A method for solving obstacle avoidance for a redundant robot is proposed in the present paper. Extra degrees of freedom (DOF) of a redundant robot are effective for realizing an objective position and orientation of its end effector (referred to hereafter as "pose"), while the robot is avoiding obstacles. The path should be planned so as that the robot can avoid obstacles and realize the desired goal pose. The models of six elementary types of obstacles are assumed, taking account of real environment The path planning method proposed herein is divided into three procedures as follows: 1) solving inverse kinematics by an analytical method for avoiding six elementary types of obstacles, 2) solving inverse kinematics by a semi-analytical method for realizing a goal pose, and 3) generating a path from a start pose to the goal one. A computational simulator for a redundant robot to avoid arbitrary obstacles based on these procedures is developed.