A multi-UAV fast search path planning algorithm research

This study provides a multi-UAV fast search path planning (MFSPP) algorithm to address the problems of unbalanced task allocation, low efficiency, and time-consuming when several unmanned aerial vehicles (UAVs) conduct search tasks. First, several places with a high probability of target occurrence are inferred based on the location data reported before the target's disappearance. Second, the K-means clustering method is used to initialize the sub-task area. Then, the conflict units of the boundary are divided according to the initial state auction of UAVs, and the search tasks of each UAV are balanced. Finally, each UAV combines the probability of target occurrence with the heuristic greedy search algorithm to design its path in the sub-region. The simulation results reveal that the proposed algorithm is highly scalable and can find the target at a smaller area coverage compared to other search algorithms, which greatly reduces the search time and improves the search efficiency.