SAOR-Based Precoding with Enhanced BER Performance for Massive MIMO Systems

Abstract

The iterative precoding scheme is a common algorithm for downlink massive MIMO systems. Due to the large number of system antennas, traditional linear precoding schemes are usually involve the large-scale matrix inversion and leads to high computational complexity. The complexity of the linear precoding algorithm greatly reduced when the iterative algorithm is proposed, but it caused a decline in Bit Error Rate (BER) performance. Improving the BER performance of the iterative algorithm and ensuring the convergence rate has always been the focus of attentions. Currently, there are many iterative methods that only have mathematical theory and not applied to precoding yet. To solve the aforementioned problem, we proposes a precoding scheme based on Symmetric Accelerated Over Relaxation (SAOR) method, which achieve the enhancement BER performance compared to other iterative algorithms. Combined with the actual system, the selection of optimal acceleration factor and relaxation factor are discussed, which is only related to system parameters. The simulation results show that SAOR-based precoding can achieve good BER performance with less iterations and guarantee the convergence rate.