correlation-aware port selection

This study addresses beam distortion and port selection failures caused by mutual coupling in densely packed fluid antenna arrays. We propose a mutual coupling-aware design paradigm that models the array as a coupled multi-port network, explicitly incorporating inactive ports. By jointly optimizing active port selection and source voltages, this approach actively exploits mutual coupling effects rather than merely compensating for them passively. Compared to traditional uncoupled models and fixed arrays, the proposed method significantly enhances average main-lobe signal-to-noise ratio and reduces peak sidelobe levels within a limited aperture. These improvements effectively overcome performance bottlenecks inherent in dense configurations, enabling precise beamforming through the constructive utilization of electromagnetic interactions in compact fluid antenna systems.

This work addresses the challenge of port activation in fluid antenna arrays under limited RF chain constraints, where existing approaches struggle to jointly optimize communication rate, aperture sparsity, and feed feasibility. To overcome this limitation, the paper proposes an Impedance-Aware Zonal Port Activation (IA-ZPA) method that uniquely integrates mutual coupling effects with real-time channel state information (CSI) into the port selection process. IA-ZPA leverages a learnable CSI-driven scoring network, a checkerboard-inspired feasibility projection, and a mutual-impedance-aware selection mechanism to achieve efficient sparse configurations. Furthermore, it incorporates current-domain regularized zero-forcing (RZF) beamforming and models mutual impedance under an induced electromotive force protocol. Experimental results demonstrate that, under average sidelobe level constraints, IA-ZPA significantly outperforms greedy strategies by achieving the highest constrained transmission rate while substantially reducing decision latency.

This work addresses the challenge of jointly optimizing sum-rate performance, sidelobe suppression, hardware constraints, and real-time computational complexity in port activation for fluid antenna arrays. To this end, the authors propose a learning-based block-wise port activation (L-BPA) method, which fixes the number of active ports within each aperture block and integrates a lightweight convolutional scoring network, a differentiable proxy for peak sidelobe level, block-wise straight-through masking, and a multi-scale geometric repulsion mechanism. This design effectively prevents port clustering while enabling low-complexity real-time beamforming without online iterative search. Experimental results demonstrate that, compared to uniform sparse, greedy, and gain-based selection schemes, L-BPA achieves comparable or slightly improved sum rates while reducing average peak sidelobe levels by 3.26 dB, 8.13 dB, and 10.10 dB, respectively.

This work addresses the significant overhead in channel state information acquisition and feedback caused by the high-dimensional radiation patterns of massive MIMO and reconfigurable intelligent surfaces (RIS), which hinders efficient beam management. To overcome this, the authors propose a training-free, compressed representation model for radiation patterns. They introduce a novel 3D pattern modeling approach that combines low-order spherical harmonics with anisotropic Gaussian kernels, and for one-dimensional azimuth slices of RIS responses, they design a hybrid sparse representation using Fourier bases and 1D Gaussians. This method yields a hardware-aware, interpretable, and high-fidelity low-dimensional parameterization. Experiments on the AERPAW platform and a public RIS dataset demonstrate reconstruction mean squared errors reduced to 1/2.8 and 1/10.4 of baseline methods, respectively. Simulations further show a 12.65% average uplink throughput gain under a fixed uplink budget.

This paper addresses the active user detection (AUD) problem in massive machine-type communication (mMTC), investigating the coupling and trade-off between signal sparsity and frequency diversity. We propose a joint framework of multi-frequency non-orthogonal pilot transmission and single-channel orthogonal matching pursuit (OMP), which for the first time quantifies the dynamic boundary between sparsity gain and frequency diversity gain. Based on this, we establish an optimal frequency diversity allocation criterion under resource constraints, breaking the conventional AUD modeling paradigm that neglects sparsity degradation effects. Experiments demonstrate that in short-pilot/multi-antenna regimes, sparsity dominates performance—yielding a 3.2× improvement in detection success rate; whereas in long-pilot/few-antenna regimes, introducing moderate frequency diversity reduces the AUD error rate by 90%.

