Parameter Calibration of Dimensionality Reduction DH ErrorModel Based on Robot Task Space

To simplify the kinematic error model, this paper proposes a kinematic parameter calibration method based on the dimensionality reduction DH error model on the robot task space. Through the global sensitivity analysis, the kinematic parameters with smaller sensitivities in the DH kinematic model of the 7-DOF robot can be eliminated, and then the dimensionality reduction kinematic error model can be derived. The space convex hull algorithm is utilized to determine the scope of the robot workspace, and the least squares method is applied in parameter calibration. The experimental results show that the parameter calibration using the dimensionality reduction kinematic error model can simplify the error model as well as reduce the pose error of the end effector. The dimensionality reduction model also reduces the computation time when compared with the standard error model. In comparison with the calibration method of selecting the calibration points from the whole workspace, selecting points from the robot task space will slightly improve the accuracy of the end effector's pose.