Low complexity user selection for MU-MIMO downlink based on MSMSE transceiver

This paper investigates low complexity user selection in multiuser multiple-input multiple-output (MU-MIMO) downlink assuming perfect channel state information. Various capacity-based user selection algorithms have been proposed, most of which usually based on null multiuser interference. There is always a performance gap between channel capacity theory and real transmission rate. We study the user selection based on minimum-sum-mean-squared-error (MSMSE) criterion which can reflects the system's performance more properly. We then propose a low-complexity user selection algorithm that doesn't eliminate the multiuser interference completely and considers power allocation during the selection procedure. Simulations show that the proposed algorithm can achieve a much lower BER, and attain the sum rate close to the brute-force search.