Machine Learning-based Direction of Arrival Estimation

Abstract

Beamforming (BF) is expected to be one of the key technologies in Sixth Generation (6G) networks. BF improves the Signal-to-Noise Ratio (SNR) of received signals and focuses the radiation pattern to specific locations by weighting the amplitude and phase of individual antenna signals. This technique provides better coverage in an indoor environment and at the edge of a cell. To make the best use of this technology, it is essential to know the location of the device to direct the antenna beam of the radio Base Station (BS). Consequently, the Direction of Arrival (DOA) method becomes crucial and essential at this time. Therefore, this study addresses the problem of accurately predicting the azimuth and elevation angles of a signal impinging on an antenna array based on Machine Learning (ML) models. Simulation results show that ML models are competitive techniques to find the azimuth and elevation angles of a signal impinging on a receiving system.