Mobile edge computing-space-air-ground integrated network (MEC-SAGIN) is emerging as a crucial component of future wireless systems. Despite its potential, addressing network fluctuations while ensuring continuous low-latency computing services in highly dynamic environments remains a significant challenge. To address this issue, this article proposes a fluid antenna (FA)-assisted MEC-SAGIN system, which enhances channel transmission conditions and reduces uplink task offloading latency by flexibly adjusting the antenna ports of edge computing users equipped with FAs. Specifically, we aim to minimize the maximum total computational delay (TCD) of edge computing tasks for ground users (GUs) and the satellite user (SU) by jointly optimizing the task offloading strategies, computational resource allocation, FA port positions, uncrewed aerial vehicle (UAV) location, and the receive beamforming matrix. To solve this nonconvex problem, we employ the block coordinate descent (BCD) technique to decompose the original problem into four subproblems. The subproblems are optimized using a combination of low-complexity iterative algorithms and the projected gradient descent (PGD) method to refine communication and computation configurations as well as FA port selection. Simulation results demonstrate that the FA-assisted scheme significantly improves the TCD performance of the MEC-SAGIN system. It maintains transmission stability and reliability in dynamic environments while outperforming conventional fixed-position antennas (FPAs) and random-port antenna schemes.

Integrated satellite-unmanned aerial vehicle (UAV)-terrestrial network (ISUTN) is promising to provide high maneuverability, versatile deployment capabilities, and pervasive connectivity for future wireless systems. However, limited frequency band and the coexistence of multi-layer communications bring new challenges for interference management. In this paper, we propose a robust multi-layer interference management scheme for spectrum sharing in ISUTN with imperfect channel state information at the transmitter (CSIT) of the low earth orbit (LEO) satellite. In the proposed scheme, hierarchical rate-splitting multiple access (H-RSMA) is utilized to mitigate interference and reduce the total energy consumption (TEC) of the system. Specifically, we investigate the TEC minimization problem by jointly optimizing the UAV receiver (UR) grouping, LEO beamforming, UAV power allocation, UAV trajectory, and rate allocation subject to the quality of service (QoS) and transmission power constraints. A robust alternating algorithm integrating weighted K-means (WK), semi-definite programming (SDP), and successive convex approximation (SCA) methods is proposed to handle the non-convexity of the original problem while ensuring TEC minimization. Numerical results validate the effectiveness and robustness of the proposed H-RSMA scheme under diverse network loads and CSIT uncertainties. Benefiting from the interference management capability, H-RSMA saves significant TEC compared to several benchmarks.

The integration of Unmanned Aerial Vehicles (UAVs) with Mobile Edge Computing (MEC) expands coverage and flexibility in low-altitude networks, enabling next-generation Ultra-Reliable Low-Latency Communications (URLLC). However, deep fading and air-to-ground channel fluctuations increase task transmission latency and degrade offloading reliability. To overcome these challenges, we propose an online resource management scheme deploying Fluid Antennas (FAs) in a UAV-enabled MEC network to mitigate propagation impairments and enhance the Quality of Service (QoS). Specifically, we aim to minimize the total energy consumption of MEC users and the UAV while meeting FA port constraints, UAV speed limits, per-user QoS guarantees, and queue-stability requirements, thereby delivering enhanced URLLC service. By leveraging Lyapunov optimization, the original stochastic problem is transformed into a series of deterministic per-slot online optimization subproblems. A Block Coordinate Descent (BCD) algorithm then decomposes each subproblem into three tractable components. Closed-form solutions are derived for computing resource allocation and uplink time-slot assignments, while UAV trajectory and FA port selection are optimized via low-complexity Successive Convex Approximation (SCA) and Projected Gradient Descent (PGD) methods. Simulation results show that the proposed algorithm significantly reduces energy consumption compared with existing benchmarks, maintains queue stability, and delivers superior URLLC performance in multi-task and dynamic scenarios. (c) 2025 The Author(s). Published by Elsevier Ltd on behalf of Chinese Society of Aeronautics and Astronautics. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

