With the advent of the sixth-generation (6G) era, the rapid growth of access points (APs) and service types in space–air–ground integrated networks (SAGIN) has posed significant challenges to efficient user association and joint resource allocation in multi-layer heterogeneous environments. To address this issue, we develop a three-layer SAGIN slicing framework composed of ground base stations (BSs), unmanned aerial vehicles (UAVs), and low earth orbit (LEO) satellites, where a shared UAV resource pool is introduced to enhance aerial resource coordination. A weighted multi-objective optimization problem is formulated to jointly support high-throughput, low-latency, and wide-coverage services. Due to its non-convex nature, a Stackelberg game pricing (SGP)-based multi-agent deep reinforcement learning (MADRL) framework, namely MATD3-SGP, is proposed. where the original problem is decomposed into user association and resource allocation subproblems. Specifically, a stackelberg game pricing mechanism is designed for user association, and the uniqueness of the stackelberg equilibrium is proven. while the resource allocation problem is modeled as a partially observable Markov decision process (POMDP), and a multi-agent twin delayed deep deterministic policy gradient (MATD3)-based slicing scheme with centralized training and decentralized execution is developed. Simulation results show that the proposed MATD3-SGP framework consistently outperforms benchmark schemes in terms of system utility, throughput, latency, and coverage ratio. In particular, the system utility is improved by approximately 12.19%, 26.32%, 75.37%, and 147.74% compared with MADDPG-SGP, MASAC-SGP, MAPPO-SGP, and the hard-slicing method, respectively.
Integrated sensing and communication (ISAC) is expected to play a key role in future sixth-generation (6G) networks, where random data-bearing signals are reused to support both communications and sensing, thereby improving time-frequency utilization. In practice, the high peak-to-average power ratio (PAPR) of conventional multicarrier signals incurs severe power-amplifier (PA) back-off and limits both communication throughput and sensing range. Constant-envelope orthogonal frequency division multiplexing (CE-OFDM) is a promising remedy due to its 0-dB PAPR. However, its random nonlinear phase mapping produces high autocorrelation sidelobes, significantly degrading ranging performance. In this paper, we develop an expectation-oriented autocorrelation-function (ACF) design framework for pulse-shaped random CE-OFDM signals. We derive a tractable analytical approximation for the average squared ACF, which decomposes into a pulse-induced pedestal determined by the shaping filter and a phase-dependent leakage term controlled by time-domain phase rotation. Based on this characterization, we formulate an expected integrated sidelobe level (EISL) minimization problem and propose a joint design of phase rotation and Nyquist pulse shaping. We further show that coherent integration across independent transmission slots suppresses residual leakage proportionally to the integration length. Numerical results demonstrate that the proposed design achieves pronounced sidelobe suppression over the prescribed delay interval while preserving the ultra-low PAPR, favorable BER performance, and spectral efficiency of CE-OFDM.
This paper proposes non-iterative nonlinear equalization (NLE) techniques under a novel minimum mean squared error (MMSE) criterion based on amplitude-phase-frequency (APF) characterization. The APF characterization based MMSE (APF-MMSE) objective is to minimize the sum of mean squared magnitude (SoMSM) of the error between the APF characteristic function of the equalized nonlinear channel and the linear characteristic basis function without requiring higher-order spectral statistics. The performance of the proposed APF-MMSE NLE is analyzed and compared with its zero-forcing counterpart for both equalizer structures modelled by expanded memory polynomial (EMP) and orthogonal EMP (OEMP) respectively. Low-complexity pseudo-APF-MMSE implementation suitable for practical systems is also investigated. Simulation and experimental results obtained from a 73.5 GHz millimeter wave system with 2.125 GHz bandwidth validate the findings from this paper and highlight a future research direction to achieve true-APF-MMSE equivalent performance under practical constraints on complexity and channel estimation accuracy.
