Federated split learning (FSL) has rapidly emerged as a promising paradigm for enabling ubiquitous intelligence in next-generation networks. However, current FSL approaches incur significant communication overhead and diminished training efficiency due to the frequent transmission of high-dimensional smashed data and gradients between devices and the base station. To address these issues, we propose a low-rank approximation (LoRA)-based FSL scheme, referred to as low-rank FSL. We analyze the convergence performance of low-rank FSL by considering the influence of LoRA rank on non-convex loss functions. To minimize a weighted sum of overall training latency and energy consumption in resource-constrained wireless networks, we formulate a long-term optimization problem by jointly optimizing computing frequency, power allocation, decoding order, LoRA rank, and split layer selection. An iterative optimization algorithm is then developed to solve this problem with low computational complexity. Numerical results demonstrate that our low-rank FSL reduces communication overhead by at least 300% while maintaining high learning performance. Moreover, our optimization algorithm achieves a low weighted cost in terms of training latency and energy consumption.

This paper studies the problem of long-term, high-quality data collection in dynamic industrial environments using autonomous mobile robots (AMRs). We formulate a max–min data-rate optimization problem that jointly designs AMR trajectories, receive beamforming, sensor association, and decoding policy under mobility, communication, and battery constraints. The problem is challenging because motion decisions directly influence link quality, which affects decoding and beamforming efficiency, while all of these factors are tightly coupled with battery dynamics, resulting in a mixed-integer, time-varying, and non-convex design. To handle this strong coupling and the need for scalable long-horizon decision-making, we propose a semi-distributed reinforcement learning framework that combines cloud-level global coordination with fog-level local adaptability. In this framework, a cloud-layer deep Q-network determines AMR-sensor associations, and a fog-layer federated actor-critic algorithm jointly designs continuous trajectories and beamforming. A Fubini–Study distance-assisted k-means clustering method enhances multi-antenna directional gains, and next-generation multiple access (NGMA) improves spectral efficiency and interference suppression. Theoretical analysis confirms the convergence and computational efficiency of the proposed algorithm. Simulations show that the proposed approach achieves faster and more stable convergence, sustains large-scale coverage through periodic recharging, and significantly improves the minimum data rate compared with orthogonal multiple access-based baselines.

Cell-free (CF) networks have attracted increasing attention for their effectiveness in mitigating inter-cell interference through cooperative transmission among distributed access points (APs). However, conventional terrestrial CF networks often lack spatial flexibility and struggle to adapt to dynamic environments. To overcome these limitations, we propose a new CF network served by unmanned aerial vehicles (UAVs) equipped with a three-dimensional (3D) rotatable antenna array. Combined with the UAV’s controllable 3D position, the resulting six-dimensional (6D) spatial reconfigurability enables the active beam steering of such aerial APs, thereby enhancing interference mitigation and dynamic user association. However, this design, referred to as 6D aerial rotatable antenna arrays (6DARAs), faces several critical challenges, such as high-dimensional coupled control variables, time-varying user positions, and increased channel state information (CSI) estimation overhead. To address these issues, we develop a two-timescale optimization framework that separates large-timescale 6DARA control (i.e., clustering, position, and rotation) from small-timescale signal processing. At the small-timescale, a closed-form team minimum mean-squared error decoder is derived using local and statistical CSI. At the large-timescale, 6DARA clustering is modeled as a local altruistic game and solved via a concurrent update algorithm, while 6DARA mobility is managed by an enhanced multi-agent reinforcement learning algorithm for efficient position and rotation adaptation under partial observability. Simulation results demonstrate that the proposed network and optimization framework significantly outperform existing baselines in terms of throughput, scalability, and robustness in dynamic environments.

As a paradigm shift towards pervasive intelligence, semantic communication (SemCom) has shown great potentials to improve communication efficiency and provide user-centric services by delivering task-oriented semantic meanings. However, the exponential growth in connected devices, data volumes, and communication demands presents significant challenges for practical SemCom design, particularly in resource-constrained wireless networks. In this work, we propose a task-agnostic semantic communication (TASC) framework capable of supporting multimodal data across diverse tasks. To investigate the interplay between communication and intelligent tasks from an information-theoretic perspective, we introduce a distributed multimodal information bottleneck (DMIB) principle, which enables the extraction of minimal sufficient unimodal and multimodal representations by eliminating redundant information while preserving task-relevant semantics. To further reduce the communication overhead, we develop an adaptive semantic feature transmission method under dynamic channel conditions. Then, TASC is trained based on federated meta-learning (FML) to learn a well-initialized model for rapid adaptation and generalization. To gain deep insights, we conduct theoretical analysis and devise resource management to accelerate convergence while minimizing the training latency and energy cost. Moreover, we develop a joint user selection and resource allocation algorithm to address the non-convex problem with theoretical guarantees. Extensive simulation results validate the effectiveness and superiority of the proposed TASC compared to baselines.

Fluid antenna system (FAS) revolutionizes wireless communications via utilizing position-flexible antennas that dynamically optimize channel conditions and mitigate multipath fading. This innovation is particularly valuable in indoor environments, in which signal propagation is severely degraded due to structural obstructions and complex multipath reflections. In this paper, we investigate the channel modeling and the joint optimization of antenna positioning, beamforming, and power allocation for indoor FAS. In particular, we propose a layout-specific channel model, and employ the novel group relative policy optimization (GRPO) algorithm for tackling the optimization problem. Compared to the state-of-the-art Sionna model, our model achieves an 83.3% reduction in computation time with an approximately 3 dB increase in root-mean-square error (RMSE). When simplified to a two-ray model, our model allows for a closed-form antenna position solution with near-optimal performance. For the joint optimization problem, our GRPO algorithm outperforms proximal policy optimization (PPO) and other baselines in sum-rate, while requiring only 50.8% computational resources of PPO, thanks to its group advantage estimation. Simulation results show that increasing either the group size or trajectory length in GRPO does not yield significant improvements in sum-rate, suggesting that these parameters can be selected conservatively without sacrificing performance.

In this paper, we propose a hybrid learning framework that combines federated and split learning, termed semi-federated learning (SemiFL), in which over-the-air computation is utilized for gradient aggregation. A key idea is to strategically adjust the learning rate by manipulating over-the-air distortion for improving SemiFL’s convergence. Specifically, we intentionally amplify amplitude distortion to increase the learning rate in the non-stable region, thereby accelerating convergence and reducing communication energy consumption. In the stable region, we suppress noise perturbation to maintain a small learning rate for improving SemiFL’s final convergence. Theoretical results demonstrate the antagonistic effects of over-the-air distortion in different regions, under both independent and identically distributed (IID) and non-IID data settings. Then, we formulate two energy consumption minimization problems, one for each region, which implements a two-region mean square error threshold configuration scheme. Accordingly, we propose two resource allocation algorithms with closed-form solutions. Simulation results show that under different network and data distribution conditions, strategically manipulating over-the-air distortion can efficiently adjust the learning rate to improve SemiFL’s convergence. Moreover, energy consumption can be reduced by using the proposed algorithms.

The rapid development of generative artificial intelligence (AI) has introduced significant opportunities for enhancing the efficiency and accuracy of image transmission within semantic communication systems. Despite these advancements, existing methodologies often neglect the difference in importance of different regions of the image, potentially compromising the reconstruction quality of visually critical content. To address this issue, we introduce an innovative generative semantic communication system that refines semantic granularity by segmenting images into key and non-key regions. Key regions, which contain essential visual information, are processed using an image oriented semantic encoder, while non-key regions are efficiently compressed through an image-to-text modeling approach. Additionally, to mitigate the substantial storage and computational demands posed by large AI models, the proposed system employs a lightweight deployment strategy incorporating model quantization and low-rank adaptation fine-tuning techniques, significantly boosting resource utilization without sacrificing performance. Simulation results demonstrate that the proposed system outperforms traditional methods in terms of both semantic fidelity and visual quality, thereby affirming its effectiveness for image transmission tasks.

To facilitate fast data aggregation in Internet of Things, over-the-air computation (AirComp) is a communication-efficient enabler by virtue of its high spectrum efficiency and low transmission latency. However, traditional AirComp faces challenges such as signal misalignment, imperfect channel state information (CSI), and noisy fading channels. In this paper, we employ a multi-functional reconfigurable intelligent surface (MF-RIS) to alleviate mean square error (MSE) of AirComp through a joint design of transceiver beamforming and MF-RIS coefficients, but it necessitates solving a mixed-integer nonlinear programming problem. To tackle this issue, we propose an alternating optimization algorithm based on the semidefinite relaxation approach and difference-of-convex programming. Numerical results underscore the performance gains achieved by the proposed algorithm, as well as the remarkable proficiency of the MF-RIS in suppressing MSE under imperfect CSI.