This work addresses the challenge of dynamic wireless blockage caused by container stacking and industrial infrastructure in smart ports, which severely hinders the efficient deployment of Mobile Integrated Access and Backhaul (MIAB) base stations. To tackle this, the authors propose DOCKING—a framework that reconstructs RSRP/SINR radio environment maps using ordinary Kriging interpolation from sparse wireless measurements and known network parameters, without requiring prior geometric information about obstacles. Strong attenuation regions are abstracted into compact cuboid models to drive backhaul-aware joint MIAB optimization. This approach uniquely integrates radio environment map reconstruction with obstacle inference for industrial MIAB deployment. Experimental results demonstrate that with only 15% spatial sampling, the 90th-percentile REM prediction error remains below 3 dB, obstacle detection achieves a true positive rate exceeding 85%, system capacity improves by up to 150% under sparse deployment, and each optimization converges within 5–15 seconds, with measured throughput closely matching predictions.

This work addresses the challenges of sidelobe and grating lobe leakage in sparse arrays for extremely large-scale MIMO systems, which, while cost-effective and aperture-preserving, suffer from interference, whereas rotatable antennas that mitigate such leakage incur high control overhead. To overcome this trade-off, the paper proposes a user-group-oriented sparse rotatable antenna architecture that partitions users into service groups, each served by a dedicated sparse subarray. It jointly optimizes aperture allocation, antenna orientation, and beamforming to maximize the minimum SINR. Leveraging group-level geometric information, the design enables low-complexity configuration and reveals an approximately decoupled beam structure—inter-group leakage suppression and intra-group orthogonalization—that guides antenna placement and allocation. A two-layer structured algorithm, integrating analysis-guided initialization, multi-start search, and closed-form orientation rules, achieves near-ideal performance with substantially reduced hardware cost, significantly outperforming both compact and omnidirectional sparse baselines.

This work addresses the limitations of conventional channel coding in non-adaptive single-RF-chain massive beamspace MIMO systems, where the inability to support noncoherent decoding degrades sensing performance. The authors formulate channel sensing as a noncoherent decoding problem and, for the first time, derive an exact expression for the subspace distance of binary linear codes under BPSK mapping, revealing a fundamental distinction between Hamming and subspace distances. Building on this insight, they propose a beamspace subspace code based on Golomb ruler-inspired sparse antenna selection, integrated with maximum-likelihood angle estimation and convolutional beamforming. This approach achieves hardware- and sampling-efficient operation while preserving theoretical performance guarantees. Results demonstrate that high-Hamming-distance codes without careful design may yield zero subspace distance and thus fail entirely, whereas the proposed method attains near-optimal subspace distance and robust sensing performance.

This study addresses the large-scale non-convex mixed-integer challenge of jointly optimizing port selection and precoding in multi-user MIMO systems. We propose a solution framework based on discrete diffusion models that reformulates combinatorial optimization as sampling from a target distribution. The core innovation lies in introducing a novel potential function learning mechanism driven by local objective differences, integrated with the Metropolis-Hastings criterion to guide parallel sampling without requiring optimal label supervision. Experimental results demonstrate that the proposed method achieves a sum rate within merely 0.3% of exhaustive search while accelerating computation by 150 times. Furthermore, it operates 18.8 times faster than greedy algorithms, exhibiting significantly superior overall performance compared to existing state-of-the-art approaches.

This work addresses the performance degradation in fluid antenna systems (FAS) caused by strong spatial correlation among densely packed ports, which diminishes port distinguishability and impairs spatial modulation efficacy. Focusing on a single RF-chain FAS transmitter paired with a multi-antenna SIMO receiver, the study proposes three correlation-aware port selection strategies: SF-EDAS enhances constellation separability, SOPS minimizes the condition number of the channel matrix, and CC-COAS jointly optimizes channel gain and decorrelation. A reliability analysis framework grounded in energy and extreme-value degrees of freedom is developed, complemented by maximum-likelihood detection and high-SNR diversity analysis. Experimental results demonstrate that the proposed schemes significantly outperform conventional spatial modulation and grouped benchmarks, with CC-COAS achieving the best trade-off between bit error rate performance and computational complexity.