A Low Complexity Linear Precoding Method for Massive MIMO

Abstract

In massive multiple-input multiple-output (MIMO), the base-station (BS) is equipped with tens or hundreds of antennas to serve many users. The demand for a low complexity precoding algorithm is significantly increased. Linear precoding schemes, such as zero-forcing (ZF), are capable to obtain a quasi-optimum performance when a favourable channel propagation occurs in downlink massive (MIMO) systems. Particularly, conventional linear precoding techniques require to calculate a matrix inverse which increases the expenses of computational complexity. In this paper, a low complexity linear precoder based on a hybrid acceleration overrelaxation (AOR) method is proposed. The optimal relaxation, acceleration parameters, and the initial solution are selected to achieve an attractive balance between the performance and the computational complexity. Numerical results show that the proposed algorithm has greatly reduced the complexity from $\mathcal{O}(K^{3})$ to $\mathcal{O}(K^{2})$ (where K is the number of users). It also outperforms the up-to-date algorithms.