Sparse-Aware Minimum Mean Square Error Detector for MIMO Systems

Abstract

In this letter, we propose a novel detector, called the sparse-aware minimum mean square error (SA-MMSE) detector, to improve detection performance for multiple input multiple output (MIMO) systems. We first convert a conventional MIMO system into a sparse system via the symbol error vector obtained from a linear detector. Then, we develop a new sparse refining method, which has a similar structure as the linear MMSE detector, to improve the detection performance. The proposed SA-MMSE detector can be performed iteratively and has low complexity similar to that of the parallel interference cancellation (PIC). Simulation results show that the SA-MMSE detector outperforms the compressive-sensing-based detector and PIC with a similar complexity, especially when the numbers of users and antennas become large.