Asymptotic Equivalent Performance of Multicell Massive MIMO With Spatial–Temporal Correlation

In this letter, we derive some asymptotic equivalent (AE) sum rates of a multicell downlink massive multiple-input multiple-output (MIMO) system with different beamforming schemes. Due to insufficient antenna spacing and user mobility, we take spatial and temporal correlation into consideration and obtain a correlated channel state information (CSI) model. Under the correlated CSI model, the AE sum rates with matched filter (MF), zero-forcing (ZF) and regularized zero-forcing (RZF) beamforming are obtained in closed-form expression based on random matrix theory. These AE sum rates provide a key insight into theoretically evaluating the effect of correlated CSI on the performance. We find that the spatial correlation has a negative effect on the sum rate and the three beamformings are all sensitive to the temporal correlation. Furthermore, RZF and ZF beamforming are more robust to the spatial correlation compared with MF beamforming.