Automatic UAV Swarm Task Planning in Cooperative Region Coverage Detection based on Greedy Policy

Unmanned Aerial Vehicle(UAVs) swarm has great advantage over traditional equipment in cooperative detection scenario for its easy-maneuverability, no human injury and low cost, etc. As a representative task in cooperative detection, region coverage has widely applications in environmental monitoring, search and rescue, etc. In UAV cooperative detection tasks, the most critical step is task planning, which has direct impact on the overall detection performance. The target of task planning is to generate planned actions and flight route for UAVs to complete specific detection task according to UAV swarm locations, sensor ability, task region, etc. However, traditional task planning methods for UAV cooperative detection that based on evolutionary computing or reinforcement learning always need plenty of time for getting planning results. In this paper, we proposed a top-down task planning algorithm based on greedy policy to tackle this problem. The core idea of the proposed method lies in that we choose optimal detection trace from all trace candidates during each planning step in a greedy manner via a predefined performance indicator. Moreover, we also proposed a simple but effective procedure for generate detection trace candidates by corner points and nearest border points extraction. To evaluate the effectiveness of the proposed method, we conducted comprehensive experiments for the representative swarm detection task region coverage. Experiment results demonstrated the effectiveness of the proposed method and superiority over traditional methods on task planning speed.