This study aims to address the issues of low positioning accuracy in industrial robot end-effectors and the high costs as-sociated with traditional calibration methods, proposing a novel robot self-calibration approach based on a line-structured light sensor.
First, the local product of exponentials (POE) method is adopted to establish the kinematic error model of the robot. The kine-matic parameter errors and hand-eye errors are concentrated in the local rigid body motion pose change matrix, which greatly simplifies the kinematic error parameters. Second, the robot error model is used to construct a calibration algorithm based on distance constraints, achieving the identification of the robot kinematic parameter errors and hand-eye errors. Finally, the accura-cy of the self-calibration algorithm is verified through simulation, and a self-calibration experimental platform is set up to carry out experiments.
Experimental results indicate that, after calibrating and compensating the robot with this method, the average distance error of the robot drops from 0.69 to 0.13 mm, a decrease of 81%.
It overcomes the disadvantages of high cost and narrow application scope of traditional calibration methods, and also circum-vents the singularity problem in D-H modeling. This is of great significance to the research on robot calibration.
