A Deep Learning-Based mmWave Beam Selection Framework by Using LiDAR Data

Millimeter Wave (mmWave) technology plays an non-substitutable role in 5G communication system because of its effectiveness to increase spectral efficiency. Furthermore, internet of things (IoT) applications will be further developed via mature 5G technology for which 5G has a competitive information interaction delay. Beam selection is one of crucial proposals that can optimize the power dissipation in 5G multi-antenna devices (includes base station and user equipment), particularly for smart mobile devices, such as autonomous vehicles. In this study, a novel 3D convolutional neural networks is proposed for beam selection in a vehicle-to-everything (V2X) application scenario, which is based on 5G mmWave communication system and Light detection and ranging (LiDAR) sensors. The experimental results demonstrate that our proposed methods outperform than the previous methods.