ABSTRACT This study proposes a federated split learning framework for large language models (FedsLLM) integrated with rate‐splitting multiple access (RSMA), aimed at enhancing the efficiency and privacy of LLM training in wireless communication systems. By leveraging low‐rank adaptation (LoRA) to distribute computational loads and a fluid antenna system to dynamically optimize channel capacity, the framework effectively reduces training latency through joint optimization of learning accuracy and communication resources. Experimental results demonstrate that the proposed framework significantly outperforms traditional time‐division multiple access including time division multiple access, frequency division multiple access (FDMA), enhanced bandwidth FDMA, and fairness‐enhanced FDMA across multiple metrics: at a transmit power of 20 dBm, RSMA reduces task completion time by 8.3%; under 20 MHz bandwidth, it achieves a 25% performance improvement; and even with a data volume of 900 Kbits, it maintains a 12% advantage. The adopted alternating optimization algorithm converges rapidly, reaching 95% of the optimal value within only 5 iterations, substantially outperforming the fixed‐point method. Overall, FedsLLM‐RSMA effectively addresses privacy, computational and communication bottlenecks in distributed LLM training. Compared to TDMA, it reduces total training latency by 28% and improves communication efficiency by 35%, while achieving higher model accuracy and faster convergence. This work provides a viable pathway for efficient and scalable deployment of LLMs in 6G networks.

In this paper, we investigate an uplink multi-modal probabilistic semantic communication (PSCom) system based on probability graph in the satellite scenario. The system consists of both: semantic computation and traditional communication. Firstly, at the user end, the transmitted data is compressed based on the probabilistic graph. Then, the compressed data is transmitted to the satellite, which uses the same probabilistic graph to recover the received data. In the considered model, this paper addresses an optimization problem for multi-modal multi-user semantic communication across multiple time slots. An optimization problems formulated aiming to minimize the total energy consumption of the PSCom system, with satisfying the transmission time, transmission power, transmission bandwidth, local computation frequency, and transmission data requirements. To solve this problem, the Lyapunov drift-plus-penalty function based on online optimization is first used to transform the multi-slot problem into a stochastic single-slot problem, thereby converting the optimization problem into a trade-off between system energy consumption and queue length. Subsequently, an alternating algorithm is proposed to iteratively optimize, semantic compression rate, local computation frequency, transmission bandwidth, transmission power, and time allocation variables. Finally, simulation experiments demonstrates the effectiveness of the proposed algorithm.

Physical layer security (PLS) provides information-theoretic confidentiality for next-generation wireless systems, yet its practical deployment is fundamentally constrained by the sensitivity and noise characteristics of classical radio receivers. Recent advances in quantum sensing, particularly Rydberg atom-based receivers (RAQRs), present a promising approach to address these limitations. This paper introduces a Rydberg atomic quantum multiple-input multiple-output (RAQ-MIMO) architecture for secure multi-user uplink communications, where an array of Rydberg vapor cells functions as a high-sensitivity, reconfigurable receiving front-end. We develop a comprehensive end-to-end signal and channel model that incorporates quantum sensor response, wireless fading, and eavesdropper behavior under passive interception. To maximize the achievable secrecy rate, we propose a joint optimization framework that adaptively tunes key system degrees of freedom, including the Rabi frequency of the local oscillator, phase configuration, digital combining matrix, and user transmit power. Extensive simulations validate that RAQ-MIMO substantially outperforms classical MIMO in terms of secrecy rate across varying transmission distances, power budgets, and user loads—particularly under low-signal to noise ratio (SNR) and extended-range regimes. Our work establishes a quantum-sensing-assisted security foundation for future high-assurance wireless networks, bridging atomic physics with scalable communication system design with scalable communication system design under narrowband operation.