This paper investigates the delay minimization problem for multi-task federated learning (MTFL) systems at the network edge. We develop a novel MTFL framework, based on which the divergence bounds are derived for both task-related and task-unrelated scenarios, systematically quantifying the effects of user importance, user participation, and inter-task correlations on convergence behavior. Building upon these insights, a long-term joint optimization problem is formulated to minimize the overall training delay under the long-term divergence bounds and energy constraints. To address the coupling in multi-slot user scheduling, the optimization problem is decomposed into a single-slot joint resource allocation and task scheduling subproblem and a cross-slot user scheduling subproblem. The former is solved using block coordinate descent (BCD) combined with Johnson’s rule, while the latter is modeled as a constrained Markov decision process (CMDP) and addressed via a dueling double deep Q-network (D3QN) with cost shaping and prioritized experience replay. Numerical results verify the effectiveness of the proposed framework and convergence analysis, demonstrating its significant improvements over baseline schemes in terms of convergence and delay reduction.

Split inference divides a global deep neural network (DNN) into several sub-models and assigns them to different computing nodes, thereby leveraging distributed computational resources at network edge for executing complex artificial intelligence (AI) algorithms. As an emerging technique, over-the-air computation (AirComp) turns a multi-access channel into a processor for distributed computing. Specifically, by exploiting the waveform superposition of the multi-access channel, a receiver directly aggregates signals transmitted by different transmitters to obtain their average (or any nomographic function). In order to accelerate inference speed and reduce communication and computation loads, we propose a novel in-network AI framework that realizes inference through communications among devices. This distinctive feature of the framework is to generalize multiple-input-multiple output (MIMO) AirComp to realize over-the-air matrix-vector multiplications (MVM), the most computation intensive operations of a DNN. As a result, the participating devices act like neurons in the DNN but they are now linked through computation capable wireless channels, which gives the name of Over-the-Air Neural Link (AirNeuralink). The proposed AirNeuralink techniques are associated with different system topologies including relay, star, and distributed topologies. Their design involves jointly optimizing the precoding and post-equalization to minimize the MVM errors of the estimated feature values at the receiver set with respect to their ground truth under the transmit power constraint for each transmitter. The optimal post-equalization is derived in the form of Wiener filter but with an aggregation matrix (weight matrix). Utilizing the property of Schur-concave function and matrix inequality, the optimal structure of precoding in each topology is designed to enable spatial-channel power allocation by efficiently solving a convex problem with scalar variables. Besides, the asymptotic MVM errors are analyzed to show that the errors can be sufficiently small if the rank of model-weight matrix is no larger than that of channel matrix. Simulations demonstrate the superior performance of the proposed framework in transmission latency reduction while maintaining the inference accuracy as compared with traditional digital transmission.

This paper proposes an over-the-air computation (AirComp)-assisted split federated learning (SFL) framework for mobile edge networks. Specifically, heterogeneous devices are assigned device-specific split points, such that each device executes the layers up to its split point locally and offloads the remaining layers to the edge server, while the split submodels are aggregated over the air. To guarantee reliable learning under noisy aggregation, we impose a mean-squared error constraint to control the aggregation distortion. We formulate a joint optimization problem over split-point selection, communication and computation resource allocation, and develop an efficient alternating algorithm that combines Karush-Kuhn-Tucker (KKT)-based updates with the interior point method to convex subproblems. Simulation results demonstrate that, under various signal-to-noise ratios, data distributions, and computing capabilities, the proposed scheme consistently achieves faster convergence and higher test accuracy than baselines.

Orthogonal time frequency space (OTFS) modulation is promising for low Earth orbit (LEO) satellite systems to combat mobility-induced channel variations. However, existing researches are confined to single-frame, single-antenna architectures, overlooking the spatio-temporal correlations inherent in the equivalent delay-Doppler (DD) domain channel in multi-frame LEO satellite multiple-input multiple-output (MIMO)-OTFS systems. Therefore, in this paper, we first establish a multi-frame LEO satellite MIMO-OTFS system with transmitter-windowed under fractional Doppler, derive the input-output relationship, and formulate a structured sparse signal recovery problem. Then, we propose an expectation propagation-based spatio-temporal channel estimation algorithm (EP-ST), which employs a spike-and-slab prior embedded in a hierarchical Gaussian process to exploit the spatio-temporal correlations of the channel, while efficiently tackling high-dimensional posterior intractability via the EP framework. Simulation results demonstrate that the proposed method effectively captures channel characteristics and substantially improves estimation accuracy.

