Abstract
Effective Forest fire management requires enhancing the early identification of fire hotspots and continuous monitoring of key drivers to mitigate risks and accelerate response strategies. This study focuses on the application of flying robots for wildfire management and introduces an efficient cooperative search and coverage framework designed for uncertain environments, where the origins and expansion patterns of forest fires are initially unknown. introduces an efficient cooperative search and coverage framework designed for uncertain environments where the origins of forest fire and expansion patterns are initially unknown. Operational constraints such as intensity of fire and smoke, limited mission duration, and number of available aerial firefighting vehicles can significantly impact search and monitoring efficiency. To surmount these hurdles, a data-driven decision-centric approach is proposed, integrating the history of aerial and satellite images with probabilistic feature matching. The methodology defines key surveillance regions and formulates past fire management missions as optimization problems to drive an optimal dataset. Utilizing a normalized cross-correlation algorithm, the historical data that best matches the probability distribution of the current fire scenario is identified. The extracted optimal parameters are then applied to the ongoing mission to improve search and coverage efficiency. The effectiveness of this approach is validated through simulations, demonstrating its potential to enhance flying-robot-based forest fire detection and tracking in uncertain and dynamic environments.