In this paper, we propose an uplink multi-modal probabilistic semantic communication (PSCom) system that considers both communication and computation. In the considered PSCom model, the users and base station share the common knowledge, which is characterized by probability graph. Based on the shared probability graph, the original large-size semantic data is compressed into the small-size semantic information, which will introduce additional computation costs for semantic compression. Besides, semantic information recovered by the base station will also bring additional computation costs. Although this method incurs additional computation costs, it effectively reduces communication energy consumption. Based on the considered model, an optimization problem is formulated to minimize the total communication and computation energy consumption, considering latency, power budget, bandwidth, and semantic compression ratio constraints. To solve this mixed integer optimization problem, we first adopt a greedy algorithm for the semantic compression level selection of each model, thereby transforming the original problem into a continuous variable optimization problem. Then, derive the optimal solution of the communication power. Finally, the Lagrange multiplier method is used to determine the optimal bandwidth, which can result in a closed-form optimal solution. Simulation results validate the effectiveness of the proposed algorithm.

This paper proposes a framework for integrating sensing, energy, and communication (ISEAC) in a single system. Specifically, the base station (BS) transmits the same signals to the information receivers (IRs), energy harvesting receivers (ERs), and sensing nodes. We aim to optimize the transmission beamforming to minimize the Cramér-Rao bound (CRB) while guaranteeing the performance of the IRs and ERs. We then solve this optimization problem by employing the semidefinite relaxation (SDR) method and Gaussian randomization algorithm. We provide a suboptimal solution with the closed-form solution for the special case with a single IR and a single ER. Finally, simulation results demonstrate the effectiveness of the proposed optimal beamforming design.

This paper investigates the performance of simultaneous transmission and reflection reconfigurable intelligent surface (STAR-RIS) assisted massive multiple-input multiple-output (MIMO) systems with direct links. We apply the two-timescale scheme to design the base station (BS) beamforming and the phase shifts of the STAR-RIS. Specifically, we derive the closed-form expression of the average achievable rate. Based on the derived rate, we theoretically draw insights from comparing STAR-RIS and conventional RIS under the same condition. Then, we optimize the phase shifts of the STAR-RIS using accelerated gradient ascent algorithm. Finally, numerical results are provided to validate our theoretical insights that STAR-RIS outperforms the conventional RIS under some special cases.

Abstract A transmission power minimization problem for a reconfigurable intelligent surface (RIS) aided non‐orthogonal multiple access (NOMA) network has been studied here, which has assumed the channel state information is imperfect. The channel error in this paper has been modeled as two different error models, that is, bounded error model and statistical error model. First, a two‐step algorithm is proposed to decouple the decoding order and optimization variables under bounded error model. Then, for the statistical error model, the alternating optimization method is adopted to decouple optimization variables. Finally, the performance difference is compared between the proposed beamforming scheme and the non‐RIS scheme from two key parameters, the number of reflected elements and the error factor. Simulation results verify that passive RIS has the advantage on the system performance compared with the scheme without RIS.

This paper investigates the performance of simultaneous transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) assisted massive multiple-input multiple-output (MIMO) systems with Rician fading channels and channel estimation errors. We adopt the two-timescale scheme to design the systems, namely, applying the instantaneous channel state information (CSI) to design the base station (BS) beamforming and leveraging the statistical CSI to design the phase shifts of the STAR-RIS. Specifically, we estimate the overall channels based on the linear minimum mean-squared error (LMMSE) estimator and derive the closed-form expression of the average achievable rate. Based on the derived rate, we analyze the power scaling laws in which the transmit power is respectively reduced inversely proportional to the number of BS antennas and STAR-RIS elements. Besides, we draw insights from the comparison between STAR-RIS and conventional RIS under the same condition and the power scaling laws of STAR-RIS and optimize the phase shifts of the STAR-RIS to maximize the sum rate using an accelerated gradient ascent-based algorithm. Finally, numerical results are provided to validate our theoretical insights. In particular, we also compare the two-timescale scheme with the instantaneous CSI scheme in the simulation. We show that STAR-RIS outperforms conventional RIS, and the two-timescale-based scheme outperforms the instantaneous CSI-based scheme. Furthermore, we draw insight into this phenomenon.

