Examining Spatial Consistency for Millimeter-Wave Massive MIMO Channel Estimation in 5G-NR

Third-Generation Partnership Project (3GPP) has determined the millimeter wave (mmWave) as one of the main technologies to achieve a high data rate communication scheme for 5G. Unlike conventional sub-6 GHz bands, mmWave communication has short areal coverage due to its propagation characteristics. As a result, mmWave strongly depends on beamforming using Massive MIMO to achieve expected optimal spectral efficiency. For this propose, accurate channel state information is essential to align the antenna arrays properly. However, mmWave channel estimation is challenging because of its sparse delay-angular characteristics. In this paper, we are examining the local Spatio-temporal correlation of the non-stationary mmWave channel to study the effect of the spatial consistency from mmWave channel estimation point. We are employing Correlation Distance Matrix (CDM) to characterize the correlation between channel ensembles in non-stationary mmWave channel model. Our observation shows that spatial consistency decreases sharply as the physical distance between the channel ensembles increases. Also, we observed that channel estimation using a sparse recovery algorithm preserves the spatial consistency between estimated channels. Our results show that the spatial consistency can be employed to design more efficient channel estimation techniques for mmWave massive MIMO system provided that the update distance is very short.