This paper introduces a successive interference cancellation (SIC)-free uplink rate-splitting multiple access (RSMA) framework with dual-polarized multiple-input and multiple-output (DP-MIMO) to overcome decoding complexity and error propagation. Different from traditional uplink RSMA, the BS manages inter-user and intra-user interference by exploiting the DP antenna architecture, which eliminates the requirement of SIC. To validate the efficacy of the proposed framework, we compare the finite-blocklength (FBL) achievable sum spectral efficiency (SE) of DP-MIMO without SIC (DP-noSIC), single-polarized MIMO without SIC (SP-noSIC) and SP-MIMO with a local-SIC strategy (SP-localSIC) over correlated Rician channels. Specifically, we first develop the minimum mean-square error (MMSE) channel estimator for correlated Rician DP-MIMO channels and obtain the channel estimates and the error covariance matrices. Then, with the instantaneous signal-to-interference-plus-noise ratio (SINR), the sum SE maximization problem is established for the three schemes under FBL constraints. To solve the formulated problem, a unified algorithm framework is developed for efficient combiner design and power control. Simulation results reveal the performance superiority of the proposed DP-noSIC scheme, which attains the highest sum SE compared to the SP-noSIC and SP-localSIC schemes under different system setups. The observed outcomes demonstrate the effectiveness and robustness of DP-noSIC in improving throughput and reducing complexity.
Stacked intelligent metasurfaces (SIMs) have emerged as a promising paradigm for wave-domain signal processing in next-generation wireless networks. In this paper, we investigate the joint optimization of antenna selection, SIM phase-shift design, and power allocation in SIM-assisted multiuser multiple-input single-output (MU-MISO) systems to maximize the achievable downlink sum rate. To handle the resulting mixed discrete-continuous constraints, we introduce differentiable reparameterizations and derive analytical gradients for the three coupled variable blocks under the cascaded multilayer propagation model. Building on these physics-aware gradient directions, we propose a model-driven unrolled full-gradient optimization (UFGO) network that unfolds the analytical joint-gradient update process into a trainable architecture with a fixed number of unfolded layers. Each unfolded layer performs learnable block-specific updates guided by the physics-aware analytical gradients and employs differentiable reparameterizations to preserve variable feasibility, thereby enabling efficient online inference while retaining the interpretability of model-based optimization. Simulation results show that UFGO achieves sum-rate performance comparable to the high-iteration FGJO benchmark while requiring substantially lower runtime than the high-iteration AO and FGJO methods. Comparisons with the corresponding variants without antenna selection further demonstrate the performance benefit of adaptive antenna selection. Additional evaluations under electromagnetic propagation model mismatch, imperfect channel state information, and frequency-selective fading demonstrate the performance resilience of UFGO under practical model and channel uncertainties, supporting its applicability to latency-sensitive SIM configuration.
Federated learning (FL) can safeguard the data privacy for air-ground networks, while simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS) can help further enhance its channel quality. However, the model exchanges in FL may be detected by wardens, and air-ground links increase this threat. In this paper, we investigate a STAR-RIS aided air-ground covert FL scheme to hide the presence of uploading, where the unmanned aerial vehicle (UAV) acts as an aerial server to collect local models from the distributed devices. To alleviate the communication burden, each device prunes the local model before transmitting. The minimum detection error probability at the warden is derived with the optimal detection threshold, indicating the worst case for the legitimate covert FL. Then, we analyze the convergence of the proposed FL scheme and provide an upper bound on the expected convergence gap of global loss function. To improve the learning performance, the convergence gap is minimized subject to the covertness and latency constraints. Specifically, the closed-forms of optimal pruning ratio and transmit power are derived, while the phase shifts of STAR-RIS and UAV position can be optimized by semidefinite relaxation and successive convex approximation, respectively. Finally, the overall problem can be solved via an alternative algorithm. Numerical results validate that the proposed covert air-ground FL scheme can achieve excellent test accuracy and convergence performance while guaranteeing the covertness assisted by STAR-RIS.
