This paper aims to present a complete relative pose error model for robot calibration, considering both the relative distance error and the relative rotation error of the robot end-effector, which can improve calibration accuracy.
In this paper, the relative distance error model and the relative rotation error model of robot calibration are derived by ignoring high-order nonlinear errors, and the two models form into a complete relative pose error model. Besides, mathematical expectation of the nonlinear errors is calculated, indicating that they have little influence on calibration accuracy.
Comparative experiments have indicated that the proposed complete relative pose error model does better in robot calibration than only the distance error model.
The main contribution of this paper lies in the derivation of the relative rotation error model, which helps to form a complete relative pose error model for calibration. The proposed method improves calibration accuracy, with avoiding identifying the transformation matrix between the measurement system frame and the robot base frame.
