Cooperative forest fire monitoring using multiple UAVs

This paper addresses the problem of using a team of UAVs to cooperatively monitor multiple forest fires (called hotspots) in a region. We deploy two types of agents: service agents that locally monitor the state of the hotspots and detector agents that search for hotspots and assign service agents to those that are detected. We develop an assignment scheme based on auctions to perform the assignment of a hotspot to a group of service agents. While monitoring a hotspot, service agents are required to be equally spaced around the perimeter. In order to achieve equal spacing around the hotspot perimeter, we develop a splay state controller that ensures the convergence of the service agents to the equally spaced configuration. We also address the problems associated with tracking hotspots that expand in size and possibly combine together. Simulation results are presented to validate the assignment and control algorithms.