Learned Blockwise Port Activation for Real Time Beamforming in Fluid Antenna Arrays

2026-07-28

AI Summary

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.

Abstract

Fluid antenna arrays (FAAs), support multiuser downlink transmission by activating a subset of reconfigurable ports. The activation mask jointly determines the effective channel and the sparse radiating aperture, which requires a balance among sum rate, sidelobe suppression, hardware constraints, and online complexity. Channel driven selection can cluster active ports and increase sidelobes, whereas sidelobe oriented synthesis is typically channel independent and can sacrifice sum rate. This paper proposes learned blockwise port activation (L-BPA), for real time sidelobe aware FAA downlink beamforming. L-BPA activates a fixed number of ports in each aperture block, which supports grouped switching hardware and limits port clustering. A lightweight convolutional network scores ports using multiuser channel features, port coordinates, and user power statistics. Training combines blockwise straight through masks with a differentiable peak sidelobe level (PSLL), surrogate. During inference, learned scores are combined with multiscale geometric repulsion, followed by regularized zero forcing precoding over the reduced effective channel. L-BPA reduces the average PSLL by 3.26 dB relative to uniform sparse activation while achieving a slightly higher sum rate. It also reduces the PSLL by 8.13 dB and 10.10 dB relative to greedy and gain based selection, respectively, without iterative online search.

Problem

• fluid antenna arrays
• port activation
• beamforming
• sidelobe suppression
• sum rate

Innovation

• fluid antenna arrays
• blockwise port activation
• peak sidelobe level
• learned beamforming
• real-time precoding