Device-free respiratory monitoring has attracted significant attention due to its potential applications in sleep disorders, psychopathology, and cardiology. It enables respiratory monitoring in a device-free and contact-free manner by analyzing the influence pattern of human respiratory on surrounding wireless signals, such as mmWave signals. Although remarkable progress has been achieved in this task when the targets remain stationary, the respiratory monitoring will fail when the target moves freely. In this paper, we propose a device-free real-time respiratory abnormality monitoring method based on mmWave signal segmentation to solve the aforementioned problem. Specifically, we design the physical state assessment strategy to obtain the real-time states of the target, including positional movement, large-scale activities in place, and micro motions in place. We propose the Doppler signal segmentation method to extract the micro motions signals when the target position remains unchanged. We present the multi-frame joint analysis method to obtain the frequency of micro motions based on the extracted micro motion signals, thereby eliminating interference and achieving real-time respiratory abnormality monitoring. To validate the effectiveness of the proposed methods, we conduct extensive experiments on a 77GHz mmWave testbed. The results indicate that the proposed method is feasible for achieving real-time respiratory abnormality monitoring even when the target moves freely.

Reconfigurable intelligent surface (RIS)-enabled integrated sensing and communication (ISAC) systems enhance spectrum efficiency and sensing accuracy. Building on this, we propose a novel self-sustainable multi-functional RIS (S-MFRIS) concept that supports multiple functionalities: reflection, refraction, amplification, energy harvesting, and target sensing. By harvesting energy from incident signals, the S-MFRIS can reflect, refract, and amplify signals without needing an external power supply, effectively overcoming double-fading attenuation. Furthermore, by deploying low-cost sensor elements, the S-MFRIS can capture echo signals from multiple targets, mitigating the signal attenuation commonly associated with multi-hop links. Then, we establish an S-MFRIS-enabled ISAC system and formulate an optimization problem to maximize the signal-to-interference-plus-noise ratio (SINR) of the sensing targets, subject to constraints on communication rate, power budget, and reflection coefficients. To solve this non-convex problem, we decompose it into three sub-problems, which are efficiently addressed using an iterative algorithm. Simulation and numerical results demonstrate the following key findings: (1) The proposed algorithm achieves better convergence and performance than the semidefinite relaxation-based and random-based algorithms. (2) The performance of the MFRIS-aided system varies under different operating protocols, with the self-sustainable MFRIS outperforming other schemes, particularly when the power budget is sufficient. (3) The proposed S-MFRIS achieves$30\%-46\%$sensing SINR gains at most for the same total power budget or element configuration. (4) The number of sensing elements improves sensing performance up to a certain point, after which further increases in the number of elements yield diminishing returns.

Automatic modulation recognition (AMR) is a key technology in cognitive radio and plays a vital role in modern wireless communication systems. Recently, numerous attention-based AMR methods have been proposed to enhance recognition performance. However, most of these methods focus on the time-domain features of modulation signals while insufficiently exploring their frequency-domain features. To address this issue, this paper proposes a novel architecture named denoising autoencoder hybrid-domain self-attention network (DAE-HDSANet). Specifically, a denoising autoencoder (DAE) is employed to extract robust signal representations through self-supervised learning, while a hybrid-domain self-attention (HDSA) mechanism, incorporating the discrete cosine transform (DCT), adaptively fuses time and frequency-domain features for joint modeling. Experimental results demonstrate that DAEHDSANet achieves superior recognition performance compared to benchmark approaches. Moreover, the introduction of HDSA significantly enhances the discriminative capability for higher-order modulation signals, effectively compensating for the insufficient frequency-domain feature modeling in existing AMR methods.

Deep neural networks (DNNs) offer intelligent solutions for communications’ automatic modulation recognition (AMR) tasks. However, DNNs are vulnerable to adversarial attacks, which can lead to incorrect predictions. To address this critical challenge, this paper proposes a feature-alignment generative adversarial network (FA-GAN) to defend against adversarial attacks targeting DNNs. FA-GAN employs a bidirectional mapping mechanism to iteratively update and learn the feature differences between original and adversarial signals. It quantifies these differences using a self-attention feature alignment (SAFA) encoder, eliminating abnormal perturbations in adversarial signals. Black-box and white-box attack-defense experiments conducted on the publicly available RML2016.10a dataset demonstrate that the proposed FA-GAN not only significantly enhances the model’s defense against adversarial perturbations but also preserves the classifier’s original performance.

