Distributed Cooperative Search Algorithm With Task Assignment and Receding Horizon Predictive Control for Multiple Unmanned Aerial Vehicles

For target search using multiple unmanned aerial vehicles (UAVs) while knowing the probability distribution of the targets, a distributed cooperative search algorithm aiming to minimize the search time is proposed. First, an importance function for the representation of the environment is designed. Second, a mission planning system (MPS) is proposed, consisting of preliminary planning, task assignment, and post-planning layers. In the MPS, the search region is divided into a series of sub-regions of different sizes by centroidal Voronoi tessellation; these are regarded as subtasks assigned to the UAVs. The loading of the MPS improves the performance of global planning of the UAVs. Finally, receding horizon predictive control is used to plan the paths of the UAVs online. Moreover, the conflict between the requirements of target search and connectivity maintenance of the UAVs is mitigated using the minimum spanning tree strategy to optimize the communication topology while considering the communication cost when evaluating the tasks. The results of Monte Carlo simulations show that the introduction of the MPS into the traditional cooperative search framework effectively improves search and coverage efficiency.