UAV-RIS Assisted Maritime Beamforming Based on Reinforcement Learning

With the growing demand for reliable, high-capacity maritime communications, the installation of Reconfigurable Intelligent Surfaces (RIS) on Unmanned Aerial Vehicles (UAVs) can provide flexible and adaptable services for maritime communications. Jointly optimizing the communication link between the base station and the RIS on the ship is a challenging problem. To address the above problem, this paper proposes a UAV-RIS-assisted maritime communication system based on a two-ray sea-air channel model, which maximizes the summed spectral efficiency by adjusting the transmitting beamforming of the base station and the phase shift of the UAV-RIS. Due to the non-convex nature of the problem, we employ the Deep Deterministic Policy Gradient (DDPG) algorithm to learn an efficient beamforming strategy. Simulation shows that the algorithm's performance is comparable to two advanced benchmarks. It has also been discovered that the algorithm's performance and rate of convergence can be significantly enhanced by using the correct network parameter values.