Abstract
This paper presents a novel near non-dominated solution approach for distributed cooperative search and coverage (CSC) in uncertain environments for wildfire emergency management. The goal is to efficiently cover the extent of the wildfire and search for unknown stationary fire hotspots using a team of flying agents. Certain management factors such as time and the number of agents can directly impact the effectiveness of cooperative wildfire coverage and fire hotspots search in environments with scattered and widespread uncertainties. To address this, a near non-dominated solutions approach for a new CSC mission in a specific forest region is proposed in this paper. This approach involves generating a dataset based on previously completed wildfire management missions and probability features matching processes. Accordingly, prime and appointed surveillance regions are defined, and previously completed CSC missions are formulated as optimization problems to generate an optimal dataset. Afterward, the normalized cross-correlation algorithm is utilized to identify data that closely matches the probability pattern of the initial probability map of the current mission. By allocating the parameters of the identified data to the operational parameters of the current mission, the performance of the distributed cooperative fire hotspot search and wildfire coverage mission can be enhanced. This allocated data to the current mission serves as the near non-dominated solution. The effectiveness of the proposed strategy is validated and demonstrated via CSC simulations.