A Calibration Method of Robots by Eliminating Redundant Parameters Based on Jacobian Matrix

This paper presents a new identification method for error parameters of robot calibration model by eliminating the redundant error parameters based on Jacobian matrix. Firstly, the kinematic error model of the robots based on modified Denavit-Hartenberg (D-H) method was established. Secondly, the error parameters were analyzed based on Jacobian matrix in detail. And a new identification method by eliminating the redundant error parameters was presented. Finally, the calibration experiments of a 6-DOF industrial robot were carried out. The results showed that the error compensation rate was more than 71%. In addition, the positon accuracy has been further improved after compensation based on the presented identification method than that based on traditional method. Hence, the calibration method presented in this study is sensible and effective, and could be used for the error compensation and improve the position accuracy of the robots.