Applications of Deep Learning for Millimeter Wave

Abstract

Millimeter Wave (mm-wave) has been considered as significant importance in various communication systems. It has achieved a greater attention to meet the capacity requirement of the future 5G network. Since mm-wave has a high frequency (30 to 300 GHz) using orthodox technologies for mm wave is more challenging. Thus advanced technology i.e. Deep Learning (DL) is a pragmatic approach to analyze a massive amount of data. Firstly, to find out how DL has beaten traditional approaches, this review briefly explores, the different methods of DL for mm wave are. Secondly, the review of the multiple applications in mm wave such as beam and blockages prediction, beam spacing, beamforming for mm wave OFDM system, precoding for mm-wave, channel estimation for mm-wave, sparse channel estimation, and hybrid precoding and fingerprinting-based indoor localization with mm wave is concisely explained. Last but not least, several studies have proved that DL has superior efficiency for mm wave than conventional approaches.