Radar-Assisted Predictive Beamforming for UAV-Aided Networks: A Deep-Learning Solution
Rui Yin, Jingwei Peng, Yunlong Cai, Celimuge Wu, Benoı̂t Champagne, Naofal Al‐Dhahir
Abstract
In the realm of Sixth-Generation (6G) mobile communication systems, unmanned aerial vehicle (UAV) technologies have gained significant attention, particularly within UAV-assisted wireless networks. This paper explores radar-assisted predictive beamforming to enhance communication in such networks. A mobile UAV equipped with multiple antennas and radar functionality is deployed to serve a group of mobile users on the ground. The primary objective is to optimize communication performance, ensuring stable and efficient data transmission rates within a specified time frame, under practical power and flight range constraints. To address this challenging problem, we introduce a novel solution framework that integrates a deep reinforcement learning (DRL) network and a dual-layer deep-unfolding network (DUN). By leveraging a two-timescale processing structure, the new framework enables and facilitates the joint optimization of mobile user tracking, transmit beamforming and UAV trajectory. Simulation results confirm the effectiveness of our proposed scheme, demonstrating improvements in movement tracking accuracy and transmission rates.