Repulsive Force Control Based on Distance Computation Algorithm

Abstract

Computer-aided repulsive force control of collision avoidance is presented in this paper. A repulsive force is artificially created using the distances between the robot links and obstacles, which are generated by a simplified distance computation algorithm. This distance computation algorithm is based on the Gilbert-Johnson-Keerthi algorithm. Control gains in the repulsive force control model are selected based on traditional control design and genetic algorithms. Results on shortest distance computation and collision detection are presented. Real-time manipulator collision avoidance control has achieved. A repulsive force gain is introduced through the approaches for definition of link coordinate frames and kinematics computations. The safety distance between objects is affected by the repulsive force gain. Safety zone can be adjustable by the repulsive force gain which is selected by a specified fitness function of the genetic algorithm.