This paper proposes a joint beamforming design for optimizing integrated sensing and communication (ISAC) systems under imperfect channel state information (CSI), leveraging a simultaneous transmission and reflection reconfigurable intelligent surface (STAR-RIS). Owing to the deployment of large-scale antenna arrays and high carrier frequencies, the Rayleigh distance can extend to tens or even hundreds of meters. This expansion leads to a fundamental paradigm shift in electromagnetic field characteristics, transitioning from the conventional far-field regime to the emerging near-field regime. As a result, the propagation characteristics experienced by users and the target may differ. Accordingly, we consider a practical scenario in which users and the target are situated in distinct fields. However, this hybrid-field model increases the system’s sensitivity to channel estimation errors (CEE). To this end, we propose a robust design that jointly optimizes beamforming for base station (BS) and STAR-RIS, aiming to maximize the achievable sum-rate of the nodes while satisfying the constraint of sensing requirements. Under a statistical CEE model, we derive the interference covariance matrix and reformulate the maximization problem as an equivalent weighted mean square error (MSE) minimization problem. Subsequently, the transformed problem is decoupled into multiple sub-problems using a block coordinate descent (BCD)-based algorithm. The algorithm capitalizes on semidefinite relaxation and Gaussian randomization to obtain an effective solution. Finally, simulation results validate the effectiveness of the proposed robust design. Compared with the baseline schemes, the proposed algorithm achieves a higher communication sum-rate and illustrates how various parameters affect the performance.

Existing monostatic integrated sensing and communication (ISAC) systems face challenges in realizing high-precision sensing due to a singular observation angle. This drives the exploration of cooperative sensing ISAC modes. However, communication and sensing heavily rely on line-of-sight (LoS) links, and the issue of obstacle blocking remains unavoidable even with collaborative efforts among multiple base stations (BSs). Simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS), can create virtual links in full space by dynamically adjusting its element coefficients. This effectively mitigates the pathloss caused by blockages. Thus, integrating STAR-RIS into cooperative ISAC networks holds substantial potential, which has not been widely investigated. In this study, a STAR-RIS-aided cooperative sensing framework is first proposed. To maximize the sensing signal-to-clutter-plus-noise ratio (SCNR), we jointly optimize the transmit beamforming at BS, transmission and reflection coefficients at STAR-RIS, and receive beamforming at multi-receivers. Also, the receivers selection is considered. To solve this nonconvex problem, we propose a joint optimization algorithm using alternating optimization (AO), fractional programming (FP), and semidefinite relaxation (SDR) based on the receivers selection mechanism, called the RS-AFS algorithm. Numerical results validate that the RS-AFS algorithm can significantly enhance the sensing SCNR.

The rapid development of the internet of things (IoT) has heightened the demand for timely data delivery in heterogeneous networks, making the age of information (AoI) a key performance metric. In this work, we consider an unmanned aerial vehicle (UAV)-assisted IoT network enhanced by a simultaneous transmission and reflection reconfigurable intelligent surface (STAR-RIS) to improve data freshness. We formulate a joint optimization problem with respect to UAV trajectory, STARRIS phase shifts, and device scheduling to minimize the AoI. To tackle the problem's complexity, we employ a proximal policy optimization (PPO)-based reinforcement learning approach that autonomously discovers effective control policies. Simulation results demonstrate that our proposed solution reduces AoI, demonstrating enhanced adaptability, robustness, and communication performance in dynamic IoT environments.

Automatically detecting surface defects in prismatic battery is crucial for ensuring quality meets established standards. Traditional methods face challenges in accurately identifying these defects due to their minute and varied shapes and high density of distribution. To address these issues, we propose an innovative network for prismatic battery surface defect (PBSD-Net), which employs shunted dynamic snake convolution and focal modulation to detect surface defects in prismatic battery. This network is integrated into the 2D-AOI system. Firstly, we introduce sliding slice amplification (SSA) as a training strategy to enhance the network’s ability to recognize densely clustered tiny defects. Secondly, we develop a novel method using the shunted dynamic snake convolution (SDSC) module and focal modulation (FM) to improve the extraction of deformation features, thereby addressing complex and sporadically scattered surface defects. By integrating the SDSC module and FM mechanism, the receptive field of the defect feature extraction network is expanded, enabling the acquisition of comprehensive defect edge features. Additionally, we introduce the quality focal loss (QFL) function to effectively tackle the issue of imbalanced sample types. Experimental results on the PBSD-RGB dataset demonstrate that our method achieves a mAP@50 of 85.8%, representing an improvement of approximately 7.7% over the baseline network. We have applied the PBSD-Net to an automatic defect detection system in a well-known battery production company. This enhancement significantly boosts the accuracy of surface defect detection in prismatic battery. The relevant code is at the https://github.com/yikuizhai/PBSD-Net.

