Maximizing Energy Harvesting with the Aid of Reconfigurable Intelligent Surface for UAV Using Proximal Policy Optimization Algorithm

Unmanned aerial vehicles (UAVs) equipped with Reconfigurable intelligent surfaces (UAV-RIS) are able to offer ubiquitous communication services in areas where communication is disabled, but it is limited by the on-board energy of UAVs. This paper presented the EH-RIS system, a novel energy harvesting (EH) strategy designed for high-performance next-generation wireless systems. By utilizing passive reflective arrays to facilitate parallel energy harvesting and information transfer, the EH-RIS system expands upon the idea of Simultaneous Wireless Information and Power Transmission. However, pedestrian mobility and rapid channel changes imposed through external factors make efficient resource allocation in wireless systems challenging. Thus, a robust model-free, on-policy, an actor-critic method called Proximal Policy Optimization algorithm based on Deep Reinforcement Learning is developed, which improves the decision-making of proposed EH-RH systems for ensuring quality of services under the dynamic wireless environment. The K-means clustering technique and K-medoids have been introduced to optimize the UAV trajectory design. Simulation results show that the provided EH-RIS-based system is both effective and efficient. It performs better than state-of-the-art systems at the moment and is nearly as efficient as exhaustive search strategies. Our proposed approach has great potential for enhancing UAV-RIS systems and enabling more connectivity in places where communication is extremely difficult.