The deployment of large language models (LLMs) at the network edge has emerged as a critical paradigm for enabling low-latency, privacy-preserving AI services. However, resource-constrained user devices (e.g., smartphones, IoT endpoints) struggle to execute computationally intensive LLM inference tasks solely. Collaborative computing frameworks that offload subtasks to edge servers offer a promising solution but face challenges in balancing task completion latency, energy consumption, and dynamic network conditions. This paper addresses these challenges by proposing a two-stage optimization framework that jointly manages task offloading decisions and computational resource allocation, aiming to minimize the total task completion latency. Simulation results demonstrate that the proposed algorithm reduces total latency by 49.6% compared to random, greedy, and genetic baselines, exhibits robust adaptability across varying energy budgets, and achieves 97.8% energy utilization in resource-constrained edge environments.

Cell-free massive multiple-input multiple-output (CF-mMIMO) networks improve spectral efficiency via coordinated transmission and flexible beam selection. However, the resource allocation in such networks presents a high-dimensional optimization challenge due to the distributed architecture with multiple access points and antennas. To address this, we first formulate a beam selection problem, and then propose an efficient Q-value mixing (QMIX)-based algorithm. Furthermore, recognizing the inherent limitations of deep reinforcement learning (DRL) in practical applications, such as costly training, risky exploration phases, and suboptimal convergence speeds, we design a data-driven digital twin (DT) framework to optimize the DRL training phase. Simulation results show that our approach achieves accelerated convergence and enhanced stability compared to conventional methods. DT-based pre-training establishes a robust performance lower bound prior to real-system deployment.

In existing deep learning-based semantic communication systems, centralized training of semantic models brings a risk of privacy leakage, whereas distributed training imposes a huge computational burden on user equipments (UEs). To address these challenges, we propose a hybrid federated learning (Hybrid-FL) framework to alleviate the computational burden on UEs while protecting the user privacy. Specifically, each UE uploads local gradients and semantic symbols to the base station for the collaborative training of global and local semantic models. Furthermore, we propose a joint communication and computation scheme for supporting the model aggregation and semantics transmission. To gain deep insights, we expose the joint impact of communication and computation on the convergence behavior of Hybrid-FL by deriving an upper bound. Then, we formulate a mixed-integer nonlinear programming problem to improve the convergence performance of Hybrid-FL, which is then effectively solved by using our proposed algorithm that developed based on alternating and matching theory. Experimental results demonstrate that Hybrid-FL outperforms conventional FL by achieving a 20% accuracy gain and a 80% latency reduction.

Over-the-air computation has emerged as a high-spectrum efficient and low-latency solution for model aggregation in federated learning (FL) by leveraging the superposition property of wireless channels. However, traditional over-the-air FL (AirFL) faces challenges such as signal misalignment, imperfect channel state information (CSI), and noisy fading channels. To enable more efficient and reliable AirFL in Internet of Things (IoT), we employ a multifunctional reconfigurable intelligent surface (MF-RIS) to alleviate mean square error (MSE) of AirFL model aggregation. By deriving the convergence analysis of AirFL in both convex and nonconvex settings, we unveil the impact of MSE on MF-RIS-aided AirFL under varying conditions. Based on the theoretical insights, we aim to minimize the MSE through a joint design of transceiver beamforming and MF-RIS coefficients, but it necessitates solving a mixed-integer nonlinear programming (MINLP) problem. To solve it efficiently, we propose an alternating optimization (AO) algorithm based on the semidefinite relaxation (SDR) approach and difference-of-convex (DC) programming. The efficacy of our approach is corroborated by numerical results, which underscore the performance gains achieved by the proposed algorithm. Additionally, the MF-RIS demonstrates remarkable proficiency in suppressing MSE and bolstering AirFL performance, even under conditions of imperfect CSI.

In this letter, we study the beamforming design for channel estimation of multi-functional reconfigurable intelligent surface (MF-RIS)-aided multi-user communications that supports simultaneous signal reflection, refraction, and amplification. A least square (LS) based channel estimator is proposed for MF-RIS by considering both the coupled MF-RIS beams and the introduced thermal noise. With the discrete fourier transform (DFT)-matrix, the MF-RIS beamforming design problem is simplified under the proposed LS channel estimator. The optimal MF-RIS beamforming design that achieves the Cramér–Rao lower bound (CRLB) of channel estimator is obtained with the proposed alternating optimization algorithm. Simulation results demonstrate the effectiveness of the proposed beamforming design in reducing the impact of thermal noise.

We propose a robust transceiver design for a covert integrated sensing and communications (ISAC) system with imperfect channel state information (CSI). Considering both bounded and probabilistic CSI error models, we formulate worst-case and outage-constrained robust optimization problems of joint transceiver beamforming and radar waveform design to balance the radar performance of multiple targets while ensuring the communications performance and covertness of the system. The optimization problems are challenging due to the non-convexity arising from the semi-infinite constraints (SICs) and the coupled transceiver variables. In an effort to tackle the former difficulty, S-procedure and Bernstein-type inequality are introduced for converting the SICs into finite convex linear matrix inequalities (LMIs) and second-order cone constraints. A robust alternating optimization framework referred to alternating double-checking is developed for decoupling the transceiver design problem into feasibility-checking transmitter- and receiver-side subproblems, transforming the rank-one constraints into a set of LMIs, and verifying the feasibility of beamforming by invoking the matrix-lifting scheme. Numerical results are provided to demonstrate the effectiveness and robustness of the proposed algorithm in improving the performance of covert ISAC systems.

Distributed over-the-air computation (AirComp) is a promising technology for fast data aggregation in wireless networks by leveraging multiple access channel to achieve communication and computation simultaneously. However, device-to-device (D2D) links applied are vulnerable to obstacles, and the performance of distributed AirComp is restricted by the device with the worst channel condition. To tackle these issues, we introduce a multi-functional reconfigurable intelligent surface (MF-RIS) to reconstruct the wireless propagation environment, where the MF-RIS can achieve signal reflection, refraction, and amplification simultaneously. Specifically, we propose an MF-RIS-aided distributed AirComp framework, where MF-RIS receives the aggregated data from all devices and then transmits it to each device for post-processing. We formulate a mean-squared error (MSE) minimization problem by jointly optimizing transmit scalar, receive scalar, and MF-RIS coefficients. To address this non-convex problem, we employ an alternating optimization (AO) technique to decompose it into three subproblems, where semi-closed form or closed form solutions are obtained. Then, we extend the single-input single-output (SISO) system into multiple-input multiple-output (MIMO) one. Next, we derive the asymptotic MSE performance for SISO and MIMO cases when the number of RIS elements and that of transmit/receive antennas are very large. Numerical results demonstrate the superiority of MF-RIS in improving MSE performance compared to the baseline without RIS. Additionally, the MF-RIS outperforms its passive counterparts, which reveals the advantages of deploying MF-RIS in distributed AirComp systems to reduce data aggregation error.

The robust beamforming design in multi-functional reconfigurable intelligent surface (MF-RIS) assisted wireless networks is investigated in this work, where the MF-RIS supports signal reflection, refraction, and amplification to address the double-fading attenuation and half-space coverage issues faced by traditional RISs. Specifically, we aim to maximize the system energy efficiency by jointly optimizing the transmit beamforming vector and MF-RIS coefficients in the case of imperfect channel state information (CSI). We first leverage the S-procedure and Bernstein-Type Inequality approaches to transform the formulated problem into tractable forms in the bounded and statistical CSI error cases, respectively. Then, we optimize the MF-RIS coefficients and the transmit beamforming vector alternately by adopting an alternating optimization framework, under the quality of service constraint for the bounded CSI error model and the rate outage probability constraint for the statistical CSI error model. Simulation results demonstrate the significant performance improvement of MF-RIS compared to benchmark schemes. In addition, it is revealed that the cumulative CSI error caused by increasing the number of RIS elements is larger than that caused by increasing the number of transmit antennas.