This paper investigates the performance of a two-timescale transmission design for uplink reconfigurable intelligent surface (RIS)-aided cell-free massive multiple-input multiple-output (CF-mMIMO) systems. We consider the Rician channel model and design the passive beamforming of RISs based on the long-time statistical channel state information (CSI), while the central processing unit (CPU) utilizes the maximum ratio combining (MRC) technology to perform fully centralized processing based on the instantaneous overall channel, which is the superposition of the direct and RIS-reflected channels. Firstly, we derive the closed-form approximate expression of the uplink achievable rate for arbitrary numbers of access point (AP) antennas and RIS reflecting elements, which can be used to obtain energy efficiency through the proposed total power model. Relying on the derived expressions, we theoretically analyze the impact of important system parameters on the rate and draw explicit insights into the benefits of RISs. Then, based on the rate expression under statistical CSI, we optimize the phase shifts of RISs by using the genetic algorithm (GA) to maximize the sum rate and minimum rate of users, respectively. Finally, the numerical results demonstrate the correctness of our expressions and the benefits of deploying large-size RISs into cell-free mMIMO systems. Also, we investigate the optimality and convergence behaviors of the GA to verify its effectiveness. To give a more beneficial analysis, we present numerical results to show the high energy efficiency of the system with the help of RISs. Besides, our results have revealed the benefits of distributed deployment of APs and RISs in the RIS-aided mMIMO system with cell-free networks.

To combat the half-space coverage and enhance the flexibility of the reconfigurable intelligent surface (RIS) technology, a simultaneously transmitting and reflecting RIS (STAR-RIS) is applied in the radar-communication co-existence (RCC) system, where the signal through STAR-RIS is transmitted to opposite spaces, and STAR-RIS is utilized to handle the interference from the base station (BS) to the radar. A radar detection probability maximization problem by optimizing the transmit beamforming vector of the BS and the transmission-and reflection-coefficient matrices of the STAR-RIS is formulated, subject to the power constraint of the BS and the communication rate constraints of users. The problem is challenging to solve due to the highly coupled variables. We convert it into two sub-problems and propose an efficient alternating optimization (AO) algorithm to solve this non-convex problem. The simulation results validate the convergence of the proposed algorithm and the performance advantages of using STAR-RIS over conventional RIS.

Abstract This paper focuses on a reconfigurable intelligent surface (RIS)‐aided full‐duplex multi‐user massive multiple‐input multiple‐output system with transceiver hardware impairments (THWIs). Different from the existing works, the two‐timescale design scheme is considered. The phase shift at the RIS is designed only based on the statistical channel state information. Firstly, the closed‐form expression of the uplink and downlink achievable rate is derived. Then, the power scaling laws are revealed. Finally, the impact of THWIs on the system performance is analysed and genetic algorithm to maximize the achievable rate is used.

Abstract This paper studies the beamforming design for an active reconfigurable intelligent surface (RIS) aided non‐orthogonal multiple access (NOMA) network. It aims to minimize the transmission power while satisfying the requirement of the quality of service (QoS). To solve this problem, an iterative algorithm based on approximate transformations and penalty functions is proposed. Finally, simulation results illustrate that an active RIS‐aided NOMA system can achieve significant performance gains over the passive RIS‐aided NOMA network.

This paper studies the worst-case robust beamforming design for a reconfigurable intelligent surface (RIS) aided non-orthogonal multiple access (NOMA) network with imperfect channels. We aim to minimize the transmission power while satisfying the requirement of the worst-case quality of service (QoS). With the worst-case QoS constraints, unit-modulus constraints and imperfect channel state information (CSI), this problem is a non-convex optimization problem. To solve this problem, we propose a two-procedure algorithm by applying penalty function and semidefinite relaxation (SDR). Finally, simulation results illustrate that the RIS-aided NOMA system has better performance than the traditional NOMA system.

