Joint Beam Management and Relay Selection Using Deep Reinforcement Learning for MmWave UAV Relay Networks

Abstract

Unmanned aerial vehicle (UAV) relays are useful in tactical millimeter wave (mmWave) networks to overcome blockages and improve link resilience. Getting the most benefits from relays, though, requires efficient strategies for relay selection and for beam management. In this paper, we use deep reinforcement learning (DRL) to jointly select unblocked UAV relays and to perform beam management. The proposed DRL-based algorithm uses rate feedback from the receiver to learn a policy that adapts to the dynamic channel conditions. We show with numerical evaluation that the proposed method outperforms baselines without prior channel knowledge. Moreover, the DRL-based algorithm can maintain high spectral efficiency even under frequent blockages.