Hybrid Beamforming for Multiuser MIMO mm Wave Systems Using Artificial Neural Networks

Abstract

Hybrid Beamforming has been used in wireless communications for many years. With the fifth generation of wireless communications or (5G) and beyond networks, the need for beamforming is ever increasing because of the use of higher frequencies and the need to provide better coverage and better spectral utilization. Although many designs have been suggested to build hybrid beamforming, the Machine Learning (ML) based designs have attracted much attention recently because of the flexibility in coping with the wireless channel variations and user mobility they can attain when directing the transmission to the right direction during the communication process. In this paper, we describe the extended design of machine learning based hybrid beamforming for multiple users in systems that use millimeter waves (mmWaves) and massive MIMO architectures. The simulation results show that with the right amount of training data samples (channel feedback), the ML based hybrid beamforming architecture can achieve the same spectral efficiency (bits/sec/Hz) as the fully digital beamforming designs with negligible error for both single user and multi-user Massive-MIMO scenarios.