The rapid evolution of edge computing and artificial intelligence (AI) paves the way for pervasive intelligence in the next-generation network. As a hybrid training paradigm, federated split learning (FSL) leverages data and model parallelism to enhance training efficiency. However, existing FSL encounters unacceptable waiting latency due to device heterogeneity and synchronous model aggregation. To address this issue, we propose a semi-asynchronous FSL (SAFSL) framework that enables personalized model splitting and aperiodic model aggregation. We derive the convergence upper bound by considering factors such as the number of devices, training iterations, and data heterogeneity. To minimize the long-term average training latency while maintaining high energy efficiency in resource-constrained wireless networks, we formulate a stochastic mixed-integer nonlinear programming problem. By decomposing it into multiple sub-problems in each round, we propose a Lyapunov-based alternating optimization algorithm to solve it in an online manner. Numerical results demonstrate that our SAFSL achieves faster convergence with reduced communication overhead while maintaining high prediction performance under non-independent and identically distributed data, outperforming state-of-the-art benchmarks. Moreover, our algorithm achieves a low training latency, highlighting its superior performance and effectiveness.

In this paper, we propose a novel multi-functional reconfigurable intelligent surface (MF-RIS) that supports signal reflection, refraction, amplification, and target sensing simultaneously. Our MF-RIS aims to enhance integrated communication and sensing (ISAC) systems, particularly in multi-user and multi-target scenarios. Equipped with reflection and refraction components (i.e., amplifiers and phase shifters), MF-RIS is able to adjust the amplitude and phase shift of both communication and sensing signals on demand. Additionally, with the assistance of sensing elements, MF-RIS is capable of capturing the echo signals from multiple targets, thereby mitigating the signal attenuation typically associated with multi-hop links. We propose a MF-RIS-enabled multi-user and multi-target ISAC system, and formulate an optimization problem to maximize the signal-to-interference-plus-noise ratio (SINR) of sensing targets. This problem involves jointly optimizing the transmit beamforming and MF-RIS configurations, subject to constraints on the communication rate, total power budget, and MF-RIS coefficients. We decompose the formulated non-convex problem into three sub-problems, and then solve them via an efficient iterative algorithm. Simulation results demonstrate that: 1) The performance of MF-RIS varies under different operating protocols, and energy splitting (ES) exhibits the best performance in the considered MF-RIS-enabled multi-user multi-target ISAC system; 2) Under the same total power budget, the proposed MF-RIS with ES protocol attains$\rm {52.2}\%$,$\rm {73.5}\%$, and$\rm {60.86}\%$sensing SINR gains over active RIS, passive RIS, and simultaneously transmitting and reflecting RIS (STAR-RIS), respectively; 3) The number of sensing elements will no longer improve sensing performance after exceeding a certain number.

Although reconfigurable intelligent surface (RIS) can improve the secrecy communication performance of wireless users, it still faces challenges such as limited coverage and double-fading effect. To address these issues, in this paper, we utilize a novel multi-functional RIS (MF-RIS) to enhance the secrecy performance of wireless users, and investigate the physical layer secrecy problem in non-orthogonal multiple access (NOMA) networks. Specifically, we derive the secrecy outage probability (SOP) and secrecy throughput expressions of users in MF-RIS-assisted NOMA networks with external and internal eavesdroppers. The asymptotic expressions for SOP and secrecy diversity order are also analyzed under high signal-to-noise ratio (SNR) conditions. Additionally, we examine the impact of receiver hardware limitations and error transmission-induced imperfect successive interference cancellation (SIC) on the secrecy performance. Numerical results indicate that: i) under the same power budget, the secrecy performance achieved by MF-RIS significantly outperforms active RIS and simultaneously transmitting and reflecting RIS; ii) with increasing power budget, residual interference caused by imperfect SIC surpasses thermal noise as the primary factor affecting secrecy capacity; and iii) deploying additional elements at the MF-RIS brings significant secrecy enhancements for the external eavesdropping scenario, in contrast to the internal eavesdropping case.

The emergence of large language models (LLMs) with multi-task generalization capabilities is expected to improve the performance of artificial intelligence (AI)-as-a-service provision in 6G networks. By fine-tuning LLMs, AI services can become more precise and tailored to the demands of different downstream tasks. However, centralized fine-tuning paradigms pose a potential risk to user privacy, and existing distributed fine-tuning methods incur significant wireless transmission burdens due to the large-scale parameter transmission of LLMs. To tackle these challenges, by leveraging the low rank feature in LLM fine-tuning, we propose a wireless over-the-air federated learning (AirFL) based low-rank adaptation (LoRA) framework that integrates LoRA and over-the-air computation (AirComp) to achieve efficient fine-tuning and aggregation. Based on multiple-input multiple-output (MIMO) and orthogonal frequency division multiplexing (OFDM), we design a multi-stream AirComp scheme to fulfill the aggregation requirement of AirFL-LoRA. Furthermore, by deriving an optimality gap, we gain theoretical insights into the joint impact of rank selection and gradient aggregation distortion on the fine-tuning performance of AirFL-LoRA. Next, we formulate a non-convex problem to minimize the optimality gap, which is solved by the proposed backtracking-based alternating algorithm and the manifold optimization algorithm iteratively. Through fine-tuning LLMs for different downstream tasks, experimental results reveal that the AirFL-LoRA framework outperforms the state-of-the-art baselines on both training loss and perplexity, closely approximating the performance of FL with ideal aggregation.

To enhance the robustness and convergence of hierarchical federated learning (HFL) in wireless networks with imperfect channel state information (CSI), a quantized HFL (QHFL) framework is proposed. Considering the local training and communication latency, the outage probability of quantized gradient transmission is modeled under imperfect CSI. Then, the convergence of the proposed QHFL with transmission outage and gradient quantization is analyzed. Simulation results demonstrate the correlation between quantization accuracy and transmission outage, along with their joint impact on the HFL convergence, which align with the insight behind the convergence analysis.

In this paper, we investigate the computation offloading problem in a distributed reconfigurable intelligent surface (RIS)-aided cell-free network, where users offload computing-intensive tasks to their associated base stations with the aid of multiple RISs. To minimize the long-term energy consumption of all users under the constraints of latency tolerance, we formulate a long-term non-convex problem by jointly optimizing the offloading strategy, transmit power at users, reflection matrix at the RIS, receive beamforming at the BS, and CPU resources at both users and the server. To solve this intractable time-varying problem with multiple coupling variables, we propose a two-layer distributed proximal policy optimization (DPPO) algorithm for solving the problem with a high-dimensional decision-making space. Simulation results show that the proposed algorithm effectively reduces the long-term energy consumption of all users while completing the computing task within a given time limit.

To facilitate edge intelligence in Internet of Robotic Things (IoRT), over-the-air federated learning (AirFL) is a communication-efficient enabler by virtue of its high spectrum efficiency and low transmission latency. Supported by multi-functional reconfigurable intelligent surface (MF-RIS), the model aggregation process of AirFL can be facilitated thanks to full-space signal amplification. However, the learning performance of AirFL may be degraded by the uncertainty of wireless channels originated from inaccurate channel estimation and robot mobility. In this paper, we investigate the model aggregation problem of MF-RIS-aided AirFL in a dynamic IoRT system with imperfect channel state information (CSI). We aim to minimize the long-term mean square error (MSE) of AirFL model aggregation by jointly optimizing MF-RIS coefficients and transceiver beamforming. To enable online decision-making in the dynamic system, we propose a novel deep reinforcement learning (DRL)-based double-agent algorithm, where one agent first decides operating modes of MF-RIS elements, and then the other agent devises transceiver beamforming and other MF-RIS coefficients referring to the mode switching strategy. Numerical results unveil the effectiveness and robustness of the proposed DRL-based algorithm in suppressing long-term MSE under hostile CSI.

Federated learning (FL) is a promising distributed learning approach which enables multiple devices to collaboratively train deep neural networks in a privacy-preserving fashion. However, computing-limited devices in Internet of Things (IoT) networks are difficult to perform local model training, which is an obstacle to implementing connected intelligence at the network edge. In this paper, we propose a semi-federated learning (SemiFL) framework to support computing-heterogeneous devices for collaborative model training, even if some local devices do not have computing capabilities. Specifically, in SemiFL, computing-limited devices upload raw data to the base station (BS), while computing-powerful devices upload model parameters to the BS. To characterize the achievable learning performance of SemiFL in practical IoT networks, we derive the convergence upper bound with non-independent and identically distributed (non-IID) data. Furthermore, we formulate an optimization problem to minimize the computation and communication latency by jointly design the transmit power of IoT devices and the receive factor at the BS. Accordingly, we propose an two-stage optimization algorithm to solve the formulated non-convex problem efficiently. Extensive simulation results show that our SemiFL framework outperforms traditional FL in both IID and non-IID settings. In addition, it is shown that the proposed algorithm achieves lower latency than benchmarks under different settings.

