Fast Collision-Free Trajectory Planning for Robot Manipulator Based on Distortional Configuration Space

Abstract

The paper describes a nobel method to plan the collision avoiding trajectory for a robot manipulator in the existence of obstacles. The method is based on Distortional Configuration Space, proposed in this paper, which a conventional C-Space is transformed into by artificial attractive potentials. In the Distortional C-Space, all the forbidden areas are eliminated by the potentials, and then a collision-free path is planned without considering the interference with them. In order to obtain a feasible trajectory, it is important to define appropriate potentials in the C-Space. The definition of the potentials depends on the positions and the shapes of forbidden areas, though it needs enormous calculation time to get exact information of C-Space. By applying the approximate forbidden area and by determining the appropriate potential function, the processing time gets shorter, and then the proposed method realizes a quick trajectory planning. The validity of the method is confirmed in the results of numerical experiments.