Cooperative Multi-Target Search with UAV Swarms: Evolutionary vs. Reinforcement Learning Strategies

Multi-UAV systems have many uses in civil applications including search and rescue, surveillance, and monitoring. In dynamic missions such as search and rescue, where multiple targets or areas of interest might need to be searched and monitored, the accuracy of target detection and information sharing among UAVs and with ground control station (GCS) play a vital role in mission success. In this work, we present a joint coverage, connectivity, and revisit time optimization framework for multi-target search missions with imperfect sensing and dynamic information exchange. We provide evolutionary algorithms (EA) and reinforcement learning (RL) based solutions to the optimization problem. Our results show that by considering cell revisit times as well as connectivity in the path planning using EA, the target detection time can be improved by 30% and we can achieve near-instant GCS inform in realistic, uncertain environments. RL-based results show worse mission time performance compared to EA methods indicating that further tuning and better representation methods might be required.