Low-Complexity MMSE Signal Detection Based on the AOR Iterative Algorithm for Uplink Massive MIMO Systems

Abstract

Massive multiple-input multiple-output (MIMO) systems can substantially improve the spectral efficiency and system capacity by equipping a large number of antennas at the base station and it is envisaged to be one of the critical technologies in the next generation of wireless communication systems. However, the computational complexity of the signal detection in massive MIMO systems presents a significant challenge for practical hardware implementations. This work proposed a novel minimum mean square error (MMSE) signal detection method based on the accelerated overrelaxation (AOR) iterative algorithm. The proposed AOR-based method can reduce the overall complexity of the classical MMSE signal detection by an order of magnitude from \( {\rm O}\left( {K^{3} } \right) \) to \( {\rm O}\left( {K^{2} } \right) \), where \( K \) is the number of users. Numerical results illustrate that the proposed AOR-based algorithm can outperform the performance of the recently proposed Neumann series approximation-based algorithm and approach the conventional MMSE signal detection involving exact matrix inversion with significantly reduced complexity.