Task-oriented semantic communication has gained increasing attention due to its ability to reduce the amount of transmitted data without sacrificing task performance. Although some prior efforts have been dedicated to developing semantic communications, the semantics in these works remains to be unexplainable. Challenges related to explainable semantic representation and knowledge-based semantic compression have yet to be explored. In this article, we propose a triplet-based explainable semantic communication (TESC) scheme for representing text semantics efficiently. Specifically, we develop a semantic extraction method to convert text into triplets while using syntactic dependency analysis to enhance semantic completeness. Then, we design a semantic filtering method to further compress the duplicate and task-irrelevant triplets based on prior knowledge. The filtered triplets are encoded and transmitted to the receiver for completing intelligent tasks. Furthermore, we apply the proposed TESC scheme to two emblematic text tasks: 1) sentiment analysis and 2) question answering, in which the semantic codec is meticulously customized for each task. Experimental results demonstrate that 1) the TESC scheme outperforms benchmarks in terms of Top-1 accuracy and transmission efficiency and 2) the TESC scheme enjoys about 150% performance gain compared to the traditional communication method.

Integrated sensing and communication integrated sensing and communications (ISAC) is an enabling technology for the sixth-generation mobile communications, which equips the wireless communication networks with sensing capabilities. In this article, we investigate transmit beamforming design for the multiple-input and multiple-output (MIMO)-ISAC systems in scenarios with multiple radar targets and communication users. A general form of multitarget sensing mutual information (MI) is derived, along with its upper bound, which can be interpreted as the sum of individual single-target sensing MI. Additionally, this upper bound can be achieved by suppressing the cross-correlation among the reflected signals from different targets, which aligns with the principles of adaptive MIMO radar. Then, we propose a multiobjective optimization framework based on the signal-to-interference-plus-noise ratio of each user and the tight upper bound of sensing MI, introducing the Pareto boundary to characterize the achievable communication-sensing performance boundary of the proposed ISAC system. To achieve the Pareto boundary, the max-min system utility function method is employed, while considering the fairness between the communication users and radar targets. Subsequently, the bisection search method is employed to find a specific Pareto optimal solution by solving a series of convex feasible problems. Finally, the simulation results validate that the proposed method achieves a better tradeoff between the multiuser communication and multitarget sensing performance. Additionally, utilizing the tight upper bound of sensing MI as a performance metric can enhance the multitarget resolution capability and angle estimation accuracy.

Many wearable Internet of Medical Things (IoMT) devices have limited computing power and small storage space. Additionally, the healthcare data sensed by a single IoMT device is not enough to train a sophisticated deep learning model. To address these challenges, we propose a federated split learning (FedSL) framework that allows for collaborative healthcare analytics on multiple IoMT devices with limited resources. Compared to centralized learning, FedSL can protect user privacy by not sending raw data over wireless networks. Furthermore, FedSL offers more flexibility than other federated learning methods. It enables even low-end IoMT devices to participate in model training and result inference. Experimental results show that our FedSL performs well on medical imaging tasks with different data distributions.

To facilitate distributed machine learning in wireless networks, over-the-air federated learning (AirFL) is proposed to provide data privacy protection and high communication efficiency by leveraging the superposition property of wireless channels. However, as a typical parameter attack method, Byzantine attack brings challenges to the stable operation of AirFL systems. In this letter, we integrate orthogonal frequency division multiplexing and SignSGD with majority vote to enhance the resilience of AirFL against Byzantine attacks by performing one-bit gradient quantization. Theoretical analysis and numerical simulations are provided to validate the effectiveness of the proposed AirFL scheme under different channel states and Byzantine attacker percentages.

Radio simultaneous localization and mapping (SLAM) is challenging due to multipath propagation. While line-of-sight (LoS) and first-order non-LoS (NLoS) paths, referred to as NLoS-1 paths, play a critical role in SLAM, no existing techniques can effectively separate them from high-order NLoS paths, i.e., NLoS-npaths (n≥ 2). This paper presents a new framework to accurately identify the LoS/NLoS-1 paths and conduct SLAM. The key idea is to define the virtual user equipment (UE) of a NLoS-npath as then-th order reflection of the UE. We discover that the centers of the circles encompassing the UE, a virtual UE associated with a LoS/NLoS-1 path, and each of some other virtual UEs are aligned in a line, if and only if those virtual UEs are all associated with NLoS-1 paths. Accordingly, we propose to identify the LoS/NLoS-1 paths using Hough transform-based line detection, and estimate the UE’s location and the environments with the identified LoS/NLoS-1 paths using maximum likelihood estimation and mean-shift clustering. We analytically confirm that the localization error asymptotically approaches the Cramér-Rao Lower Bound. Simulations show that our approach outperforms the state of the art in localization accuracy by up to 91.93%, even when the latter assumed all NLoS-1 paths are perfectly identifieda-priori.

Semantic communication (SemCom) has received considerable attention for its ability to reduce data transmission size while maintaining task performance. However, existing works mainly focus on analog SemCom with simple channel models, which may limit its practical application. To reduce this gap, we propose an orthogonal frequency division multiplexing (OFDM)-based SemCom system that is compatible with existing digital communication infrastructures. In the considered system, the extracted semantics is quantized by scalar quantizers, transformed into OFDM signal, and then transmitted over the frequency-selective channel. Moreover, we propose a semantic importance measurement method to build the relationship between target task and semantic features. Based on semantic importance, we formulate a sub-carrier and bit allocation problem to maximize communication performance. However, the optimization objective function cannot be accurately characterized using a mathematical expression due to the neural network-based semantic codec. Given the complex nature of the problem, we first propose a low-complexity sub-carrier allocation method that assigns sub-carriers with better channel conditions to more critical semantics. Then, we propose a deep reinforcement learning-based bit allocation algorithm with dynamic action space. Simulation results demonstrate that the proposed system achieves 9.7% and 28.7% performance gains compared to analog SemCom and conventional bit-based communication systems, respectively.

Due to the distributed deployment and inter-network dependence, distributed cloud network (DCN) is vulnerable to large-scale damage, making emergent system recovery of vital importance. Given limited resources at an early stage of network recovery, we propose a computation-aware link repair (CALR) algorithm to meet the computation demands of data centers in heavily damaged DCNs. Taking into account both network structure and traffic dynamics, we formulate a total system cost minimization problem to guarantee network repair performance. To tackle this challenging mixed-integer programming problem, we leverage the Benders decomposition (BD) to transfer it into an iteration problem with the mutually independent master problem and subproblem, which are solved by the cutting plane and the minimum cost flow algorithms, respectively. To accelerate the convergence speed of the proposed BD-based approach, we apply a small perturbation on the subproblem for facilitating the recovery of large-scale networks. Moreover, the computational complexity is reduced significantly by generating maximal non-dominated Benders cuts. Numerical simulations demonstrate that the proposed approach outperforms benchmarks under different settings such as network scale, data significance, available resources, and topology.

Aiming to circumvent the severe large-scale fading and the energy scarcity dilemma in high-altitude platform (HAP) networks, this paper investigates the benefits of the reconfigurable intelligent surface (RIS) and simultaneous wireless information and power transfer (SWIPT) on HAP communications. Specifically, we propose a concept of multi-layer refracting RIS-assisted receiver to achieve concurrent transmission of the information and energy, which is conducive to overcoming the severe fading effect induced by extreme long-distance HAP links and fully exploits RIS’s degrees-of-freedom (DoFs) for the SWIPT design. Based on the RIS-enhanced receiver, we then formulate a worst-case sum-rate maximization problem by considering the channel state information (CSI) error, the information rate requirements, and the energy harvesting constraint. To handle the intractable non-convex problem, a scalable robust optimization framework is proposed to obtain semi-closed-form solutions. Specifically, a discretization method is adopted to convert the imperfect CSI into a robust one. Then, by utilizing the LogSumExp inequality to smooth the objective and constraints, we develop a dual method to obtain the optimal solution for the HAP transmit precoder. In addition, a modified cyclic coordinate descent (M-CCD) is adopted to update the block-wise RIS coefficients. Moreover, closed-form solutions for power splitting (PS) ratios and the receive decoder are derived. Finally, the asymptotic performance of our proposed RIS-enhanced receiver is provided to reveal the substantial capacity gain for HAP communications. Numerical simulations demonstrate that the proposed architecture and optimization framework are capable of achieving superior performance with low complexity compared to state-of-the-art schemes in HAP networks.

