A dual-layer task planning algorithm based on UAVs-human cooperation for search and rescue

To address the issues of low efficiency and difficult localization in search and rescue, a dual-layer task planning algorithm based on UAVs-human cooperation for search and rescue is proposed, which mainly includes the search layer for UAVs and the execution layer for rescuers. Firstly, in the search layer, to solve the problems of uneven task allocation and redundant coverage paths of heterogeneous UAVs, a coverage path optimization based on cluster algorithm (CPOC) is adopted. It applies the K-means algorithm with the proportional constraint to allocate the appropriate task-area for each UAV, and uses the non-redundant exact cellular decomposition method to achieve more efficient planning of the subregion coverage paths, meanwhile, those paths are connected by the Min-Max Ant System. Secondly, in the execution layer, the Rapidly-exploring Random Tree star with dynamic guidance mechanism (DG-RRT*) is introduced to improve the performance of path planning for rescuers in the indoor environment. By comparing the different levels of the target locations, this mechanism guides the RRT to explore purposefully to avoid the algorithm being trapped in the local optimum. Finally, compared with the classical algorithm, the total task time of CPOC in the two examples is reduced by 7.3 % and 27.8 % respectively. DG-RRT* can obtain the effective solution in a shorter time under the premise of ensuring the optimal path length. The results indicate that our algorithm can improve the efficiency of search and rescue route planning as well as the accuracy of the solutions.