In this study, the authors study the achievable rate of the full‐duplex massive multiple‐input multiple‐output (MIMO) relaying system with low‐resolution analogue‐to‐digital converters (ADCs) over Rician fading channels, where maximum ratio combining/maximum ratio transmission are adopted. They first investigate the approximate rate with perfect channel state information by considering the effect of the quantisation error, the loop interference (LI) and the Rician K ‐factor. Then, they analyse the relationship between the achievable rate and several important variables in the Rician fading channel. It is shown that a large number of antenna arrays at the relay can compensate for the rate loss caused by quantisation at the relay end, and the quantisation bits at the destination have a greater influence on the rate than the quantisation bits at the relay end. In particular, as the number of antennas increases, the rate determined by the quantisation bits at the destination tends to a finite value. They also conclude that the Rician K ‐factor is proportional to the rate and the attenuation of the rate is very obvious when the LI is large. Finally, the simulation results are presented to verify their analysis.

Different from conventional key‐based cryptography schemes, physical‐layer security (PLS) techniques have drawn much attention recently to realise unconditional security from the information theory perspective. As an important performance metric in PLS, the ergodic secrecy rate (ESR) for a multi‐input multi‐output wireless communication network over a Nakagami fading channel is analysed. The network is consisted of a multi‐antenna transmitter (Alice), a multi‐antenna legitimate receiver (Bob), and a multi‐antenna eavesdropper (Eve). By using the selective transmission (ST) at Alice and the maximum ratio combining (MRC) at Bob and Eve, an exact expression of the ESR is derived. However, due to the infinite summation, it is very hard to evaluate the ESR performance. To reduce computational complexity and obtain more insights, a lower bound of the ESR is then obtained, which is in a closed form. As special cases, the lower bounds of the ESR for the signal‐antenna scenario and Rayleigh fading channel are also obtained, respectively. Numerical results show that the derived expressions of the ESR and its lower bound are very accurate to evaluate system performance.

In this paper, we investigate the physical-layer security for a spatial modulation (SM)-based indoor visible light communication (VLC) system, which includes multiple transmitters, a legitimate receiver, and a passive eavesdropper (Eve). At the transmitters, the SM scheme is employed, i.e., only one transmitter is active at each time instant. To choose the active transmitter, a uniform selection (US) scheme is utilized. Two scenarios are considered: one is with non-negativity and average optical intensity constraints and the other is with non-negativity, average optical intensity, and peak optical intensity constraints. Then, lower and upper bounds on the secrecy rate are derived for these two scenarios. Besides, the asymptotic behaviors for the derived secrecy rate bounds at high signal-to-noise ratio (SNR) are analyzed. To further improve the secrecy performance, a channel adaptive selection (CAS) scheme and a greedy selection (GS) scheme are proposed to select the active transmitter. Numerical results show that the lower and upper bounds of the secrecy rate are tight. At high SNR, small asymptotic performance gaps exist between the derived lower and the upper bounds. Moreover, the proposed GS scheme has the best performance, followed by the CAS scheme and the US scheme.

We study the uplink achievable rate of full-duplex (FD) multiuser multi-input multi-output (MIMO) systems with a base station (BS) and users over Rician fading channels, based on maximum ratio combining (MRC) receivers. Contrary to previous related works over Rayleigh fading channels, this paper assumes that the fast fading MIMO channel matrix follows the Rice distribution which will be more common in future 5G wireless communication systems. A lower bound expressions of the uplink achievable rate for perfect channel state information (CSI) can be derived when the number of BS antennas grows large. Based on the theoretical analysis, it is found that when the antennas of the BS are large enough and the power scaling law is applied properly, the impact of multi-user interference (MUI), loop interference (LI) and noise can be suppressed. In addition, the simulation results show that the uplink sum achievable rates increase with the number of BS antennas and they will converge to fixed values with the increasing Rician K-factor.

