Performance Analysis of ZF-Precoded Scheduling System for MU-MIMO with Generalized Selection Criterion

In this paper, we present performance analysis of a zero forcing (ZF)-precoded scheduling system for multiuser multiple input multiple output (MU-MIMO) broadcast transmission. The ZF-precoded scheduling system selects a user set with selection criterion of the effective signal to noise ratio (SNR). We consider generalized selection criterion of user set which is the ordinal number of order statistics for the effective SNR and show insights into how the selection criterion affects performance of the scheduling system. The tight closed-form expressions of the ZF-precoded scheduling system for symbol error rate (SER) and ergodic sum-rate capacity are derived for the generalized selection criterion. We also present asymptotic analysis to quantify diversity order and SNR gain, and show that the ordinal number of order statistics for the effective SNR is the factor of the diversity order. From the analytical results, we show that the scheduling system with selection criterion of the maximum effective SNR provides the best performance of SER and ergodic sum-rate capacity by multi-user diversity (MUD) over the conventional ZF-precoded system, while degrading performance for the selection criterion of the minimum effective SNR. This result means that the performance of the ZF-precoded scheduling system mainly depends on selection criterion.