Robust Hybrid Control for Unknown Obstacle Avoidance
Michael Naderhirn, Peter Langthaler, Luigi del Re · 2009
In this paper, we solve the problem of robust collision free motion planning for a nonholonomic planar vehicle in a stationary unknown envoirnment. The vehicle tracks its own position, the heading, and is equipped with a sensor with limited sensor range and uncertainties to detect stationary obstacles along the path.The vehicle moves with constant velocity and a bounded curvature. A robust hybrid control policy is proposed which ensures stationary obstacle avoidance in the presence of measurement uncertainties e.g. a radar cone. Furthermore, the proposed policy is geometric explainable. The robustness requires no numerical optimization, it simply results in a min-max algorithm.