Cyclic redundancy check-aided tail-biting convolutional codes (CRC-TBCCs) have shown strong potential for hyper-reliable and low-latency communications (HRLLC), particularly due to their excellent performance in short block lengths. However, their ability to support variable input lengths remains underexplored. This work addresses that gap by introducing a novel framework for flexible-length encoding using CRC-TBCC. The proposed scheme combines a fixed-length cyclic redundancy check (CRC) with bit repetition, enabling the encoder to accept variable-length input sequences while maintaining a fixed output length. Two novel techniques are proposed to fully exploit the potential of CRC. First, we prioritize the repetition of CRC bits. Second, the CRC bits, including their repetitions, are evenly multiplexed with the information bits. To support the proposed encoding method, we introduce a tailored pre-processed list Viterbi decoding algorithm. In the proposed decoder, preprocessing is first applied to enhance the reliability of the edges of the TBCC trellis by exploiting the constraint of repetition. After pre-processing, the serial list Viterbi algorithm (SLVA) is then executed to identify the candidate paths, which are subsequently checked by both the repetition and CRC constraints.We also propose an erasure-aided pre-processed SLVA to further improve the decoding performance. Extensive simulation results demonstrate the effectiveness of the proposed repetition and multiplexing techniques and the two-stage decoder. In particular, compared to existing flexible coding schemes, the proposed multi-rate CRC-TBCCs with the pre-processed SLVA show notable performance gains. These results make CRC-TBCC a promising solution for 6G HRLLC, where high flexibility and high performance are both required.
Deploying uncrewed aerial vehicles (UAVs) in vehicular networks overcomes the inherent limitations of terrestrial infrastructure by dynamically enhancing coverage and line-of-sight (LoS) connectivity. However, further enhancement of the data rate necessitates the deployment of additional terrestrial base stations (BSs) and UAVs, which incur high costs and consume a large amount of energy. To address these challenges, synergistically integrating UAVs equipped with intelligent reflecting surfaces (IRS) and multiple-input multiple-output (MIMO) techniques can effectively improve the channel capacity of air-to-ground integrated vehicular networks, while maintaining low cost and low power consumption. The key idea is to partially replace some of the required terrestrial relays with high-flexibility and energy-efficient UAVs equipped with IRS. Specifically, in IRS-MIMO air-to-ground integrated vehicular networks, we formulate a joint precoding design, phase shift optimization, and UAV deployment problem with the objective of maximizing the weighted sumrate. To tackle this non-convex problem, an iterative optimization algorithm with polynomial-time complexity is developed, which enables a gradual approximation of a feasible solution to the formulated problem. Finally, simulation results demonstrate that the proposed scheme is superior to the state-of-the-art schemes in terms of the weighted sum-rate. Additionally, the impact of network parameters on transmission performance and the convergence of iterative optimization are thoroughly analyzed.
Symbiotic radio (SR) is a promising technology for next-generation wireless networks, yet its broadcast nature exposes it to critical security risks, particularly from passive eavesdroppers. To overcome the reliance of conventional physical-layer security (PLS) schemes on impractical prior information about the eavesdropper’s location, this paper proposes an integrated sensing and communication (ISAC)-enhanced secure SR framework. By exploiting the backscattering characteristics of secondary transmitters, we endow the SR system with passive sensing capabilities to localize passive eavesdroppers. Specifically, we design a hybrid time division multiple access (TDMA) and code-domain superposition strategy for multi-backscatter device (BD) scenarios. This protocol coordinates assistive BDs to passively modulate orthogonal variable spreading factor (OVSF) sequences, which act as structured artificial noise (AN) to jam the eavesdropper while simultaneously establishing separable observation paths for localization. Furthermore, we formulate a joint transmit and receive beamforming optimization problem to maximize the secondary secrecy throughput, subject to the constraints on the primary system’s throughput and sensing Cramér-Rao lower bound (CRLB). To solve the resulting non-convex problem, we develop an algorithm based on Riemannian manifold optimization and alternating method. The simulation results demonstrate that the proposed framework effectively translates the sensing gains into PLS advantages, significantly outperforming existing benchmarks.
