Sensor-based 2-D Potential Panel Method for Robot Motion Planning

Abstract

Summary A potential panel method is proposed to solve simultaneously the path planning and collision avoidance problem for a mobile robot operating in an uncertain obstacle filled environment. The problem is solved in three steps: (1) transform the arbitrary shaped obstacles in the 2-D workspace into simple convex polygons; (2) generate a local minima-free potential field on the workspace; (3) generate a streamline from the starting position towards the goal position in the artificial potential field. The computational complexity of the pertinent algorithms justify the applicability of the approach in real-time. Simulation results are included.