As a large variety of applications emerge in vehicle-to-infrastructure (V2I) networks, the explosion of data places greater demands on communications and computing. Sensing technology offers accurate data by collecting real-time environmental information. Meanwhile, mobile edge computing (MEC) can significantly reduce communication latency and enhance computational efficiency. Therefore, integrating the two novel technologies into V2I applications to improve overall performance has become a research hotspot. In this paper, we focus on the optimization problem for determining caching, offloading, and matching strategies to minimize system cost under the joint communications, sensing, and MEC framework of V2I networks. First, we analyze the positive impact of sensing on signaling overhead and age of information (AoI), and derive a linear relationship between delay and AoI. Next, we formulate the optimization function of system cost, which is defined as the weighted sum of AoI and energy consumption and is proved to be NP-hard. To address this problem, we leverage an improved quantum particle swarm optimization (QPSO) algorithm to acquire a suboptimal solution of caching and offloading strategies. This significantly reduces computational complexity compared to the optimal solution obtained via the branch-and-bound (B&B) method. According to the vehicles’ AoI, we design a matching algorithm for the roadside unit (RSU) with vehicles. Building on these, we propose a QPSO-based algorithm under the joint communications, sensing, and MEC framework (QJCSM). Simulation results demonstrate that the QJCSM algorithm outperforms other baseline algorithms in terms of AoI and energy consumption and achieves near-optimal performance with low complexity.

Recently, Deep Reinforcement Learning (DRL) has been applied to enhance the Quality of Experience (QoE) of Adaptive Bitrate Streaming (ABR) by adjusting the video quality level in real time based on instantaneous network conditions. To build a state-of-the-art DRL-based ABR (DRLABR) algorithm, it must learn from the clients’ actual network and video streaming behavior. However, collecting such data directly from clients introduces several challenges, including privacy concerns, high bandwidth consumption, and the straggler effect—where poor network conditions of certain clients delay the training process, as DRLABR’s performance is highly dependent on network interactions. To overcome these limitations, we propose a decentralized training approach for DRLABR using a Federated Learning (FL) framework. Instead of gathering raw data, clients train their local DRLABR models independently and send only model updates to the central server. To address the straggler issue, we propose to desynchronize the FL update rules, allowing clients to contribute their model updates at their own pace, regardless of varying network conditions. In addition, we design a DRL-based client selection mechanism to prevent oversampling of high-bandwidth clients, which could lead to model divergence, thereby ensuring balanced participation and improving the overall training efficiency. We validate our approach through a comprehensive simulation encompassing diverse video content and real-world network traces, simulating a wide range of streaming activities. Our results show that the proposed framework significantly outperforms conventional FedAvg and FedAsync methods, achieving the highest average QoE score of 2.02 and reducing the total training latency by 21.26%.

Abstract User pairing plays an important role in device‐to‐device (D2D) relay communication, contributing significantly to maintaining low energy consumption, high throughput, and overall energy efficiency in the communication system. To achieve these purposes, an attention‐based long short‐term memory motion prediction model (AT‐LSTM) and propose a joint power control algorithm. Leveraging these techniques, we also propose a D2D user pairing algorithm, distance–power–SINR pairing algorithm (DPSPA), which comprehensively considers factors such as D2D communication distances, transmit power, and signal‐to‐interference‐plus‐noise ratio. Initially, the AT‐LSTM model is utilized to predict the location of users. Subsequently, the distance between the user terminal device and each communication point and the base station, filtering cache points, and non‐cache points within the D2D communication radius are calculated. Then, based on the distance, required transmission power, and signal‐to‐interference‐plus‐noise ratio of each point, the evaluation index (the best matching product) is obtained. Finally, the point with the maximum best matching product is selected for D2D direct communication mode, D2D relay communication mode, or cellular communication mode. Simulation results demonstrate that DPSPA effectively reduces system energy consumption, enhances system throughput, and improves overall energy efficiency.

Man-made workpiece counting is a routine job for manufactory workers; however, this is an error-prone task. In this article, we are interested in detecting and counting arbitrary workpieces in industrial manufacturing. Therefore, we construct a comprehensive and large-scale open-world public benchmark dataset for workpiece counting, called workpiece counting dataset, which includes 121 475 instances of workpieces from 351 different categories. We also propose a novel method for workpiece detection and counting, named two-stage workpiece counting network. The first stage of the network is to develop a class-agnostic detector to localize each workpiece instance, followed by the second stage to employ an unsupervised deep clustering strategy with the backbone network pretrained in a workpiece convolutional autoencoder for decision boundary prediction, achieving workpiece clustering under unknownKvalues. Finally, our experiments show that the proposed method outperforms current mainstream methods, greatly enhancing the efficiency of factory operations.

