UAV Obstacle Avoidance Algorithm Based on Ellipsoid Geometry

An avoidance algorithm is proposed in this paper which works by generating avoiding waypoints, within the original predefined waypoints, when the vehicle finds obstacles obstructing its flightpath. The approach developed here bases the search for avoidance path on the utilization of ellipsoid geometry as a restricted zone containing the obstacle. The restricted ellipsoid zone is established by considering the identified obstacle geometry information, and further the ellipsoid becomes the basis for computing the new waypoints for avoiding the obstacle. These avoiding waypoints determined by computing the contact points between the ellipsoid and planes the normal vector of which are corresponded to the vehicle velocity vector. The information about geometry and dimension of the ellipsoid are computed from the information about obstacle geometry, which is assumed to be available, either from mission database or predicted from UAV's ground detection system. In the development process, the algorithm is constructed in MATLAB environment and then simulated and analyzed in some scenarioed cases representing possible situations when an UAV has to avoid obstacles during its flight. This algorithm is intended to be integrated into the guidance system of UAV.