A Robust UAV-UGV Collaborative Framework for Persistent Surveillance in Disaster Management Applications
Md Safwan Mondal, R. Subramanian, James Humann, James M. Dotterweich, Jean-Paul Reddinger, Marshal Childers, Pranav A. Bhounsule
Abstract
Unmanned Aerial Vehicles (UAVs) are fast, agile, and capable of covering large areas quickly but are constrained by their limited fuel capacities. In contrast, Unmanned Ground Vehicles (UGVs) have longer battery lives but move at slower speeds. By combining UAVs with UGVs, which serve as mobile recharging stations, we can harness the strengths of both: UAVs can achieve rapid task execution over extended periods by refueling from UGVs. This synergy makes the collaborative routing of UAVs and UGVs well-suited for modern disaster management applications. However, their varied operational constraints require a sophisticated planning framework to ensure optimized coordination and task execution. In this paper, we introduce a robust multi-agent framework leveraging asynchronous planning to optimize the routes of UAVs and UGVs in a persistent surveillance task, considering their individual limitations like fuel, speed, and charging constraints. The framework is designed to scale effectively with the number of vehicles and accommodates diverse team configurations. The effectiveness of this framework is demonstrated through a simulation of a 4-hour mission covering 30 task points across five different team compositions, showing significant improvements in route efficiency. Additionally, a detailed cost analysis identifies the optimal UAV-UGV team composition by effectively balancing mission performance and cost, thus serving as a valuable tool for optimizing disaster response strategies.