Comparative Study of Two Pose Measuring Systems Used to Reduce Robot Localization Error
Marek Franaszek, Geraldine S. Cheok, Jeremy A. Marvel
Accurate pose determination of an industrial robot end-effector is critical in many applications. The pose is usually derived from a forward kinematic model of the robot's arm and is dependent on how well the model parameters are determined. Localization error still remains even after robot calibration and, different compensation techniques have been developed. These techniques require independent determination of the end-effector pose by external pose measuring systems. In this paper, a method for reducing robot position and orientation error is used for two datasets acquired with a camera-based motion capture system and a laser tracker. Both systems have very different characteristics which affect accuracy of the measured poses. Experimental results show that the method can reduce the median position error comparably well for both pose measuring systems (97% and 92%, down to 0.3 mm and 0.43 mm). However, the reduction of the median orientation error is different for both systems (57% and 88%, down to 0.27° and 0.04°). Simultaneous reduction of both position and orientation robot error is possible when the pose measuring system with right characteristic is used.