Paper Title
ANALYSIS OF HYBRID PRECODER USING DEEP LEARNING FOR 5G MASSIVE MIMO MMWAVE COMMUNICATIONS

Abstract
Beam forming (BF) has been one of the most important enabling techniques for millimeter wave (mmWave) communications and massive multiple-in-multiple-out (MIMO) systems.Hybrid beamforming, which divides BF operation into radio frequency (RF) and baseband (BB) domains, will play a critical role in MIMO communication at millimeter-wave (mmWave) frequencies. For the purpose of sending multiple data streams through the channel, a set of precoding and combining weights are derived from the channel matrix. In this research work, the precoding weight will be designed by using deep learning-based method (Long Short-Term Memory network). It is assumed that the base station (BS) has the channel statistics only and feeds the channel covariance matrix (CCM) into a LSTM to obtain hybrid precoders (FBB and FRF). To train the proposed LSTM network, 50 different scenarios for channel matrix are considered in this research. Different number of antennas (NT= 16, 32, 64) are also considered in this paper.After training the data, the hybrid precoders values are predicted andtheRMSE values are calculated. Through simulation, this paper demonstrates that the proposed approach has achieved near-optimum solutions. Keywords - Hybrid Precoder, LSTM Network, CCM, Massive MIMO, mmWave