Abstract
It is difficult to solve inverse kinematics of an articulated redundant robot, since this problem is nonlinear. In this paper, a method for solving this problem using genetic algorithm (GA) is proposed. It is aimed to achieve an precise solution within allowable short time by applying a heuristic search to GA. In this search, three weights of position, orientation, and collision avoidance in an evaluation function are gradually changed as the search progresses. Also, the search range of each joint angles are gradually changed as the search progresses. As a preliminary example, the effectiveness of this method is confirmed by computer simulation in case of that a 8 DOF robot passes through a gap on a thick wall.
