Kinematic Parameter Identification for a Parallel Robot with an Improved Particle Swarm Optimization Algorithm

Abstract

The spacecraft docking motion simulation system for on-orbit docking plays a very important role in some theoretical research and engineering application fields. The parallel robot utilized in the spacecraft docking simulation system requires high positioning and orientation accuracy to achieve better simulation results. A novel kinematic parameter identification method with an improved particle swarm optimization (PSO) algorithm is proposed to enhance positioning and orientation accuracy of the parallel robot. A fitness function is established using these residuals between the measured and computed poses by a coordinate measuring machine and forward kinematics. The kinematic parameter identification problem is turned into a high-dimensional nonlinear optimization in which the unknown kinematic parameter errors are regarded as optimal variables. The optimal variables are solved by the proposed improved PSO algorithm. The mean values of the positioning and orientation errors are reduced from 4.3268 mm and 0.2221 deg to 0.7692 mm and 0.0674 deg, respectively. The proposed kinematic parameter identification method increases the positioning accuracy mean by 22.26% and the orientation accuracy mean by 32.80% compared with the least squares method. The kinematic parameter identification method with the improved PSO algorithm can effectively enhance positioning and orientation accuracy of the parallel robot for docking motion simulation.