Low-complexity Optimized Version of AOR Algorithm for Signal Precoding in Large-scale MIMO Systems

In recent years, there has been a growing focus on research concerning wireless communication technologies, with a particular emphasis placed on the emerging field of massive MIMO systems. In these systems, precoding performed at the base station (BS) is a crucial signal processing task which ensures reliable downlink transmission. In this paper, we propose a new modified accelerated overrelaxation (AOR) approach to enhance signal precoding in large-scale MIMO downlink systems. This approach uses distinctive matrix decompositions along with optimally selected relaxation and acceleration parameters. Specifically, the proposed method, termed "optimized symmetric accelerated over-relaxation (OSAOR)", exhibits two key advantages: low complexity (compared to the near optimal zero forcing (ZF) precoder) and iterative nature, with its parameters optimized by means of the particle swarm optimization (PSO) algorithm that is capable of boosting convergence and improving precoding precision. Simulation results are given to confirm the superiority of the proposed algorithm, as it may outperform conventional AOR and other existing solutions.