SENSOR-BASED 2-D POTENTIAL PANEL METHOD FOR ROBOT MOTION

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 w orkspace 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 potentia l field. The computational complexity of the pertinent algorithms just ify the applicability of the approach in real-time. Simulation results are included.