Deep neural networks (DNNs) offer advanced intelligent solutions for communications' automatic modulation recognition (AMR) tasks. However, AMR models are vulnerable to adversarial attacks, wherein carefully crafted perturbations are injected into input signals, causing the models to make incorrect predictions. Existing defense methods face challenges balancing recognition accuracy while effectively countering varied attacks. This paper proposes a dual-stage defense framework based on collaborative vector quantized masked autoencoder (VQMAE) to address this. In the first stage, collaborative VQMAE achieves feature compression and reconstruction through vector quantization. To further refine the adversarial signal by aligning features and minimizing the discrepancy between adversarial and original signals, feature alignment is employed to quantify their differences. In the second stage, a bidirectional Transformer is utilized to learn the prior distribution of tokens indices, thereby enhancing signal reconstruction accuracy. Blackbox and white-box attack-defense experiments conducted on the publicly available dataset RML2016.10a demonstrate that the proposed method significantly strengthens the model's resilience to adversarial perturbations while maintaining the classifier's original performance.

The simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) aided integrated sensing and communications (ISAC) framework holds promise for wide-range coverage, sensing, and communication. However, allocating multiple resources becomes challenging due to the coupled time, frequency, and space resources. Additionally, echo interference from communication users to the sensing target degrades sensing performance. To tackle these challenges, a performance optimization problem in a STAR-RIS-aided ISAC system is formulated, and an alternating optimization scheme is investigated to balance the defined sensing mutual information and communication performance by optimizing the active transmit beams, STAR-RIS reflecting and transmit beams, and multicarrier distribution variables under the energy splitting protocol. Simulation results demonstrate the convergence and effectiveness of the proposed scheme, highlighting its advantages in improving sensing performance compared to benchmark schemes.

ABSTRACT The rapid development of the Internet of Things (IoT) has led to an increasingly prominent issue of spectrum resource scarcity. To effectively address this shortage, automatic modulation classification (AMC) has emerged as one of the critical factors. Most existing deep learning‐based AMC methods rely on supervised attention models. However, these approaches have not fully accounted for the inherent characteristics of modulation signals and feature sparsity. In response, this paper proposes a weight non‐negative constraint recurrent self‐attention (WNRSA) model. This model incorporates a recurrent attention module (RAM) within an autoencoder architecture, creating a recurrent self‐attention extraction mechanism that enhances multi‐dimensional feature representations. RAM comprises three types of attention modules: spatial, frequency, and temporal. The point attention model (PAM) extracts local spatial information to emphasize critical regions in the image. The frequency attention model (FAM) captures salient features at different scales in the frequency domain, reducing noise sensitivity to details and high‐frequency information. The time attention model (TAM) captures temporal information, strengthening the ability to extract dynamic features. Additionally, we introduce weight non‐negative constraint and KL‐divergence regularization term to optimize the WNRSA model's loss function, achieving sparser feature representations and reducing sensitivity to noise. Experimental results demonstrate that the WNRSA model achieves superior performance across various signal‐to‐noise ratio (SNR) levels.

Vehicle-to-Everything (V2X) communication and computation encounter challenges in achieving ultrareliable, low-latency communication, and optimizing energy consumption in dynamic vehicular environments. To overcome these issues, intelligent reflecting surfaces (IRSs) are introduced to boost communication efficiency and reliability while lowering latency and energy use. This article presents an IRS-enhanced V2X system, employing multiple IRSs to improve vehicle communication and computation offloading through spectrum reuse principles. An effective utility function is developed to quantify total latency and energy consumption, facilitating precise system evaluation and optimization. The complex optimization problem is divided into four subproblems: 1) vehicle operation mode selection; 2) spectrum reuse allocation; 3) computation offloading decision; and 4) beamforming design. A swap-matching-based tabu-search method solves the mode selection subproblem, while semi-definite relaxation and penalty functions address the other issues integrated through an alternating optimization algorithm. The optimized system achieves efficient and reliable communication with reduced latency and energy consumption. Simulations reveal that effective utility function minimization significantly enhances system efficiency compared to benchmarks. Strategic IRS deployment reduces channel losses, resulting in substantial performance improvements and supporting intelligent, sustainable transportation network advancement.

