Scheduling UAV Surveillance Tasks, Lessons Learned from Trials with Users

Multiple unmanned aerial vehicles (UAVs) operating in the same area can divide up a set of observation tasks between them to optimize the number of tasks that they can perform. QinetiQ researchers have developed a scheduler as part of a large assessment and test system used for developing new operational concepts. This paper describes a scheduler based on an Anytime A* search using greedy search to generate both heuristic values and candidate schedules. We describe how the scheduler is used as part of a decision support system which makes real-time adjustment of task assignments. In a series of trials the use of the scheduler has been shown to increase the overall number of tasks completed by 50%.