As a novel statistical model, Málaga distribution is proposed recently, which covers many commonly used fading models as special cases. The secure transmission of information over the Málaga fading channels has not been discussed in literature. In this paper, the secrecy outage probability (SOP) over the Málaga fading channels is investigated. Initially, an exact expression of the SOP is derived, which is with an integral term and can not be easily used in practice. To reduce its computational complexity, a closed-from expression for the lower bound of the SOP is then obtained. Numerical results that illustrate the effect of the Málaga fading on system performance are provided. The accuracy of the performance analysis is verified by simulations.

In this paper, we study widely linear precoding techniques to mitigate in-phase/quadrature-phase (IQ) imbalance (IQI) in the downlink of large-scale multiple-input multiple-output (MIMO) systems. We adopt a real-valued signal model, which considers the IQI at the transmitter, and then develop widely linear zero-forcing (WL-ZF), widely linear matched filter, widely linear minimum mean-squared error, and widely linear block-diagonalization (WL-BD) type precoding algorithms for both single- and multiple-antenna users. We also present a performance analysis of WL-ZF and WL-BD. It is proved that without IQI, WL-ZF has exactly the same multiplexing gain and power offset as ZF, while when IQI exists, WL-ZF achieves the same multiplexing gain as ZF with ideal IQ branches, but with a minor power loss, which is related to the system scale and the IQ parameters. We also compare the performance of WL-BD with BD. The analysis shows that with ideal IQ branches, WL-BD has the same data rate as BD, while when IQI exists, WL-BD achieves the same multiplexing gain as BD without IQ imbalance. Numerical results verify the analysis and show that the proposed widely linear type precoding methods significantly outperform their conventional counterparts with IQI and approach those with ideal IQ branches.

This paper presents a unified, optimization-driven solution for designing IIR and FIR notch filters with prescribed, possibly varying notch levels in the given stop-bands, and near unit magnitude frequency response at the pass-bands. Although the original IIR notch filter optimization problem is non-convex, we show that it can be well approximated by a convex problem, by replacing a non-positive semi-definite 2 × 2 Hermitian matrix with its nearest positive semi-definite counterpart. With this approach, the IIR filter design can be efficiently solved via Newton iteration. The same approach can be directly applied to the FIR filter design since it is a degenerated case of the IIR filter. Moreover, we show that the FIR design problem is convex and therefore can be solved optimally. Numerical examples are presented to verify the effectiveness of the proposed design.

In this paper, we consider the problem of maximizing the weighted sum energy efficiency (WS-EE) of a multi-input single-output (MISO) ad hoc network. To solve this problem, we develop one low-complexity distributed beamforming algorithm based on pricing mechanism. Specifically, each node updates its price information and broadcasts it to all the other nodes. Having collected all the information, each node selects its beam-vector in closed-form with low computational complexity. The convergence of this algorithm is strictly proved. Compared with the existing two-layer optimization algorithm, our algorithm has lower computational complexity without performance loss. Simulation results show that the proposed algorithm performs slightly worse than the centralized algorithm, but requires much less information exchange overhead.

This paper studies the impact of hardware mismatch (11M) between the base station (BS) and the user equipment (UE) in the downlink (DL) of large-scale antenna systems. Analytical expressions to predict the achievable rates are derived for different precoding methods, i.e., matched filter (MF) and regularized zero-forcing (RZF), using large system analysis techniques. Furthermore, the upper bounds on achievable rates of MF and RZF with 11M are investigated, which are related to the statistics of the circuit gains of the mismatched hardware. Moreover, we present a study of 11M calibration, where we take zero-forcing (ZF) precoding as an example to compare two 11M calibration schemes, i.e., Pre-precoding Calibration (Pre-Cal) and Post-precoding Calibration (Post-Cal). The analysis shows that Pre-Cal outperforms Post-Cal schemes. Monte-Carlo simulations are carried out, and numerical results demonstrate the correctness of the analysis.