Orthogonal time frequency space (OTFS) modulation is promising for low-Earth-orbit (LEO) satellite systems due to its resilience to mobility-induced channel variations. However, existing estimators fail to explicitly model the non-uniform block-sparse structure induced by fractional Doppler shifts in the effective delay- Doppler (DD) domain channel of LEO satellite OTFS systems. In this paper, we first derive an element- wise input–output relationship that accurately characterizes this structure. Then, we introduce a novel noise-adaptive pattern-coupled variational Bayesian inference algorithm for sparse Bayesian learning (termed NNCP-VSBL), which effectively exploits the non-uniform block sparsity while maintaining robustness against noise. Simulation results demonstrate that the proposed method efficiently captures the channel characteristics and improves the estimation accuracy of the LEO satellite OTFS channels.

The high mobility in low-earth-orbit (LEO) satellite communications results in severe Doppler spreads, for which orthogonal delay-Doppler (DD) division multiplexing (ODDM) modulation has emerged as a promising solution. In this paper, we investigate the Tomlinson-Harashima precoding (THP) scheme for downlink multi-user (MU) multiple-input single-output (MISO) ODDM transmissions. This scheme facilitates simplified equalization, and numerical results confirm its superior performance in mitigating the inter-symbol coupling in the DD domain. In addition, we characterize the achievable sum-rate of the proposed scheme by establishing an information-theoretic equivalent model. In particular, we derive the scaling laws for the achievable rate with respect to the number of antennas and users. Numerical simulations corroborate the accuracy of our theoretical derivation.

Orthogonal delay-Doppler division multiplexing (ODDM) modulation offers a promising solution to the problem of severe Doppler effects in low earth orbit (LEO) satellites communications. It has been suggested in the recent literature that ODDM modulation can extract full delay-Doppler (DD) diversity, yet a rigorous analysis has not been presented. In this paper, we present a formal analysis of the DD diversity achieved by ODDM modulation along with supporting simulations. Specifically, the analysis and simulations reveal that the asymptotic DD diversity order of ODDM modulation is one, and this order is achieved at lower bit error rate (BER) values for increased frame sizes. We also present low-complexity detector for ODDM modulation based on the Orthogonal approximate message passing (OAMP) algorithm, and show that this detector extracts full DD diversity with the reduced complexity.

This paper investigates mobile edge computing (MEC)-aided multi-static integrated sensing and communication (ISAC) systems. In the considered system, each device offloads a task to the MEC server for computation. Meanwhile, a sensing receiver (SR) has to detect a target. To enhance the detection capabilities, the devices and base station (BS) collaboratively transmit radar signals, and the SR detects the target based on the received radar echoes. Specifically, the task execution period is divided into two phases. In the first phase, the devices offload part of their tasks while simultaneously transmitting radar signals for target sensing. In the second phase, the MEC server at the BS processes the offloaded tasks, and meanwhile devices along with the BS transmit radar signals for sensing. Thus, the system entities form a multi-static ISAC framework. Radar signal processing schemes are designed for both scenarios—with and without knowledge of the target response amplitudes. The corresponding cooperative detection probabilities under a constant false alarm probability constraint are derived. To ensure high energy efficiency, optimization problem is formulated to minimize energy consumption across all devices, subject to task computation and detection probability constraints. This is achieved by adjusting beamforming at the devices and BS, as well as task partitioning and phase duration allocation. For scenarios with knowledge of target response amplitudes, the problem is efficiently solved by iteratively optimizing beamforming subproblem, time and task division subproblem using weighted minimum mean square error method, successive convex approximation, and golden section search method. For scenarios without knowledge of target response amplitudes, the formulated bi-level optimization problem is first equivalently transformed into a single-level problem, which is then tackled using alternating optimization method, semidefinite programming, and Lagrangian dual method. Simulations validate the benefits of the proposed MEC-aided cooperative sensing scheme compared with numerous benchmarking schemes including traditional non-cooperative sensing scheme.

