An evolutionary algorithm for multiple waypoints planning with B-spline trajectory generation for Unmanned Aerial Vehicles (UAVs)

Abstract

In this work, we focus on multiple waypoints planning that take into account the stealthiness of autonomous UAVs; generating stealthy waypoints through a region laden with enemy radars. The important goal of waypoint planning for Unmanned Aerial Vehicles (UAVs) is to generate an optimal path for UAVs to visit all of the specified waypoints. We use the Traveling Salesman Problems (TSP) optimization to solve the visitation sequence planning problem for UAVs. The aim is to find the optimal solution that minimizes the distance travelled by the UAVs with low radar risk; under the strict constraints that each waypoint is visited exactly once by the UAVs, and the flying altitude does not exceed the specified threshold so as to minimize the radar exposure. The generated solution is then refined using a series of cubic splines in lieu of the straight line segments, resulting in a cost-effective path traceable by the UAVs. The shape of the trajectory is approximated using cubic B-splines. Simulated results are presented, illustrating the potential benefits of such approach. ©Memetic Computing Society.