Collision avoidance planning of a robot manipulator by using genetic algorithm. A consideration for the problem in which moving obstacles and/or several robots are included in the workspace

Abstract

There have been proposed various approaches for solving the collision avoidance problem of a robot manipulator. However, unfortunately, almost all of the research in this area has so far only dealt with the collision avoidance problem in which moving obstacles are not included in the workspace of a robot manipulator. In this paper, it is shown that path planning and the collision avoidance planning problem of a robot manipulator whose workspace includes moving obstacles can be solved successfully by using a genetic algorithm. Following a brief explanation of Simple GA, the new concept of pseudo-potential is introduced in order to detect the possibility of collision with moving obstacles in the future steps. This concept is utilized in order to evaluate a fitness function of a chromosome which represents a path connecting the starting point with the goal in the workspace of a robot manipulator and construct an appropriate collision avoidance planning. A computer simulation dealing with the collision avoidance problem of two robot manipulators confirms the effectiveness of our approach. >