Collision Avoidance of Robots by Artificial Force Field Around Geometric Primitives Using Two Range Image Sensors

This article presents an approach to collision avoidance of industrial robots. It is based on artificial potential fields or force fields generated by virtual charges which are placed on the obstacles. The obstacles are detected by range image sensors. To avoid fault detections e.g. as a result of occlusions and distortions two cameras are used to supervise the robot work space. An algorithm is worked out which eliminates the distortions in the camera images. After this, a method is developed to calculate a 3D image of the environment. For fast computation of the artificial forces the obstacles are classified into geometric primitives. The virtual charges are placed in 3D on the objects. From the charge positions and their potential value an artificial force can be calculated which acts on the end-effector. The force corrects the primary planed robot path to avoid collisions with the obstacles.