A low complexity MIMO detector based on combination of MRC and MMSE algorithm

Abstract

In the absence of exhaustive search and matrix inversion, this paper proposes a low complexity multi-input multi-output (MIMO) detector for coded systems based on the combination of maximal ratio combining (MRC) and minimum mean square error (MMSE) algorithm. The proposed scheme takes advantage of MMSE for the first iteration to guarantee reliable soft information derived from channel decoder. With these soft information and interference cancellation (IC), the MIMO system is converted to a single-input multi-output (SIMO) system. Hence, MRC, which is a simple space signal processing technology only for SIMO systems, is used in rest iterations to avoid matrix inverse associated with MMSE-based detector. Because only linear operations are required in rest iterations, the proposed method remarkably reduces the computational complexity of MIMO detector. The simulation results show that the proposed method does not suffer the performance degradation in bit error rate nor convergence compared with MMSE-based detector. Meanwhile, the computational complexity of the former one is at least reduced from O(n2.376) to O(n) except the first iteration compared to the latter. Consequently, the proposed method is a very attractive MIMO detection scheme for practical applications.