802.16e OFDMA system's CE under noise and interference is presented first. Then the error of CSI estimation by LMMSE channel estimator is derived. The proposed CE model covers the main parts of CSI error. Simulation results show it is suitable for SLS by providing a good approximation to estimated CSI instead of performing true LMMSE CE.