Study on Multi-UAV Task Clustering and Task Planning in Cooperative Reconnaissance

For multi-UAV cooperative reconnaissance to enemy´s multi-task points, because of multi-task, reasonable clustering is needed and the task clustering model should be established. In this paper, the task planning model is established according to task clustering of each UAV, and the sequence of task execution is determined. Reasonable task clustering optimization index is put forward. Task allocation is proposed based on improved K-means clustering algorithm of simulated annealing. The shortest path task planning is designed using the simulated annealing algorithm, which makes multi-UAV relatively balanced in the assignments, the task group in the group centralized distribution, inter-group distribution scattered and the total cruise time shortest. Simulation results show that the task clustering is well achieved and the optimum task planning program is obtained. The validity of the model and algorithm is verified and the algorithm has certain theoretical and practical value.