Practical obstacle avoidance using potential field for a nonholonmic mobile robot with rectangular body

A real-time obstacle avoidance algorithm based on sensor data is required for mobile robots. When a mobile robot is nonholonomic and it has polygonal body, the method using configuration space is usually applied. But, it is complex and it needs much computing power. On the other hand, artificial potential field is often used for real-time obstacle avoidance, however, most of them consider a mobile robot as an ominidirectional movable point. Therefore, we propose a new method of practical obstacle avoidance for a mobile robot with rectangular body. Action points of repulsive forces from obstacles are located in both the front and rear of robotpsilas body. Their forces are generated according to the distances between obstacles and robotpsilas body. Rear forces are transferred to the front by inverting and the resultant force moves the robot. This method is very simple and effective.