Internet of Robotic Things (IoRT) emphasizes the integrated robotic, artificial intelligence computing, and communication technologies, enabling more sophisticated operations and decision-making. As a crucial element of IoRT, mission-critical applications, such as industrial manufacturing and emergency services, impose stringent requirements on ultra-reliable and low-latency communication (URLLC). The paper focuses on addressing URLLC challenges in the context of IoRT, particularly when autonomous mobile robots (AMRs) coexist with static sensors. We prioritize safe and efficient AMRs’ travel through trajectory design and communication resource allocation in IoRT systems without the need of any prior knowledge. To enhance network connectivity and exploit diversity gains, we introduce the flexible decoding and free clustering as the next-generation multiple access technologies in spectrum-limited downlink IoRT system. Then, aiming at minimizing the decoding error probability and travel time, we formulate a long-term multi-objective optimization problem by jointly designing AMRs’ trajectory and communication resource. To accommodate the inherent dynamics and unpredictability in the IoRT system, we introduce a multi-agent actor-critic deep reinforcement learning (DRL) framework, offering four distinct implementations, each accompanied by comprehensive complexity analyses. Simulation results reveal the following insights: 1) in terms of DRL implementations, off-policy algorithms with deterministic policies outperform their on-policy counterparts, achieving approximately a 67% increase in rewards; 2) In terms of communication schemes, our proposed flexible decoding and free clustering strategies under designed trajectories can effectively reduce decoding errors; and 3) In terms of algorithm optimality, our DRL framework shows superior flexibility and adaptability in communication environments compared to traditional A* search and heuristic methods.

Integrated sensing and communication (ISAC) opens up new service possibilities for sixth-generation (6G) systems, where both communication and sensing (C&S) functionalities co-exist by sharing the same hardware platform and radio resource. In this paper, we investigate the waveform design problem in a downlink multi-user and multi-target ISAC system under different C&S performance preferences. The multi-user interference (MUI) may critically degrade the communication performance. To eliminate the MUI, we employ the constructive interference mechanism into the ISAC system, which saves the power budget for communication. However, due to the conflict between C&S metrics, it is intractable for the ISAC system to achieve the optimal performance of C&S objective simultaneously. Therefore, it is important to strike a trade-off between C&S objectives. By virtue of the multi-objective optimization theory, we propose a weighted Tchebycheff-based transformation method to re-frame the C&S trade-off problem as a Pareto-optimal problem, thus effectively tackling the constraints in ISAC systems. Finally, simulation results reveal the trade-off relation between C&S performances, which provides insights for the flexible waveform design under different C&S performance preferences in MIMO-ISAC systems.

In this paper, we propose and study a multi-functional reconfigurable intelligent surface (MF-RIS) architecture. In contrast to conventional single-functional RIS (SF-RIS) that only reflects signals, the proposed MF-RIS simultaneously supports multiple functions with one surface, including reflection, refraction, amplification, and energy harvesting of wireless signals. As such, the proposed MF-RIS is capable of significantly enhancing RIS signal coverage by amplifying the signal reflected/refracted by the RIS with the energy harvested. We present the signal model of the proposed MF-RIS, and formulate an optimization problem to maximize the sum-rate of multiple users in an MF-RIS-aided non-orthogonal multiple access network. We jointly optimize the transmit beamforming, power allocations as well as the operating modes and parameters for different elements of the MF-RIS and its deployment location, via an efficient iterative algorithm. Simulation results are provided which show significant performance gains of the MF-RIS over SF-RISs with only some of its functions available. Moreover, we demonstrate that there exists a fundamental trade-off between sum-rate maximization and harvested energy maximization. In contrast to SF-RISs which can be deployed near either the transmitter or receiver, the proposed MF-RIS should be deployed closer to the transmitter for maximizing its communication throughput with more energy harvested.

In existing federated learning (FL), the base station (BS) coordinates devices to collaboratively train a shared model by avoiding the transmission of raw data. To achieve communication-efficient model uploading, over-the-air computation (AirComp) is often employed to aggregate model parameters. However, in conventional AirComp assisted FL, the BS’s abundant computation resources are underutilized due to its non-involvement in model training. Meanwhile, transmission failures resulting from fluctuating wireless channels impair the quality of model aggregation. In this paper, we propose a retransmission-based semi-federated learning (SemiFL) framework, wherein devices upload model parameters and public privacy-free data for enabling a hybrid implementation of FL and centralized learning (CL). In our new framework, the BS leverages its abundant computation resources to aid CL model training, which mitigates the resource wastage while alleviating local computational burden of devices. Moreover, the proposed new retransmission mechanism effectively overcomes detrimental transmission failures resulting from the fluctuating quasi-static channel, aiming to guarantee improved learning performance of SemiFL. Successful transmission probabilities of both retransmission-based AirComp and retransmission-based digital communication are provided in closed forms. To attain deep insights, we derive an optimality gap to capture the convergence behavior of retransmission-based SemiFL. Then, we formulate a non-convex long-term problem to minimize a weighted sum of overall latency and energy consumption by jointly optimizing communication, computation, and learning parameters. Extensive experimental results show that our retransmission-based SemiFL obtains 21.9%, 30.5%, and 44.1% accuracy gains on three datasets, while efficaciously reducing latency and energy consumption compared to benchmarks. Meanwhile, our scheme enhances learning performance on the fluctuating quasi-static channel compared to state-of-the-art schemes.

In this paper, we explore robust transceiver design for an integrated sensing and communications system with bounded channel estimation error. To maximize the minimum sensing performance of multiple targets while satisfying communications requirements, we study the worst-case robust op-timization problem by jointly optimizing the transmitter and receiver variables. The formulated problem is challenging due to the non-convexity arising from the semi-infinite constraints (SICs) and coupled variables. To overcome these difficulties, we adopt the S-procedure to convert the SICs into finite convex linear matrix inequalities (LMIs). Using the alternating opti-mization technique, we decouple the robust transceiver design problem into feasibility-checking subproblems. By exploiting matrix lifting, we transform the rank-one constraints into a set of LMIs, which is leveraged to further check the feasibility of the obtained beamforming scheme. Numerical results are provided to demonstrate the robustness and effectiveness of the proposed algorithm in combating channel errors and improving the performance of ISAC systems.

In this paper, we propose a semi-federated learning (SemiFL) framework to ameliorate the performance of conventional federated learning. The base station and devices are coordinated to collaboratively train a shared model. However, due to the rapidly fluctuating channels and irrationally assigned local learning workloads, SemiFL encounters excessive latency. To overcome the challenges, we propose a retransmission-based over-the-air computation mechanism to facilitate model aggregation and data mixing over quasi-static channels. The closed-form probability of successful aggregation is derived, while the communication latency is modeled based on the Pascal distribution. Further, we establish an optimality gap to characterize the convergence performance of SemiFL, wherein the minimum number of iterations for attaining a specific local target accuracy is identified. Next, a joint resource allocation and local target accuracy assignment problem is formulated to minimize the latency of each round, subject to the decay rate, central processing unit (CPU) frequency, and transmit power. To address this non-convex problem, we develop an algorithm using the closed-form solutions for the normalizing factors and CPU frequencies. Simulation results on two real-world datasets confirm the superiority of SemiFL over benchmarks in terms of latency and learning performance.

In this paper, we propose a semi-federated learning (SemiFL) framework that allows computing-limited clients to collaboratively train a shared model with resource-abundant clients. Specifically, by supporting the coexistence of model-updating and data-offloading, the SemiFL framework enables both centralized and federated learning in a hybrid fashion. Due to the decoding error, we consider the practical case with residual interference. To improve uplink throughput for centralized learning while reducing aggregation distortion for federated learning, we formulate a non-convex optimization problem to jointly optimize the transmit power and receive strategy. Then, we propose an efficient algorithm to solve the challenging problem by using successive convex approximation. Simulation results demonstrate the effectiveness of our SemiFL framework for heterogeneous networks, and reveal the impact of imperfect signal decoding on communication rates.