Simultaneously transmitting and reflecting reconfigurable intelligent surfaces (STAR-RISs) can dynamically adjust wireless channels through joint controls. The recent development of dual STAR-RISs (D-STAR) architecture provides capacity improvement by utilizing signals impinging from 360-degree for full-plane service coverage. In this paper, we consider an active D-STAR (AD-STAR) architecture in a full-duplex (FD) multi-input-multi-output (MIMO) system. Each AD-STAR element comprises a power amplifier and an on-off controller. We aim to maximize energy efficiency (EE) by optimizing transmit beamforming of the base station (BS) and uplink users as well as the AD-STAR configurations, while ensuring quality-of-service (QoS) for both downlink/uplink. We design a joint optimization of transmit beamforming and AD-STAR configurations (JOTA) algorithm. By adopting Dinkelbach’s and Lagrangian dual transformation, we partition the problem into three sub-problems and utilize successive convex approximation, convex upper bound, abstract Lagrangian duality, difference of two concave functions approximation, and penalty convex-concave programming methods to solve these sub-problems. Our numerical results verify the effectiveness of the proposed AD-STAR architecture in FD systems. The proposed JOTA scheme achieves at least 50.27% higher EE than the considered algorithmic benchmarks under the same AD-STAR architecture and adopted simulation settings.
The combination of unmanned aerial vehicles (UAVs) and reconfigurable intelligent surfaces (RISs) forms a cooperative framework that exploits the mutual advantages of both technologies. In this work, multiple RISs mounted on fixed-wing UAVs are adopted to adaptively reconfigure the transmission environments of communications. Given the practical significance of the proposed research, this study incorporates the phase-dependent amplitude response under realistic discrete phase shifts in RISs and accounts for statistical channel state information (CSI). In view of the practical designs of fixed-wing UAVs’ mobilities, the flight turning and pitch angles are considered in it to enhance the application value. Specifically, an optimization framework is designed to maximize the average data rate under statistical CSI over the UAVs’ flight duration by simultaneously optimizing BS’s transmit precoding, practical phase shifts of RISs, UAVs’ 3D trajectories, velocities, accelerations, and flight timeslot durations. To address this highly coupled and non-convex problem, an algorithm based on penalty dual decomposition (PDD) is introduced, incorporating weighted minimum mean square error (WMMSE) and successive convex approximation (SCA). The non-convex objective function is initially converted into a more manageable form by using WMMSE and analysis under statistical CSI. Then, variables are decoupled via auxiliary variables and alternately optimized in the inner iteration process, while operating the update of penalty and dual factors in the outer iteration process. Simulation results demonstrate the effectiveness of the proposed method.
Intelligent transportation systems face dual challenges in achieving efficient synergy between sensing and edge computing, arising from dynamic environments and unstable communication channels. Unlike existing approaches that rely on perfect channel state information (CSI) or assume static environments, this paper proposes a deep deterministic policy gradient (DDPG)-based offline-online joint optimization (OOJO) framework for trajectory optimization and resource allocation in a reconfigurable intelligent surface-assisted unmanned aerial vehicle (UAV) integrated sensing and edge computation system. Imperfect CSI and highly dynamic vehicle mobility are explicitly modeled, formulating a mixed-integer non-convex problem to minimize the total sum of the maximum end-to-end latency across all UAVs. The proposed centralized single-agent OOJO framework employs an actor-critic network to output continuous control actions, enabling closed-loop optimization from sensing to scheduling. During offline learning, general policies are learned from historical data; during online learning, strategies are rapidly adapted to dynamic environmental changes. Simulation results show that the proposed DDPG algorithm outperforms the Deep Q-Network by 21.2% while maintaining low latency across varying vehicle speeds and imperfect CSI, validating its practicality and efficiency in complex dynamic traffic scenarios.
