Near-optimal signal detection with low complexity based on Gauss-Seidel method for uplink large-scale MIMO systems

Minimum mean square error (MMSE) algorithm is near-optimal for uplink large-scale MIMO systems, but involves high-complexity matrix inversion. In this paper, based on a special property of uplink large-scale MIMO systems that the filtering matrix of the MMSE algorithm is symmetric positive definite as we will prove, we propose to exploit the Gauss-Seidel method to iteratively realize the MMSE algorithm without the complicated matrix inversion. The proposed signal detection algorithm can reduce the complexity by one order of magnitude. Simulation results verify that the proposed algorithm outperforms the recently proposed Neumann series approximation algorithm, and achieves the near-optimal performance of the classical MMSE algorithm with a small number of iterations.