Self-calibration of the 6-DOF parallel robot TRIPLANAR by identification of the geometry parameters

This paper presents a model-based procedure for the self-calibration of the parallel robot, TRIPLANAR, that gets by on just a relatively simple measuring system. The measuring device continuously quantifies the distance between the tool-center point (TCP) that is adjustable in all 6 degrees of freedom and a fixed reference point in the inertial coordinate system (ICS). After just one run over a suitable measuring trajectory and recording of the measuring data (distance TCP-ICS and actuator positions) a parameter identification is performed by minimizing the deviation between the measured pose (position and orientation) and the one that was determined from a model. The identified parameters are taken into account in an inverse kinematics model that is used for the feedforward control of the TriPlanar; thus positioning accuracy can be greatly increased.