FRPSO: Inverse Kinematics Using Fully Resampled Particle Swarm Optimization

摘要

This paper proposes a novel approach for the popular Particle Swarm Optimization (PSO) algorithm to solve the inverse kinematics problem for a robotic manipulator. Our proposed approach is based on a full resampling of the particles, which gave name to the method: Fully Resampled PSO (FRPSO). By doing that, we reduced the dimension of the problem by a half since the velocity vector used by the PSO algorithm was no longer needed, and we improved the robustness of the algorithm to local minima. In order to evaluate our proposed method, we compared it to three PSO versions by solving the inverse kinematics of a planar 4-DOF robot and a 3D 7-DOF robot. The results demonstrate the capability of the FRPSO algorithm since it performed better in all tests.