Precoding with the Assistance of Attitude Information in Millimeter Wave MIMO System

Abstract

Digital beamforming (DBF) is considered as an efficient method to overcome the high propagation loss of millimeter wave (mmWave) communication, but the acquisition of channel state information (CSI) brings huge training overhead, especially in high mobility scenarios. To tackle this challenge, we consider using attitude information from motion sensors to reduce the training overhead of DBF in this paper. We first analyze the characteristics of mmWave uplink channel when the attitude of user equipment (UE) rotates, and it shows that only the precoder needs to be redesigned after the rotation. Therefore, we develop a novel attitude information aided precoding algorithm, which approaches the performance of conventional singular value decomposition (SVD) algorithm. The proposed algorithm reduces the channel estimation and feedback overhead significantly compared to the conventional one. Finally, the simulation results show that the proposed algorithms allow mmWave systems to approach their performance limits.