HEXA-parallel-structure calibration by means of angular passive joint sensors

Abstract

Geometric calibration is a well known technique to improve absolute accuracy of robotic structures. This is to adjust the kinematic parameters in the robot controller model to the actual robot parameters which differ from the nominal values due to e.g. manufacturing errors or thermal effects. Usually redundant measurement which is required for calibration is obtained by external measurement devices. Parallel robots offer the opportunity to perform a more elegant calibration by means of sensors placed in the passive joints. This approach which offers several advantages compared to conventional calibration is known as self-calibration since the passive joint sensors are part of the robot system itself. This paper presents a technique for self-calibration of the HEXA-parallel-structure. Due to consequent utilization of the passive joint sensors the approach works computationally more efficient than others. Particular problems arising during the calibration process are addressed and corresponding solutions are proposed. Finally calibration success is demonstrated by means of simulation studies.