Distributed Area Partitioning for Multi-UAVs in Cooperative Monitoring Task

Abstract

In resource and task allocation problems, reasonable allocation can make full use of the UAVs(Unmanned Aerial Vehicles) and bring massive improvement in efficiency. Specifically, the region distribution policies are studied in this paper. In order to make the workload match the ability of the UAVs, the environment is expected to be divided into several subregions of certain sizes. And the group of UAVs should always keep monitoring the whole region no matter how many UAVs are left, which means that the distribution of the area is supposed to react timely even when some UAVs quit the task. Additionally, for the convenience of single UAV's monitoring, the region of one partition is expected not to be too scattered. Therefore, both the area and the shape of the partition are considered in the partition strategies. In this paper, a power diagram based area partitioning method is proposed to rationalize both the allocation of the areas and the shape of each partition. By adjusting the weights of the partitions and the positions of the generators iteratively, the areas and shapes of the partitions can reach an expected and stable state finally. Simulation results are provided to verify the effectiveness and feasibility of the proposed method.