In this paper, we propose a new multi-functional reconfigurable intelligent surface (MF-RIS) architecture. Different from conventional RIS that only reflects signals, MF-RIS supports multiple functionalities on one surface, including reflection, transmission, amplification, and energy harvesting. As such, MF-RIS is capable of overcoming the double-fading attenuation and achieving full-space coverage by harvesting energy from the base station (BS). The physical implementation and the signal model of MF-RIS are introduced from the perspective of wireless communications. Then, we formulate a sum rate (SR) maximization problem in an MF-RIS-aided non-orthogonal multiple access network. By jointly optimizing the transmit strategy of the BS and the coefficient of the MF-RIS, we design an iterative algorithm to solve the formulated non-convex problem efficiently. Simulation results show that: i) MF-RIS provides up to 98.8% higher SR gain than self-sustainable RIS. ii) There exists a non-trivial trade-off between throughput improvement and self-sustainability, due to the limited number of RIS elements.

In industrial control system, applications perform time-critical operations based on the observations of multiple processes. We consider a request-based scenario, where a cache-enabled base station (BS) stores the most recent status observed by energy harvesting (EH) sensors, and delivers the cached status to applications upon request. Due to the time-varying nature of processes, cached status may be outdated which affects the accuracy of operations. Frequent cache update improves the status freshness but leads to high energy consumption of EH sensors. Furthermore, the freshness on the application side is simultaneously determined by multiple status, which requires joint updating of multiple sensors to improve update efficiency. Age of Information (AoI) is employed to measure the freshness of status. We adopt the maximum AoI among the responded status as age of response (AoR). A long-term average AoR minimization problem is formulated, subject to the number of wireless channels and the energy constraint of each EH sensor. The problem is challenging due to the random arrivals of requests and energy harvesting as well as the random association between requests and sensors. By introducing AoR reduction as the reward of each schedule, the problem is decomposed into a per-slot reward maximization problem, and then transformed into a knapsack problem. Then, an online correlated cache update algorithm is proposed. Numerical experiments illustrate that our solution outperforms the traditional greedy policy and achieves 16% performance gains.

Implementing existing federated learning in massive Internet of Things (IoT) networks faces critical challenges, such as imbalanced and statistically heterogeneous data and device diversity. To this end, we propose a semi-federated learning (SemiFL) framework to provide a potential solution for the realization of intelligent IoT. By seamlessly integrating the centralized and federated paradigms, our SemiFL framework shows high scalability in terms of the number of IoT devices even in the presence of computing-limited sensors. Furthermore, compared to traditional learning approaches, the proposed SemiFL can make better use of distributed data and computing resources, due to the collaborative model training between the edge server and local devices. Simulation results show the effectiveness of our SemiFL framework for massive IoT networks. The code can be found athttps://github.com/niwanli/SemiFL_IoT.

In this article, we propose a multi-functional reconfigurable intelligent surface (MF-RIS) to address the half-space coverage and double-fading attenuation issues faced by existing RISs. By simultaneously reflecting, refracting, and amplifying the incident signal, the proposed MF-RIS is capable of realizing full-space coverage with the mitigated signal degradation. The operation principle of the MF-RIS is first provided, and then an efficient beamforming scheme is proposed for MF-RIS-aided wireless communications. Simulation results show that, through combining multiple functions on one surface, the MF-RIS achieves significant throughput improvement over existing RISs.

Rapid and efficient sensor data acquisition plays a critical role in the decision-making process of each robot in a multi-robot smart factory. This paper investigates the trajectory design of autonomous mobile robots (AMRs) and communication resource allocation problems in industrial Internet of Things. Specifically, by exploiting both power and spatial domains, we adopt non-orthogonal multiple access to improve network connectivity in a spectrum-efficient manner, while the multi-antenna technique is employed to enhance diversity gain. The average sum rate is maximized by jointly optimizing the transmit power of sensors and the trajectory of AMRs. To deal with prior knowledge and dynamic channel conditions, we reformulate the long-term maximization problem as a Markov decision process, and further develop a provably efficient multi-agent reinforcement learning algorithm with a near-optimal regret bound. Our theoretical analysis reveals that both the decentralized execution and the experience exchange method are beneficial to accelerate convergence. Simulation results show that our proposed algorithm can reduce at least 80% convergence time compared to the centralized baseline, and can gain better rewards than the conventional$\epsilon $-greedy exploration.

Although reconfigurable intelligent surfaces (RISs) can improve the performance of wireless networks by smartly reconfiguring the radio environment, existing passive RISs face two key challenges, i.e., double-fading attenuation and dependence on grid/battery. To address these challenges, this paper proposes a new RIS architecture, called multi-functional RIS (MF-RIS). Different from conventional reflecting-only RIS, the proposed MF-RIS is capable of supporting multiple functions with one surface, including signal reflection, amplification, and energy harvesting. As such, our MF-RIS is able to overcome the double-fading attenuation by harvesting energy from incident signals. Through theoretical analysis, we derive the achievable capacity of an MF-RIS-aided communication network. Compared to the capacity achieved by the existing self-sustainable RIS, we derive the number of reflective elements required for MF-RIS to outperform self-sustainable RIS. To realize a self-sustainable communication system, we investigate the use of MF-RIS in improving the sum-rate of multi-user wireless networks. Specifically, we solve a non-convex optimization problem by jointly designing the transmit beamforming and MF-RIS coefficients. As an extension, we investigate a resource allocation problem in a practical scenario with imperfect channel state information. By approximating the semi-infinite constraints with the$\mathcal {S}$-procedure and the general sign-definiteness, we propose a robust beamforming scheme to combat the inevitable channel estimation errors. Finally, numerical results show that: 1) compared to the self-sustainable RIS, MF-RIS can strike a better balance between energy self-sustainability and throughput improvement; and 2) unlike reflecting-only RIS which can be deployed near the transmitter or receiver, MF-RIS should be deployed closer to the transmitter for higher spectrum efficiency.

Under the organization of the base station (BS), wireless federated learning (FL) enables collaborative model training among multiple devices. However, the BS is merely responsible for aggregating local updates during the training process, which incurs a waste of the computational resources at the BS. To tackle this issue, we propose a semi-federated learning (SemiFL) paradigm to leverage the computing capabilities of both the BS and devices for a hybrid implementation of centralized learning (CL) and FL. Specifically, each device sends both local gradients and data samples to the BS for training a shared global model. To improve communication efficiency over the same time-frequency resources, we integrate over-the-air computation for aggregation and non-orthogonal multiple access for transmission by designing a novel transceiver structure. To gain deep insights, we conduct convergence analysis by deriving a closed-form optimality gap for SemiFL and extend the result to two extra cases. In the first case, the BS uses all accumulated data samples to calculate the CL gradient, while a decreasing learning rate is adopted in the second case. Our analytical results capture the destructive effect of wireless communication and show that both FL and CL are special cases of SemiFL. Then, we formulate a non-convex problem to reduce the optimality gap by jointly optimizing the transmit power and receive beamformers. Accordingly, we propose a two-stage algorithm to solve this intractable problem, in which we provide closed-form solutions to the beamformers. Extensive simulation results on two real-world datasets corroborate our theoretical analysis, and show that the proposed SemiFL outperforms conventional FL and achieves 3.2% accuracy gain on the MNIST dataset compared to state-of-the-art benchmarks.

This paper proposes a novel concept of multi-functional reconfigurable intelligent surfaces (MF-RISs). Different from conventional single-functional RISs (SF-RISs) that only reflect signals, the proposed MF-RIS simultaneously supports multiple functionalities, namely, reflection, amplification, and energy harvesting. Specifically, by harvesting energy from incident signals, MF-RIS is able to simultaneously reflect and amplify signals without an external power supply, which is beneficial for overcoming the double-fading attenuation in a flexible manner. A new operation protocol of MF-RIS is presented, and then a sum rate (SR) maximization problem is formulated for an MF-RIS aided multi-user network. Next, an efficient iterative algorithm is proposed to solve this non-convex problem. Furthermore, through theoretical analysis, we determine the number of reflection elements required for MF-RISs to outperform self-sustainable RISs. Finally, our numerical results show that: 1) MF-RISs are able to provide up to 81.1% higher SR than the self-sustainable RISs. 2) Unlike the SF-RIS, which prefers to be deployed near the transmitter or receiver, MF-RISs should be deployed closer to the transmitter for better performance.