Reconfigurable Intelligent Surface (RIS) is applied to physical layer security to enhance the security of task offloading in mobile edge computing (MEC) systems. However, due to the “double fading” effect on the reflection link, the deployment of passive RIS can only achieve limited performance gains. In view of these, active RIS is regarded as a prospective candidate for compensating for the “double fading” effect of passive RIS due to its low-power integrated amplifiers. This paper proposes a novel secure MEC system, which improves the communication and computation performance of the system by introducing active RIS. Specifically, we aim to minimize the weighted total energy consumption (TEC) in an active RIS-assisted MEC system, subject to the constraints of secure offloading and the phase shift of the active RIS. To address this non-convex problem, block coordinate descent (BCD) and successive convex approximation (SCA) methods are employed to alternately optimize the subproblems. Under the premise of ensuring the secure transmission rate for users, the computation offloading volume, the edge computing resource allocation, beamforming at the receiver, and reflection coefficient matrix at the active RIS are jointly optimized. Simulation results demonstrate that the active RIS significantly enhances the weighted TEC performance in MEC systems while ensuring secure transmission rates, effectively mitigating the impact of the “double fading” effect, and showing notable advantages over existing passive RIS and non-RIS schemes.

The Orthogonal Time Frequency Space (OTFS) system, has brought new opportunities and challenges. However, the OTFS system is also troubled by the problem of high Peak-to-Average Power Ratio (PAPR). In this paper, the research mainly focuses on the problem of high PAPR in the OTFS system, and a joint DCT-SLM algorithm based on the OTFS system is proposed. This algorithm first uses the Discrete Cosine Transform (DCT) to replace the Inverse Symplectic Finite Fourier Transform (ISFFT) and the Heisenberg transform in the traditional OTFS system, which simplifies the modulation that originally required two steps into one step. Secondly, the Selected Mapping (SLM) algorithm is used to further reduce the PAPR of the OTFS system. The simulation results show that when the Complementary Cumulative Distribution Function (CCDF) of the PAPR drops to 10-3 while keeping a similar Bit Error Rate (BER), the PAPR threshold can be reduced by about 1.5 to 1.8 dB.

This paper investigates the joint optimal allocation of radio and computational resources aiming to minimize global average task offloading age (TOA) over all time slots and mobile devices (MDs) for long-term multi-cell MEC systems with continuous arrival of MDs. TOA represents the total number of offloading time slots, including both transmission and computation. The joint resource allocation problem cannot be solved online because its objective function is long-term average of TOA over all time slots. We transform the long-term resource allocation problem into an online one by the Lyapunov method, then an iterative algorithm is proposed to solve the online problem. The idea of this algorithm is computing iteratively the two sub-problems which optimize sub-channel allocation and offloading power and computational resources joint allocation based on an initial resource allocation scheme. The alternating direction method of multipliers (ADMM) method is employed to solve the first sub-problem. For the second sub-problem, a closed-form expression of optimal power is deduced by solving a convex optimization problem using the Lagrange multiplier method, then the sub-problem is simplified into a linear programming (LP) problem about computational resource allocation. The improved iterative greedy (IIG) algorithm is applied to solve the LP problem. Simulation results demonstrate that the proposed algorithm approaches the performance of the optimal branch-and-bound (BnB) algorithm in the MEC systems with one-time arrival of MDs, and outperforms two benchmark schemes such as first in first out (FIFO) and Chang’s algorithm.

This paper addresses the requirement for ambiguity processing in the localization of wireless communication source node by proposing a particle swarm optimization (PSO)-based beamforming design method that maximizes the squared position error bound (SPEB) for the monitoring party. First, we derive the SPEB of TOA/TDOA localization methods for source node equipped with array antennas. Subsequently, with the objective of maximizing the SPEB of the monitoring party’ s localization performance, an iterative beamforming optimization algorithm based on PSO is developed. Numerical simulations validate the effectiveness of the proposed method.

