Asymptotic Equivalent Performance of Uplink Massive MIMO Systems With Spatial-Temporal Correlation

This paper investigates the achievable sum rates of uplink massive multiple-input multiple-output systems with matched-filter (MF) and zero-forcing (ZF) receivers, under the spatial-temporal correlation channel scenario. First, we give the lower and upper bounds of the system capacity. Then, the asymptotic equivalent (AE) expressions of the bounds are obtained based on the random matrix theory. Moreover, the power scaling laws of MF and ZF receivers are exploited. Specifically, it is shown that the transmit power of each user can be scaled down by both MF and ZF receivers. Besides, the existence of the optimal number of users for the ZF receiver is proved based on the AE lower bound, and an efficient algorithm is proposed to obtain the near-optimal number of users when the number of receive antennas is fixed. Meanwhile, some simplified expressions of key parameters are obtained in the low-signal-to-noise-ratio (SNR) regime (i.e., the minimum normalized transmit energy per information bit and wideband slope) and the high-SNR regime (i.e., high-SNR slope and power offset). The simulation results and complexity analysis validate the proposed bounds and AE expressions.