This paper proposes a novel semi-federated learning (SemiFL) paradigm, which integrates centralized learning (CL) and over-the-air federated learning (AirFL) into a unified framework, with the aid of a simultaneous transmitting and reflecting reconfigurable intelligent surface (STAR-RIS). In particular, this SemiFL framework allows computing-scarce users to participant in the learning process by using non-orthogonal multiple access (NOMA) to transmit their local dataset to the base station for model computation on behalf of them. During the uplink communication, scarce spectrum resources are shared among AirFL users and NOMA-based CL users, using a STAR-RIS for interference management and coverage enhancement. To analyze the learning behavior of SemiFL, closed-form expressions are derived to quantify the impact of learning rates and noisy fading channels. Our analysis shows that SemiFL can achieve a lower error floor than the CL or AirFL schemes with partial users. Simulation results show that SemiFL significantly reduces communication overhead and latency compared to CL, while achieving better learning performance than AirFL.

As a distributed machine learning paradigm, federated learning (FL) has been regarded as a promising candidate to preserve user privacy in Internet of Things (IoT) networks. Leveraging the waveform superposition property of wireless channels, over-the-air FL (AirFL) achieves fast model aggregation by integrating communication and computation via concurrent analog transmissions. To support sustainable AirFL among energy-constrained IoT devices, we consider that the base station (BS) adopts simultaneous wireless information and power transfer (SWIPT) to distribute global model and charge local devices in each communication round. To maximize the long-term energy efficiency (EE) of AirFL, we investigate a resource allocation problem by jointly optimizing the time division, transceiver beamforming, and power splitting in SWIPT-enabled IoT networks. Considering such multiple closely-coupled continuous valuables, we propose a deep reinforcement learning (DRL) algorithm based on twin delayed deep deterministic (TD3) policy to smartly make downlink and uplink communication strategies with the coordination between the BS and devices. Simulation results show that the proposed TD3 algorithm obtains about 41% EE improvement compared to traditional optimization method and other DRL algorithms.

Ultra-reliable low-latency communication (uRLLC) is envisioned to efficiently support mission-critical scenarios, especially for industrial Internet of Things (IIoT). Considering the requirements of high throughput and massive connectivity in machine-type communications, uRLLC traffic is usually coexisted with enhanced mobile broadband (eMBB) services for the data-intensive industrial cases. To strike a balance between the two distinct tasks, this paper investigates a multi-objective optimization problem by taking into account the performance of both uRLLC and eMBB. Specifically, we aim at maximizing eMBB data rate and uRLLC reliability, while minimizing the communication overhead of control channels caused by uRLLC puncturing. To solve this challenging problem, an analytic hierarchy process method is adopted to estimate the importance of each objective with expert knowledge. Then, a coalitional game is invoked to evaluate the preference degree of resource blocks allocated to uRLLC devices. Following this, an improved Gale-Shapley algorithm is proposed for efficient traffic scheduling. Simulation results demonstrate that the proposed algorithm can achieve better performance in terms of eMBB throughput and uRLLC reliability with the reduced signal overhead.

By exploiting the superiority of nonorthogonal multiple access (NOMA), NOMA-aided mobile-edge computing (MEC) can provide scalable and low-latency computing services for the Internet of Things. However, given the prevalent stochasticity of wireless networks and sophisticated signal processing of NOMA, it is critical but challenging to design an efficient task offloading algorithm for NOMA-aided MEC, especially under a large number of devices. This article presents an online algorithm that jointly optimizes offloading decisions and resource allocation to maximize the long-term system utility (i.e., a measure of throughput and fairness). Since the optimization variables are temporary coupled, we first apply Lyapunov technique to decouple the long-term stochastic optimization into a series of per-slot deterministic subproblems, which does not require any prior knowledge of network dynamics. Second, we propose to transform the nonconvex per-slot subproblem of optimizing NOMA power allocation equivalently to a convex form by introducing a set of auxiliary variables, whereby the time-complexity is reduced from the exponential complexity to$\mathcal {O} (M^{3/2})$. The proposed algorithm is proved to be asymptotically optimal, even under partial knowledge of the device states at the base station. Simulation results validate the superiority of the proposed algorithm in terms of system utility, stability improvement, and the overhead reduction.

This article integrates nonorthogonal multiple access (NOMA) and over-the-air federated learning (AirFL) into a unified framework using one simultaneous transmitting and reflecting reconfigurable intelligent surface (STAR-RIS). The STAR-RIS plays an important role in adjusting the decoding order of hybrid users for efficient interference mitigation and omnidirectional coverage extension. To capture the impact of nonideal wireless channels on AirFL, a closed-form expression for the optimality gap (also known as the convergence upper bound) between the actual loss and the optimal loss is derived. This analysis reveals that the learning performance is significantly affected by the active and passive beamforming schemes, as well as wireless noise. Furthermore, when the learning rate diminishes as the training proceeds, the optimality gap is explicitly shown to converge with a linear rate. To accelerate convergence while satisfying quality-of-service requirements, a mixed-integer nonlinear programming (MINLP) problem is formulated by jointly designing the transmit power at users and the configuration mode of STAR-RIS. Next, a trust-region-based successive convex approximation method and a penalty-based semidefinite relaxation approach are proposed to handle the decoupled nonconvex subproblems iteratively. An alternating optimization algorithm is then developed to find a suboptimal solution for the original MINLP problem. Extensive simulation results show that: 1) the proposed framework can efficiently support NOMA and AirFL users via concurrent uplink communications; 2) our algorithms achieve a faster convergence rate on independent and identically distributed (IID) and non-IID settings compared to the existing baselines; and 3) both the spectrum efficiency and learning performance are significantly improved with the aid of the well-tuned STAR-RIS.

The fundamental communication paradigms in the next-generation mobile networks are shifting from connected things to connected intelligence. The potential result is that current communication-centric wireless systems are greatly stressed when supporting computation-centric intelligent services with distributed big data. This is one reason that makes federated learning come into being, it allows collaborative training over many edge devices while avoiding the transmission of raw data. To tackle the problem of model aggregation in federated learning systems, this article resorts to multiple reconfigurable intelligent surfaces (RISs) to achieve efficient and reliable learning-oriented wireless connectivity. The seamless integration of communication and computation is actualized by over-the-air computation (AirComp), which can be deemed as one of the uplink nonorthogonal multiple access (NOMA) techniques without individual information decoding. Since all local parameters are uploaded via noisy concurrent transmissions, the unfavorable propagation error inevitably deteriorates the accuracy of the aggregated global model. The goals of this work are to 1) alleviate the signal distortion of AirComp over shared wireless channels and 2) speed up the convergence rate of federated learning. More specifically, both the mean-square error (MSE) and the device set in the model uploading process are optimized by jointly designing transceivers, tuning reflection coefficients, and selecting clients. Compared to baselines, extensive simulation results show that 1) the proposed algorithms can aggregate model more accurately and accelerate convergence and 2) the training loss and inference accuracy of federated learning can be improved significantly with the aid of multiple RISs.

With the aim of integrating over-the-air federated learning (AirFL) and non-orthogonal multiple access (NOMA) into an on-demand universal framework, this paper proposes a reconfigurable intelligent surface (RIS)-aided hybrid network by leveraging the RIS to flexibly adjust the decoding order of heterogeneous data. A new metric of computation rate is defined to measure the performance of AirFL users. Upon this, the objective of this work is to maximize the achievable hybrid rate by jointly optimizing the transmit power, controlling the receive scalar, and designing the reflection coefficients. Since the concurrent transmissions of all computation and communication signals are aided by the discrete phase-shifting elements at the RIS, the formulated problem (P0) is a challenging mixed-integer programming problem. To tackle this intractable issue, we decompose the original problem (P0) into a non-convex problem (P1) and a combinatorial problem (P2), which are characterized by the continuous and discrete variables, respectively. For the transceiver design problem (P1), the power allocation subproblem is first solved by difference-of-convex programming, and then the receive control subproblem is addressed by successive convex approximation, where the closed-form expressions of simplified cases are derived to obtain deep insights. For the reflection design problem (P2), a relaxation-then-quantization method is adopted to find a suboptimal solution for striking a trade-off between complexity and performance. Afterwards, an alternating optimization algorithm is developed to solve the non-linear non-convex problem (P0) iteratively. Finally, simulation results reveal that i) the proposed RIS-aided hybrid network can support on-demand communication and computation efficiently, ii) the system performance can be improved by properly selecting the location of the RIS, and iii) the designed algorithms are also applicable to conventional networks with only AirFL or NOMA users.