Single carrier frequency division multiple access (SC-FDMA) is a multicarrier modulation technique known for its low peak-to-average ratio (PAPR) and high spectral efficiency. However, SC-FDMA is more sensitive to carrier frequency offset than single carrier systems. In this paper, a method is proposed to analyse the frequency offset tolerance of SC-FDMA systems under different constellation modulations, and to use it as a theoretical guide for selecting the frequency offset estimation method. An none-data-aided (NDA) method is proposed and the estimation performance of data-aided (DA) and NDA methods are compared. It is found that the proposed carrier frequency offset synchronization method can satisfy the packet error rate (PER) performance with a difference of about 0.2 dB from the ideal case.

With the dramatically increasing number of antennas, wireless channel becomes sophisticated and the pilot overhead becomes intolerable. In this letter, we construct the last-bounce cluster (LBC) based channel model to capture the spatial-temporal characteristics in extra-large multiple-input-multiple-output (XL-MIMO) systems. Then, we formulate the channel prediction problem to conserve the pilots. Leveraging the advantages of deep learning (DL), we propose convolutional neural network (CNN)-Transformer based channel prediction method (CTCP) to enhance the spatial-temporal information extraction of the channel. Simulation results validate that CTCP effectively extracts the spatial-temporal correlations of the channel, and enhances the trade-off between accuracy and comlexity in channel prediction for XL-MIMO systems.

In this paper, we investigate a joint communication and computation resource allocation strategy for an unmanned aerial vehicle (UAV)-assisted mobile edge computing (MEC) system employing fluid antenna (FA). Specifically, each user is equipped with an FA to offload the entire computation tasks to the MEC server deployed on the UAV. By dynamically selecting antenna ports, users can achieve latency-efficient edge computing services, especially advantageous in dynamic environments. To minimize the maximum execution delay of all the users, we jointly optimize the UAV location, FA port selection, and computation resource allocation, subject to computational capacity constraints. The original non-convex optimization problem is decomposed into three tractable subproblems within a block coordinate descent (BCD) algorithm. The optimal computing frequencies are derived in closed form, while the UAV location and FA port selection are optimized using low-complexity iterative algorithms based on successive convex approximation (SCA) and linear programming (LP) techniques. In addition to conventional benchmarks with fixed-position antennas (FPAs), we also introduce a reconfigurable intelligent surface (RIS)-assisted system as a comparative baseline. Simulation results demonstrate that the proposed FA-assisted scheme significantly outperforms both FPAs and RIS-assisted counterparts, with performance gains becoming more pronounced in multi-task and highly dynamic scenarios, establishing FA-assisted UAV-MEC as a promising solution for future deployments.

This paper focuses on minimizing the total energy consumption of a long-term delay-sensitive multi-cell mobile edge computing (MEC) system that serves continuously arriving mobile devices (MDs). The energy consumption minimization is achieved by jointly optimizing the task offloading proportions, transmit power allocations, and computational resource distributions while ensuring the overall deadline constraints and the minimum processing size requirements in each scheduling cycle. The optimization problem is then formulated as a multi-agent Markov decision process (MAMDP) to enable sequential optimization across multiple scheduling cycles. To efficiently solve the formulated problem, we develop a multi-agent deep reinforcement learning (MADRL) algorithm that integrates the actor-critic (AC) framework, the embedding techniques, and the centralized training and decentralized execution (CTDE) framework. Simulation results show that the proposed algorithm converges 14%-23% faster than benchmark methods and significantly outperforms benchmark methods in reducing the total energy consumption under specific constraints by up to 10%.

A robust frequency minimum mean square error (MMSE) estimator based on the weighted average of multipath signal for OTFS systems is proposed in this paper. Previous frequency estimators are based on the strongest single path signal or equally weighted average of multiple path signals so that the accuracies of estimators are limited. We generalise the weight coefficients to be taken arbitrarily, and their optimal values are determined such that the MSE of estimator is minimum. Simulation results show that the proposed method outperforms existing frequency estimators in terms of MSE performance, particularly when the normalized carrier frequency offset (CFO) is less than 0.5 and the symbol signal-to-noise ratio (SNR) of received signal is below 2 dB. Under static channel conditions, the proposed estimator can achieve a reduction in pilot energy of approximately 21.6 dB.