Enhancing Autonomus Aerial Monitoring of Large-Scale Photovoltaic Plants Using a Deep Reinforcement Learning Approach for Multi-Aerial Robot Systems

Eficient inspection of large-scale photovoltaic (PV) power plants can be achieved using a multi-aerial robot system powered by Deep Reinforcement Learning (DRL). This study introduces an innovative algorithm for online coverage path planning to optimize the efficiency, scalability, and adaptability of multi-aerial robot systems. By dynamically adapting to real-time environmental changes, the algorithm ensures comprehensive inspection coverage while reducing computational overhead. Simulations conducted in both grid-based and realistic environments demonstrate the algorithm's robustness and its potential to revolutionize inspection tasks in dynamic and complex scenarios.