Modern waveform design faces a fundamental trade-off between the computational simplicity of channel estimation (CE) and the maximization of diversity gains. While orthogonal frequency division multiplexing (OFDM) enables low-complexity one-tap equalization, it inherently suppresses multipath diversity by collapsing resolvable delay components into a single frequency-domain coefficient. Conversely, although affine-domain processing captures full channel diversity, its performance is often bottlenecked by estimation errors. This paper proposes a hybrid domain processing framework that aggregates path diversity to enhance the effective signal-to-noise ratio (SNR) without altering the existing frame structures. Firstly, we highlight the limitations of conventional iterative-search CE in the affine domain, demonstrating that fractional delay spreading induces severe estimation error floors. To resolve this, we propose a cross-domain framework that decouples estimation from data multiplexing. Our design utilizes frequency-domain demodulation reference signals (DMRS) for robust, low-complexity CE, followed by a derived full-rank mapping to an affine-domain data block for coherent path aggregation. Furthermore, we analyze the impact of Doppler-induced channel aging and subcarrier spacing constraints to mitigate inter-carrier interference (ICI). Simulation results under standardized 3GPP channel models demonstrate that the proposed scheme eliminates the error floors of native affine CE and achieves bit-error-rate (BER) performance approaching multi-antenna MIMO baselines, while maintaining the hardware simplicity of a single-antenna transceiver.
Turbo product codes have emerged as promising candidate channel codes for hyper-reliable low-latency communication scenarios owing to the inherent parallelism of product code decoding. Recently proposed product-coded MIMO systems with polar component codes demonstrate both performance gains and latency reductions compared to traditional 1-D coded systems. However, the existing successive cancellation list (SCL)-based decoding algorithm in polar-product-coded MIMO systems suffers from high memory costs due to storing internal LLRs and still exhibits potential for further performance improvements. To enhance the performance of product-coded MIMO systems and alleviate storage overhead, this paper investigates decoding schemes based on list ordered statistics decoding (List-OSD). A hardware-friendly soft-output computation method for List-OSD is proposed, which exploits both candidate codewords and pruned paths generated during decoding. Two soft-output list ordered statistics (SOLOS) decoding schemes are presented, respectively incorporating the Pyndiah method and the proposed soft-output computation technique. Simulation results show that the two proposed decoding schemes outperform conventional decoders across various code rates and constructions in product-coded MIMO systems. For the (64, 42)2 polar-product code, the proposed decoders with a list size of 32 achieve a 0.4 dB performance gain over the SCL-based decoder with a list size of 8 at comparable latency, while reducing the byte-level memory usage by 51.2%.
The minimum mean square error (MMSE) detector is widely adopted for massive MIMO systems, but its reliance on matrix inversion results in prohibitive computational complexity. To solve this problem, linear iterative detectors, e.g., weighted Neumann series approximation (wNSA), can circumvent matrix inversion and substantially reduce complexity. However, their performance degradation in high-dimensional high-order MIMO scenarios remains unacceptable. In this paper, we propose WeFTNet, a deep neural network (DNN)-aided weighted Fourier transform (FT)-based approximate MMSE detector for massive MIMO systems. Specifically, the proposed WeFTNet integrates two unexplored features: weighted FT-based MMSE approximation and DNN-driven weighting factor optimization. Furthermore, a geometrically decreasing weighting factor model is theoretically derived and introduced to enhance the training stability and performance robustness. Numerical results demonstrate that the WeFTNet consistently achieves near-MMSE bit error rate (BER) performance in various MIMO scenarios. Compared with the other linear iterative detectors, the proposed WeFTNet succeeds in delivering up to 3.5 dB performance gains while achieving 57% complexity reduction in 256 × 128 256-QAM MIMO systems.
Automatic modulation recognition (AMR) is a cornerstone of dynamic spectrum allocation and cognitive radio. Despite the promise of deep learning-based methods, their reliance on large labeled datasets limits their practicality in scenarios with scarce labeled samples. To address this, we propose a confidence-aware semi-supervised learning (CASL) framework tailored for AMR. CASL adopts a two-stage approach: first, a time-warping-based data augmentation strategy is used to extract general features through contrastive learning. In the second stage, a self-adaptive threshold mechanism generates high-quality pseudo-labels while mitigating confirmation bias using class-level information. To enhance computational efficiency and performance, we design a pyramid learnable filter-based encoder that captures multi-scale features from large-scale, long-sequence signals, reducing inference time by about half compared with the self-attention-based encoder variant. Simulation results demonstrate substantial performance gains on four public datasets, achieving a 20.43% average accuracy improvement over state-of-the-art methods in label-scarce scenarios and up to 38.64% in low signal-to-noise ratio (e.g., 0 dB) conditions. These results highlight CASL’s potential for practical AMR applications in challenging environments.
