Optimal Waypoint Placement for Road Surveillance using Unmanned Aerial Vehicles Suzanne McDonald ∗ University of Maryland, College Park, Maryland, 20742, USA Raymond W. Holsapple † and Derek Kingston ‡ Air Force Research Laboratory, Wright-Patterson AFB, Ohio, 45433, USA In this paper we consider an intelligence, surveillance and reconnaissance (ISR) mission where a human operator and a team of unmanned aerial vehicles (UAVs) are tasked to patrol a road. There are numerous inherent ISR benefits for patrolling a road. For example it might be beneficial to track the suspected movements of enemy troops in a forward operating location. Another example might be to gather intelligence that could then be used, possibly not in real time, to perform change detection, a method that might help with the detection of improvised explosive devices (IEDs). The layout of the road, with respect to a fixed known reference point, is assumed to be known a priori. For example the latitude and longitude coordinates of various points along the road would be sufficient. Most UAV autopilots make use of waypoint navigation to control the vehicle. This work investigates one method of optimally choosing the waypoints for the UAVs along the road. The cost function used minimizes the cross-track area error. I. Introduction F or the purpose of explanation, let us consider a specific scenario. Let us suppose that a military UAV unit is tasked with patrolling a network of roads in a forward operating location. One of the main tasks of this unit is to detect IEDs. The task of detecting roadside bombs and IEDs is one that has plagued the modern warfighter for many years. The use of IEDs has drastically increased in the last decade and is now at the forefront of modern terror tactics. The detection of IEDs is critical to success in the battlefield, and the development of technology that might assist in this detection is equally critical. We can suppose that to realistically perform their job, this unit wishes to visit each point along the known road once every hour. This will help minimize the amount of time that each point along the road is away from the camera. This unit is responsible for patrolling 120 miles of road, 60 miles down and 60 miles back. Multiple UAVs may be used over the same road segment in order to achieve the time objective of one flyover per hour. This research specifically deals with the optimization of the placement of waypoints for a single autopilot in order to follow a known road segment. Most UAV autopilots use GPS and waypoints for navigation. The need for waypoint optimization is a direct result of the limitation of some autopilots. For instance, an autopilot may only store 100 waypoints, which may not be able to be used in reverse order. Due to communication latency, it may also be undesirable to change the waypoints during flight. If we eliminate a few waypoints that are needed for takeoff and landing and taking into consideration that the autopilot cannot use points in reverse order, the scenario described leaves only about 46 waypoints at our disposal to navigate the aircraft over possibly 60 miles of road. A road would not need to be excessively curvy for a naive placement of the waypoints to be insufficient. The location of these points is optimized to keep the UAV as close to the road as possible. As the number of turns in the road increases dramatically, a mountainous road for example, this number of waypoints begins to appear increasingly smaller. ∗ Graduate Research Fellow, Department of Electrical and Computer Engineering, College Park, MD 20742. † Research Mathematician, Control Science Center of Excellence, Wright-Patterson AFB, OH 45433, AIAA Member. ‡ Research Engineer, Control Science Center of Excellence, Wright-Patterson AFB, OH 45433, AIAA Member. 1 of 8 American Institute of Aeronautics and Astronautics