Backscatter communications (BC) have recently emerged as a promising technology for sixth-generation (6G) networks with device-to-device (D2D) communications to support energy-efficient Internet of Things communications. This paper investigates the power allocation problem for energy-efficient 6G networks with backscatter-enabled D2D (BC-D2D) communications. To this end, we formulate the power allocation problem as a biobjective optimization problem that jointly maximizes the sum data rate and minimizes the power consumption of the networks, subject to energy-harvesting, reflection coefficient (RC) and power constraints. To solve this problem, suboptimal RCs are first analytically obtained based on the energy-harvesting constraints of the BC-D2D transmitters. Next, the biobjective optimization problem is transformed into a convex, single-objective power allocation optimization problem using the weighted sum approach, which is then solved using convex optimization. Results show that the proposed scheme outperforms the baseline schemes in energy efficiency under scenarios with different numbers of cellular user equipment and BC-D2D pairs.

With the rapidly increasing data traffic in vehicle-to-infrastructure (V2I) networks, the demand for communication and computing rises proportionally. As such, joint communication, computing and sensing becomes crucial for addressing the demands in V2I networks. This paper proposes an energy-efficient joint communication, sensing and mobile edge computing (MEC) system for V2I networks. In the proposed system, the roadside unit (RSU) is equipped with the functions of communication, sensing, and cache-aided computing. By combining beam prediction and tracking, the optimization problem of system energy consumption is formulated. The problem is divided into three parts, which are solved using the joint communication, sensing and MEC online (JCSMO) strategy. Simulation results show that the proposed JCSMO strategy outperforms the baseline schemes in system energy consumption.

In a wireless device-to-device (D2D) network, mobile edge caching can reduce transmission cost for network traffic, but it may also cause outdated caching information. How to reduce the transmission cost and improve the freshness of information becomes an important issue. A dedicated caching device called helper which has a large cache can be used in a wireless D2D network, which can significantly improve the performance of network. In this paper, to better model the file-centric data transmission, we proposed a concept called age of file (AoF), which is defined as the duration from the latest updating of the file. We analyzed the AoF, energy cost and updating cost of the files in the network level. We comprehensively consider the AoF and energy cost through the maximum and minimum normalization methods, and proposed an AoF-based accessing strategy. In the strategy, users can adjust the access of files according to their demand for file freshness and improve the quality of service. The results indicate that this strategy can reduce the energy cost for file transmission and reduce the total cost of the entire network while satisfying the freshness of files.

In wireless powered communication networks (WPCNs), the wireless energy acquisition is weak due to the fading characteristic of wireless link, which is hard to supply the energy. Reconfigurable intelligent surface (RIS) is a promising performance enhancement technology for WPCNs, which can improve the energy and spectral efficiency, expand the network coverage, and thus improve the throughput. This paper proposes a new framework for energy harvesting and information transmission of non-orthogonal multiple access (NOMA)-based WPCNs aided by the RIS. The total transmission capacity is maximized by joint optimization of energy precoding, information precoding, and the phase shift of RIS, which turns out to be a non-convex problem. So we propose a joint optimization algorithm combing alternating optimization (AO), semidefinite relaxation (SDR), and successive convex approximation (SCA) methods, which is called ASS algorithm for short. Numerical results prove that the throughput of the multi-user system can be significantly improved.

With practical finite alphabet inputs, usually the throughput of a practical communication system cannot reach the capacity based on assumption of Gaussian inputs. In this paper, we study the resource allocation strategy for multiple-input multiple-output (MIMO) full-duplex backscatter assisted wireless-powered communication network (FD-BAWPCN) with finite alphabet inputs, to maximize the sum-throughput. Firstly, we propose a gradient-based resource allocation strategy (GBRA), which alternately optimizes the precoder and time allocation based on the two-block alternating direction method of multipliers (ADMM). Secondly, to reduce the computational complexity and improve the feasibility of the strategy in practical applications, we further propose a codebook-based resource allocation strategy (CBRA), which can run more than three orders of magnitude faster than GBRA at the expense of a small sum-rate degradation. Then, the performance of the full-duplex (FD) and half-duplex (HD) system with or without backscatter assistance using GBRA are compared and analyzed. Numerical results demonstrate that the GBRA strategy has great robustness under various conditions but suffers a high computational complexity, and the CBRA strategy is effective and converges speedily.

