Deep Learning-Based Beam Alignment and Tracking Mechanism for mmWave Aerial Base Station

Millimeter wave (mmWave) aerial base stations offer distinct advantages such as enhanced flexibility and mobility compared to ground base station communication. Additionally, the utilization of mmWave technology enables the provision of high-speed, directional services to ground users while mitigating interference with existing ground networks. To establish a good communication link between mmWave aerial base stations and terrestrial mobile users, it is essential for base stations to form directional beam to cover terrestrial users. However, the movement of users and the change of environmental factors will lead to the misalignment between the mmWave beam formed by the base stations and the users, resulting in the instability of the communication link, which will seriously affect the quality of communication service. Therefore, this paper proposes a deep learning-based beam tracking scheme for mmWave aerial base stations. Simulation results show that our proposed schemes achieve higher beamforming gain and require less beam training overhead compared to existing conventional and deep learning-based approaches.