Subarray-level Robust and Fast Variable-Step-Size Parallel LMS Adaptive Beamforming

This paper addresses the issues of high computational complexity, severe grating lobe effects, and insufficient noise suppression in large-scale array adaptive beamforming. We propose a robust and fast variable step-size parallel LMS (RFVss-pLMS) algorithm for subarray beamforming. The RFVss-pLMS algorithm enhances the convergence speed and interference suppression capability of the traditional LMS algorithm through optimized step-size adjustments. It also utilizes parallel computation to accelerate convergence and reduce errors, making it suitable for real-time processing of large-scale arrays. Extensive simulation results show that the proposed algorithm performs excellently in both linear array and planar array configurations, particularly in highinterference environments. It achieves significant improvements in sidelobe suppression, signal-to-interference-plus-noise ratio (SINR), and mean square error (MSE) convergence speed. This method provides a robust solution for adaptive beamforming in large-scale arrays, especially for applications requiring highresolution beam scanning and strong interference suppression.