Over-the-air federated learning (AirFL) allows devices to train a learning model in parallel and synchronize their local models using over-the-air computation. The integrity of AirFL is vulnerable due to the obscurity of the local models aggregated over the air. This paper presents a novel framework to balance the accuracy and integrity of AirFL, where multi-antenna devices and base station (BS) are jointly optimized with a reconfigurable intelligent surface (RIS). The key contributions include a new and non-trivial problem jointly considering the model accuracy and integrity of AirFL, and a new framework that transforms the problem into tractable subproblems. Under perfect channel state information (CSI), the new framework minimizes the aggregated model’s distortion and retains the local models’ recoverability by optimizing the transmit beamformers of the devices, the receive beamformers of the BS, and the RIS configuration in an alternating manner. Under imperfect CSI, the new framework delivers a robust design of the beamformers and RIS configuration to combat non-negligible channel estimation errors. As corroborated experimentally, the novel framework can achieve comparable accuracy to the ideal FL while preserving local model recoverability under perfect CSI, and improve the accuracy when the number of receive antennas is small or moderate under imperfect CSI.

This paper proposes a new, compatible, unified framework which integrates non-orthogonal multiple access (NOMA) and over-the-air federated learning (AirFL) via concurrent communication. In particular, a simultaneous transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) is leveraged to adjust the signal processing order for efficient interference mitigation and omni-directional coverage extension. With the aim of investigating the impact of non-ideal wireless communication on AirFL, we provide a closed-form expression for the optimality gap over a given number of communication rounds. This result reveals that the learning performance is significantly affected by the resource allocation scheme and channel noise. To minimize the derived optimality gap, a mixed-integer non-linear programming (MINLP) problem is formulated by jointly designing the transmit power at users and configuration mode at the STAR-RIS. Through developing an alternating optimization algorithm, a suboptimal solution for the original MINLP problem is obtained. Simulation results show that the learning performance in terms of training loss and test accuracy can be effectively improved with the aid of the STAR-RIS.

Robots have been widely used in the era of Inter-net of Things (IoT) and become an important driver behind 6G systems. However, due to the mobility of robots and the complicated wireless environment, it is challenging to control robots without prior knowledge, especially for the dynamic multi-robot systems. This paper investigates data transmission and path planning problems in an indoor multi-robot navigation system, where non-orthogonal multiple access (NOMA) is adopted to provide massive connectivity and improve spectrum efficiency. The considered system is designed for the long-term throughput maximization by jointly designing the downlink transmit power at the access point (AP) and the motion paths of the robots, while satisfying the power budget and mobility constraints. In this paper, a novel federated deep reinforcement learning approach called federated deep Q-network learning (F-DQN) algorithm is proposed to tackle the formulated problem. Simulation results demonstrate that the proposed algorithm not only speeds up the convergence rate and promotes the system throughput, but also is scalable for the number of robots.

This paper investigates the problem of secure communication in the unmanned aerial vehicle (UAV) enabled net-works aided by a reconfigurable intelligent surface (RIS) from the physical layer security perspective. Specifically, the RIS is deployed to assist the wireless transmission from the UAV to the ground user in the presence of an eavesdropper. The objective of this work is to maximize the secrecy rate by jointly optimizing the phase shifts at the RIS as well as the transmit beamforming vector and location of the UAV. However, the formulated problem is difficult to solve directly due to the non-linear and non-convex objective function and constraints. By invoking the successive convex approximation and fractional programming techniques, the intractable original problem is transformed into convex ones, then an alternating algorithm is proposed to solve the challenging problem effectively. Simulations results demonstrate that the designed algorithm for RIS-aided UAV communications can achieve higher secrecy rate than benchmarks.

This paper investigates the model aggregation process in an over-the-air federated learning (AirFL) system, where an intelligent reflecting surface (IRS) is deployed to assist the transmission from users to the base station (BS). Since the successive interference cancellation (SIC) is adopted as a basis to decode local model parameters and analyze their statistic characteristics for detecting malicious devices, the quality-of-service (QoS) requirement is ensured. The objective of this paper is to minimize the mean-square-error by jointly optimizing the receive beamforming vector at the BS, transmit power allocation at users, and phase shift matrix of the IRS, subject to the transmit power constraint for devices, unit-modulus constraint for reflecting elements, SIC decoding order constraint and QoS constraint. To address this complicated problem, alternating optimization is employed to decompose it into three subproblems, where the optimal receive beamforming vector is obtained by solving the first subproblem with the Lagrange dual method. Then, the convex relaxation method is applied to the transmit power allocation subproblem to find a suboptimal solution. Eventually, the phase shift matrix subproblem is addressed by invoking the semidefinite relaxation. Simulation results validate the availability of IRS and the effectiveness of the proposed scheme in improving federated learning performance.

This article proposes a novel framework of resource allocation in multi-cell intelligent reflecting surface (IRS) aided non-orthogonal multiple access (NOMA) networks, where an IRS is deployed to enhance the wireless service. The problem of joint user association, subchannel assignment, power allocation, phase shifts design, and decoding order determination is formulated for maximizing the achievable sum rate. The challenging mixed-integer non-linear problem is decomposed into an optimization subproblem (P1) with continuous variables and a matching subproblem (P2) with integer variables. In an effort to tackle the non-convex optimization problem (P1), iterative algorithms are proposed for allocating transmission power, designing reflection matrix, and determining decoding order by invoking relaxation methods such as convex upper bound substitution, successive convex approximation, and semidefinite relaxation. In terms of the combinational problem (P2), swap matching-based algorithms are developed for achieving a two-sided exchange-stable state among users, BSs and subchannels. Numerical results demonstrate that: i) the sum rate of multi-cell NOMA networks is capable of being increased by 35% with the aid of the IRS; ii) the proposed algorithms for multi-cell IRS-aided NOMA networks can enjoy 22% higher energy efficiency than conventional NOMA counterparts; iii) the trade-off between spectrum efficiency and coverage area can be tuned by judiciously selecting the location of the IRS.

Due to their high mobility, flexible deployment and stable maneuverability, unmanned aerial vehicles (UAVs) have been deemed as a promising and indispensable role for various emerging applications (e.g., dangerous area detection, dynamic target tracking, and map remote sensing). Compared to the static monitoring equipments, UAV-mounted high-definition camera and signal transceiver can be used cost-effectively as an on-demand aerial platform to detect the unknown region and send the real-time data back at the same time. However, these highlighted limitations of battery capacity and communication resource extremely affect the UAV’s performance such as flight endurance and data transmission. Motivated by the above conflicts, this paper aims to minimize the total energy consumed by the UAV during the region detection mission through jointly optimizing the collected data size, transmission time, and flying trajectory. Toward this end, we derive the optimal data collection and transmission time in closed forms via convex optimization, and propose a model-free reinforcement learning-based algorithm for training the UAV to plan its trajectory without knowing the environment information in advance. Simulation results validate the performance of our designs in terms of convergence, energy consumption, and energy efficiency.

Mobile edge computing (MEC), which enables the wireless devices to offload their computation tasks to the edge servers, has been considered as a promising technology to offer low-latency computing services and prolong lifetime for the Internet of Things (IoT) devices. To further improve the system efficiency, non-orthogonal multiple access (NOMA) is exploited to the MEC system. In this paper, we investigate a NOMAbased MEC system in ultra dense network and pursue an energy efficient offloading strategy for resource-limited devices, while meeting the execution latency constraint Due to the non-convexity of the optimization problem, we propose a low complexity energy efficient task offloading algorithm based on alternating direction method of multipliers (ADMM) decomposition technique to transform it into multiple parallel convex subproblems and obtain the optimal solution. Numerical results show that the proposed scheme can significantly decrease energy consumption for user and achieve lower complexity.

For providing highly demanding services with powerful computational ability and ultra low-latency communication, mobile edge computing (MEC) has been recognized as a bright rising star among key technologies for the next-generation networking. Generally, jointly optimizing offloading decision and resource allocation in one multi-variable problem is complicated. To decrease computational scale and develop practicable strategy by splitting problems, we divide the workflow of MEC-enabled base station into two stages. First, through formulating a task offloading problem, we propose a low-complexity improved simulated annealing-based heuristic offloading decision (SAHOD) algorithm to maximize network revenue from the perspective of mobile network operator. Then, the optimal fine-grained resource allocation solution is obtained in closed forms via Lagrange duality decomposition method. Furthermore, an effective realtime sub-gradient-based resource allocation (SGRA) algorithm is presented to converge to a specific optimal allocation strategy within the adjustable accuracy. For given users, simulation results show that our SAHOD algorithm can earn about 20.5% more revenue than value-based greedy algorithm. Besides, our SGRA algorithm can converge within 4 iterations and obtain approximately 19.3% more sum rates than static scheduling method.