Transformer-Based Hybrid Beamforming with Reconfigurable Pixel Antenna for HAPS Communications

2026-05-18

AI Summary

This work addresses the high computational complexity and inadequate radiation pattern adaptability of conventional hybrid beamforming in massive MIMO systems employing reconfigurable pixel antennas for High-Altitude Platform Station (HAPS) communications. To overcome these limitations, the authors propose PR-HBFNet, a Transformer-based hybrid beamforming framework that, for the first time, integrates the Transformer architecture into the radiation pattern design of reconfigurable pixel antennas. By combining singular value decomposition (SVD) initialization with a model-driven residual learning structure, PR-HBFNet efficiently generates both analog and digital precoders. Experimental results demonstrate that the proposed method achieves spectral efficiency close to that of greedy algorithm benchmarks while substantially reducing computational complexity, thereby striking an effective balance between high performance and low complexity.

Abstract

This paper proposes a Transformer-based hybrid beamforming framework for reconfigurable pixel antenna (RPA)-equipped massive multiple-input multiple-output (MIMO) in high-altitude platform station (HAPS) communications. The proposed pattern reconfigurable hybrid beamforming network (PR-HBFNet) comprises two key components: 1) a pattern reconfigurable network that leverages a Transformer encoder to determine the radiation pattern for each antenna element, and 2) a hybrid beamforming network that employs model-driven residual learning to compute analog and digital precoders over SVD-based initializations. Simulation results demonstrate that the proposed PR-HBFNet closely approaches the spectral efficiency of a greedy benchmark while significantly reducing computational complexity.

Problem

• Hybrid Beamforming
• Reconfigurable Pixel Antenna
• HAPS Communications
• Massive MIMO
• Spectral Efficiency

Innovation

• Transformer-based beamforming
• reconfigurable pixel antenna
• hybrid beamforming
• HAPS communications
• model-driven residual learning