Fast Beamforming Strategy: Learning the AoD of the Dominant Path

Abstract

Small cell networks and directional beamforming have been regarded as the solutions to achieve high data rates and compensate for the high path loss in the millimeter-wave (mmWave) communications. Large antenna array in mmWave communications causes considerable overhead to select the beam using exhaustive beam search, which significantly affects the efficiency of these mobile communications. In this paper, we propose a fast beamforming strategy by estimating the angle of departure (AoD) of the dominant path by exploiting the position of the mobile station, leveraging the deep neural network. Simulation results, based on accurate ray-tracing, show that we can achieve up to 0.58° of angle accuracy.