Curvature-constrained traveling salesman tours for aerial surveillance in scenarios with obstacles

The curvature-constrained traveling salesman problem with obstacles deals with finding a minimum length tour which includes a set of landmarks and avoids obstacles, for a kinematically constrained vehicle. Its great practical importance is mainly due to surveillance tasks of unmanned aerial vehicles. The problem constitutes a combination of the well-studied Dubins traveling salesman problem and the flight path planning problem. We present heuristic algorithms that are based on different strategies of extending a tour by inserting new landmarks. Each insert operation comprises the optimization of overflight directions for the given sequence of landmarks. Path finding between landmarks is done by a discrete routing model. It allows arbitrary flight directions and turn angles as well as maneuvers of different strengths, thus fully exploiting the flight capabilities of the aircraft. The performance of the algorithms is evaluated for agile and less agile aerial vehicles, using randomly generated scenarios with obstacles of different size and number.