Wireless body area network (WBAN) has become a promising technology, which can be widely applied in health monitoring, and so on. However, the performance of a practical WBAN may severely suffer from the degradation caused by dynamic nature of wireless channels with the movements of human body. Traditional communication frameworks cannot catch up with the channel variation of dynamic WBANs, which may severely degrade the performance, so an accurate channel prediction model is necessary for developing an efficient transmission strategy. In this paper, we propose a DeepBAN communication framework for dynamic WBANs. In our proposed framework, a temporal convolution network (TCN) based deep learning approach is adopted for channel prediction, the computationally intensive task of which is processed by mobile edge computing (MEC), to reduce the response time. Given the predicted channel conditions, we propose a joint power control, time-slot allocation, and relay selection algorithm to maximize the energy efficiency of the system, taking into account the transmission reliability and end-to-end latency requirements. We evaluate the performance of DeepBAN, and the results show that it can achieve energy-efficient, reliable, and low-latency data transmission in dynamic WBANs, which can improve the system energy efficiency by 15% compared with the stochastic scheduling scheme.

In current research of device‐to‐device (D2D) caching strategies, when terminals request and receive files within the transmission range, the effect of file redundancy and idle loss rate of the cache in terminals are rarely considered. In this study, the authors establish a motion and the transfer model of D2D terminals, and define a loss function considering cache idle loss and file redundancy. Then the relationship between the cache size and the number of terminals is analysed, and the optimal cache size is calculated according to the loss function. Through the analysis, the authors can see that the performance of the D2D caching network is almost constant when the product of the buffer size and the number of terminals is constant. Therefore, inspired by the above analysis and results, the authors devise an adaptive caching strategy, which determines the file caching method according to the comparison of the actual size of the terminal cache and the optimal cache size with respect to the loss function. Simulation results indicate that, compared with other caching schemes, the proposed adaptive strategy yields a significantly higher cache‐hitting rate and less delay.

Due to the lack of explicit expression of instantaneous mutual information (IMI), online throughput optimization for energy harvesting (EH) MIMO systems with finite alphabet inputs suffers from heavy computation burden and large feedback overhead. We propose a two-layer predictive strategy with the aid of context information, by constructing a precoding codebook offline at both the transmitter and receiver, to relax the time pressure for the online strategy. Two schemes were provided to accelerate the codebook construction. One is to estimate IMI during optimization by using a reference function. The other is by starting the optimization of each precoder from an adjacent precoder. With the channel distribution information (CDI), the feedback overhead is reduced by minimizing the codebook size. By the methods above, the building speed of the precoding codebook can be up to two orders of magnitude faster than no accelerating strategies. Meanwhile, through topology analysis, we propose a higher layer predictive strategy, which determines the index of allocated energy at each slot within a frame. The higher layer strategy takes full advantage of the constructed precoding codebook and context information when estimating the future throughput. Simulation results demonstrate the effectiveness and practicality of the proposed two-layer strategy.

User-centric ultra-dense network (UUDN) has been envisioned as a promising network paradigm due to its advantages over traditional cell-centric structure. This paper introduces full duplex (FD) technology into UUDN and presents a joint scheme of user access, subchannel allocation and power control for FD UUDN. The scheme firstly maximizes the total network capacity under given rate requirements and transmit power constraints. Then if there are no feasible solutions, the system finds the minimal total power of the network to satisfy the rate requirements. The capacity maximization problem is solved by MAPEL algorithm and the power minimization problem is solved by a proposed heuristic algorithm. Simulation demonstrates the spectrum efficiency enhancement and the consumed power reduction of UUDN compared with cell-centric UDN.

In this paper, we maximize the sum-throughput of a multi-user MIMO full-duplex wireless powered communication network (FD-WPCN) by jointly optimizing the precoding matrices and time allocation vector. In particular, we consider a more practical scenario with energy causality and finite alphabet inputs. The original optimization problem is non-convex and intractable. We prove that the time allocation problem is convex with the precoding matrices fixed. We also simplify the optimal precoder design to a convex power allocation problem for a given time allocation vector. By optimizing the precoding matrices and the time allocation vector alternately, we propose a two-step iterative algorithm to find the optimal solution. The simulation results show that our scheme outperforms other benchmark schemes significantly.

In this paper, a fast time-varying channel estimation method for Single Carrier Frequency division Multiple Access (SC-FDMA) communications on the Long Term Evolution (LTE) uplink is proposed. In uplink SC-FDMA, fast variations of the channel result in inter-carrier interference (ICI) which cannot be estimated properly with conventional estimation method. The proposed estimation method is performed all in the frequency domain and only part of channel frequency response for each uplink user is needed to estimate. The time variations of the frequency domain transmission function are modeled by the Basis expansion model (BEM). Simulation results demonstrate the performance and effectiveness of the proposed method over fast time-varying channels.