An Improved Hand-eye Calibration Method Based on 3D Position Information

Abstract

Hand-eye calibration can unify the robot coordinate system and the measurement coordinate system, but the process is easily affected by the geometric parameter error of the robot body and the drift of joint angles. Therefore, this paper proposed an improved 3D-position-based hand-eye calibration method, a refinement of the existing process. By controlling the manipulator to only rotate it's single joint, this method fitted the axis of the rotating joint and used the axis properties to obtain part of the rotation information of the hand-eye relationship. As the process was not affected by the geometric parameters of the robot body and the drift of joint angles, this part of rotation information was used to correct the hand-eye calibration results. A simulation and experiment were conducted to prove the ability of the proposed method to effectively improve the accuracy of the hand-eye calibration.