DRL Enhanced Reconfigurable Intelligent Surface for Efficient Air-Ground Vehicle Communications

This article studies the reconfigurable intelligent surface (RIS) controlled by deep reinforcement learning (DRL) that enable reliable links between aerial platforms and ground vehicles. Particularly, we propose a model for unmanned aerial vehicle (UAV)-assisted Internet of vehicles (IoV) networks using conformal car RIS as relays. The DRL agent learns to adaptively control the intelligent reflecting surfaces to optimize wireless channels specifically between aerial nodes like UAVs and ground nodes like connected vehicles. By leveraging dynamic beamforming and 3D propagation environments, the DRL-RIS approach can focus signals to intended vehicle receivers while minimizing interception. This strategy will provide efficient connectivity for air-ground vehicle communication amidst complex real-world conditions. As unmanned aerial systems and intelligent vehicles continue to proliferate, the use of AI-enabled reconfigurable metamaterials for robust security will become increasingly important.