Indoor localization has emerged as a critical enabling technology for various smart applications, yet its performance is constrained by multipath propagation, signal blockage, and environmental dynamics. While recent deep learning (DL)-based approaches have demonstrated promising improvements over traditional techniques, their effectiveness is often limited by high architectural complexity, limited robustness against electromagnetic-environment variations, and weak interpretability. To address these limitations, this paper proposes a novel indoor localization framework that leverages the structural characteristics of dual-polarized channel state information (CSI) through disentangled representation learning. By decoupling the shared and exclusive latent features embedded in the two polarization channels, the proposed framework is able to isolate location-relevant properties from device-specific and electromagnetic distortions. Furthermore, a three-stage joint optimization strategy is developed to ensure effective feature disentanglement and robust localization performance. Compared with various representative localization models, the proposed framework achieves over 19% enhancement in localization accuracy, demonstrating improved adaptability under imperfect CSI, clock-synchronization impairments, different obstacle densities, dynamic scenarios, different antenna configurations, and reduced training data conditions, with better model interpretability and low inference complexity. Ablation studies and feature visualization further confirm the viability of the proposed network structure design, highlighting the effectiveness of information-theoretic disentanglement in extracting location-relevant features.
Cross-silo federated reinforcement learning (FRL) trains a shared policy across silos that must simultaneously respect a per-round uplink budget and on-device privacy. Although differential privacy (DP) and gradient compression each have mature solutions in supervised federated learning, naively stacking them in the policy-optimisation regime fails, for two structural reasons. First, whole-client DP injects noise whose ratio to the signal scales as √d/N, which for neural policies exceeds one at every practical ε, so the policy collapses to random. Second, a single shared compression mask cannot represent each client’s descent direction under non-i.i.d. environments, driving the compressor’s contraction constant toward zero and destroying learning. In this paper, we propose FedDPRL, a federated policyoptimisation algorithm built from two coupled designs. Specifically, a record-level (per-trajectory) Gaussian mechanism clips and noises each trajectory, lifting the effective sample size to R = NB and removing the √d/N barrier so that DP becomes viable on neural policies. Additionally, an error-feedback Top-k compressor with per-client masks preserves each client’s own descent direction while cutting the uplink by an order of magnitude. Experimental results across classic control, continuous control (MuJoCo), and vision with N = 20 clients show that FedDPRL matches dense DP-FedAvg-RL utility at 13–14× less uplink, while a shared mask collapses to near-random.
The design of millimeter-wave (mmWave) communication systems requires comprehensive research on mmWave channels. Although notable exceptions such as New York University and 3GPP have developed well-validated models based on extensive measurements, many other existing mmWave channel models still rely predominantly on theoretical analysis and lack sufficient support from substantial real-world measurements, particularly across multiple scenarios. Based on the measured mmWave channel data, this paper proposes an improved delay-domain clustering algorithm and a novel and more accurate cluster-based Saleh-Valenzuela (SV) channel model for both line-of-sight (LOS) and obstructed line-of-sight (OLOS) cases. This model separately extracts the first cluster amplified by the directional antenna for analysis, which is more suitable for the mmWave characteristics and has higher accuracy. The eight key cluster parameters and their optimal fitting distributions are extracted, and cluster characteristics in multiple scenarios are analyzed, which reveal the mmWave propagation mechanism. The generated fitting parameters are synthesized into simulated channels, and the root-mean-square delay spread, Kolmogorov-Smirnov (KS) distance and Quantile-Quantile analysis are used to validate this model in multiple scenarios. Finally, the overall process of mmWave channel modeling is presented. This paper provides valuable guidance and a novel channel model applicable to multiple scenarios for the simulation, design, and development of mmWave communication systems.