The maritime joint sensing and communication system (MSCS) has recently emerged as a promising solution to address the maritime spectrum scarcity issue for high-efficiency communication and environmental sensing. To mitigate the significant path loss experienced over the complex sea surface, the intelligent reflecting surface (IRS) is integrated into the MSCS to enhance the signal quality by dynamically adjusting the phases of its reflecting elements. Building upon this foundation, we aim to improve the sensing performance by maximizing the sensing signal-to-noise ratio while maintaining normal maritime communication, which involves optimizing active and passive beamforming vectors. Under the unpredictable environmental information and high-complexity and high-overhead channel information estimation in MSCS, we propose a heuristic algorithm based on genetic evolution to tackle this problem. To ensure the algorithm convergence and the feasibility of available solutions, we introduce an individual processing approach and an elitist reservation strategy in each genetic generation. Numerical results and simulations validate the convergence and efficacy of the proposed algorithm. Additionally, we analyze the effects of critical parameters and IRS structure on the algorithm performance.

Abstract The rapid development of the Internet of Things has exacerbated issues such as spectrum resource scarcity, poor communication quality, and high communication energy consumption. Automatic modulation recognition (AMR), a key technology in cognitive radio, has emerged as a crucial solution to these challenges. Deep neural networks have been recently applied in AMR tasks and have achieved remarkable success. However, existing deep learning‐based AMR methods often need to consider the sensitivity of models to noise fully. This study proposes a masked autoencoder multi‐scale attention feature fusion model (MAE‐SigNet). This model integrates a MAE, multi‐scale feature extraction module, bidirectional long short‐term memory module, and MAM to accomplish the AMR task under low signal‐to‐noise ratio. Additionally, we optimize the cross‐entropy loss of the MAE‐SigNet model by introducing MAE decoder reconstruction error, which enhances the model's sensitivity to noise while achieving more accurate feature representation. Experimental results demonstrate that the MAE‐SigNet model achieves average recognition rates of 63.77%, 65.28%, and 75.26% on the RML2016.10a, RML2016.10b, and RML2016.04c datasets. Mainly, MAE‐SigNet exhibits outstanding performance at various levels of low signal‐to‐noise ratios from −6 to 4 dB.

The evolution of computing paradigms has been significantly influenced by the emergence of wireless-powered mobile edge computing (WP-MEC), fundamentally transforming resource-efficient processing at the network edges. Intelligent reflecting surfaces (IRSs) integrated with WP-MEC offer new opportunities by enhancing the channel quality while addressing the complex resource management challenges. To address this, a collaborative transmission and resource management scheme for communication, energy, and computation is provided in this article. In particular, a novel performance evaluation index, named energy cost utility is presented first to capture the IRS-aided coupling performances thoroughly. Subsequently, a collaborative transmission mechanism addressing communication, energy transmission, and edge computing issues is developed, leveraging adaptive IRS association. Furthermore, to enhance the system performance, a hierarchical optimization framework for the resource management with limited computation capability is proposed, which includes the upper-layer optimization-based IRS association and resource allocation, along with the lower-layer deep reinforcement learning-based active and passive beamforming, aimed at maximizing the energy cost utility. Compared to the other benchmark schemes, our proposed collaborative transmission and resource management approach demonstrates the ability to learn from the environment and improve behavior gradually and exhibits superiority in enhancing transmission quality and reducing energy consumption. Also, appropriate neural network parameters will significantly improve the performance and convergence rate of the proposed algorithm. Finally, the advantages of the IRS association regarding quantity and configuration for enhancing the energy cost utility are explored, highlighting its potential to shape the future of the Internet of Things.

Millimeter wave (mmWave) communication combined with massive multiple input multiple output (MIMO) system is one of the most promising technologies in future wireless networks due to the characteristics of high bandwidth and narrow beam. For the strong channel attenuation, the more accuracy channel state information (CSI) is needed to make sure that the signal is received accurately in mmWave system. In this paper, a hardware-friendly channel estimation algorithm, named modular image denoising approximation message passing (MIDAMP), is proposed by combining the real image denoising network (RIDNet) and the learning approximation message passing network. The gap between the estimated channel and the real channel can be greatly reduced through using the powerful denoising ability of MIDAMP. The results of simulation demonstrate that the proposed MIDAMP algorithm has better performance in estimation accuracy and achievable sum rate (ASR) compared with some existing algorithms.

A new pulse for cognitive radio-ultra wideband (CR-UWB) using multiple modified transform domain communication system (M-TDCS) is presented in this paper. Utilizing multiple M-TDCS to sense the electromagnetic environment and UWB as a versatile PHY layer to adapt various wireless channel conditions, this pulse can avoid interference to/from the existing wireless systems or noise. This pulse can be implemented using surface acoustic wave (SAW) devices and some other accessorial devices. The bit error ratio (BER) performance of system using this pulse can be observably improved for UWB systems.