Using Multi-Stage MMSE Detection to Approach Optimum Error Performance in Multiantenna MIMO Systems

This paper proposes a multi-stage minimum mean-square error (MS-MMSE) detector and two low-complexity reliability measurement schemes, namely the Type-L and Type-A schemes, in order for the multiple-input multiple-output (MIMO) systems using binary phase-shift keying (BPSK) modulation to attain near single-user bit-error-rate (BER) performance. In our MS-MMSE detector, signals to be detected are ranked according to their real-time reliabilities measured by the Type-L or Type-A scheme. The Type-L scheme adopts both the signal-to-interference-plus-noise ratio (SINR) and magnitudes of the MSMMSE detector´s decision variables to measure the reliabilities, while the Type-A scheme only makes use of the magnitudes of the MMSE detector´s decision variables for reliability measurement. The BER performance of the MS-MMSE detector is investigated associated with the proposed Type-L or Type-A scheme. Our study and performance results show that the MS-MMSE detector is a low-complexity and high-flexibility detector. Both the Type-L and Type-A schemes are high-efficiency reliability measurement schemes, which make it possible for the MS-MMSE detector to achieve near single-user BER performance.