Abstract: A method based on an improved particle swarm optimization algorithm was proposed to address the non-uniqueness problem of the inverse kinematics of a gravity compensation redundant manipulator and to obtain the optimal inverse kinematics solution for the manipulator with minimum displacement of the counterweight mass. First, the relationship between the displacement of the counterweight mass, joint angles, and servo compensation torque was analyzed using forward kinematics. Based on this, an inverse kinematics equation was constructed with the minimum displacement of the counterweight mass as the optimization objective and the position error accuracy as the constraint condition. An improved particle swarm optimization algorithm was used for the solution. The experimental results show that the optimal inverse kinematics solution for the manipulator with the minimum displacement of the counterweight mass as optimization objective was obtained, and that the non-uniqueness problem of the inverse kinematics was solved. This result is significant for follow-up experiments on the gravity-compensated redundant manipulator.

Key words: gravity compensated redundant manipulator, counterweight mass displacement, inverse kinematics, particle swarm optimization algorithm