本論文提出一個基於粒子群最佳化演算法之運動學校正的方法來改善機械手臂的定位誤差與提升機械手臂的絕對精度。由於機械手臂在長時間使用後會有機械性偏移或是在機械手臂維修後會有精度誤差,這些現象將導致機械手臂末端點的實際姿態與命令姿態會有定位誤差,所以機械手臂之運動學校正是一個重要的研究議題。常用之機械手臂的運動學校正方法大部分利用定位誤差資訊來反推機械手臂的連桿長度誤差及關節角度誤差,然而機械手臂末端點的實際姿態是很難被直接測量出來。因此,本論文提出一個基於粒子群最佳化的方法來求機械手臂之D-H參數表上參數的近似最佳解,這個方法可以有效的減少D-H參數表上參數之實際值與理論值兩者之間的誤差。最後,一些實驗結果可以說明,本論文所提出之機械手臂運動學校正方法確實可以有效的減少機械手臂的定位誤差,進而提升機械手臂的絕對精度。
In this thesis, a kinematic calibration method based on a Particle Swarm Optimization (PSO) algorithm is proposed to improve the position error and enhance the absolute accuracy of robot manipulator. The robot manipulator will have some precision error due to it works for a long time or after maintenance. They will cause some positioning errors between the actual posture of the end-effector of robot manipulator and the command posture. Thus, the kinematic calibration of the robot manipulator is an important topic. Most kinematic calibration methods for the robot manipulator use the reverse method to correct the link length error and the joint angle error of the robot manipulator. However, the actual posture of the end-effector of robot manipulator is difficult to be measured directly. Therefore, a PSO-based method is proposed to select approximate optimal values of parameters in the D-H parameter table for the robot manipulator. It can effectively reduce the error between the actual value and the theoretical value of the D-H parameter table. Finally, some experimental results are presented to illustrate that the proposed method can effectively reduce the positioning errors and enhance the absolute precision of the robot manipulator.