Parallel region coverage using multiple UAVs

Low flying, small endurance UAVs are well-suited for region coverage over airbases or in urban zones since they are cheap, highly maneuverable and expendable. In this paper we consider the problem of minimizing the time needed to cover the region of interest, a contiguous rectilinear polygonal workspace, /spl Pscr/, using /spl eta/ UAVs. Our approach is based on partitioning /spl Pscr/ into /spl eta/ interior-disjoint polygons, P/sub i/ and making a one-to-one assignment of the polygons to the UAVs. The partition comprises polygons that are (a) rectilinear and (b) contiguous (c) whose area-ratios equal to the ratio of rates at which the UAVs can do coverage (d) and whose boundaries include points at which the respective UAV begins coverage. Our work appears to be the first that directly generates partitions that satisfy all the aforementioned conditions. We discuss the operational requirements that motivate this particular partitioning problem. We prove that our algorithm runs in O(N log N + N + /spl eta//sup 2/N) time; N is the complexity of /spl Pscr/.