Offline and online strategies to improve pose accuracy of a Stewart Platform using indoor-GPS

Abstract

The pose accuracy of a robot manipulator may be improved by assessing and correcting systematic errors. Both offline and online strategies can be considered. To date, there has not been a solution for the online pose error correction of parallel manipulators. Moreover, offline strategies using indoor-GPS as reference measurement system have not yet been investigated. In this paper an optimization-based kinematic calibration method and an online correction technique are proposed and implemented for a low-cost Stewart Platform. In both cases, an indoor-GPS system was used as reference measurement equipment. Performance of both strategies are compared to a kinematic calibration method based on direct parameter measurement. Pose errors are evaluated for each strategy using a robotic total station. Performance of the optimization-based calibration and the online correction technique were similar and better than the direct parameter measurement calibration. Both techniques resulted in average pose errors less or equal to 0.3 mm and 0.05°. The proposed strategies may be adapted to other similar parallel manipulators and are applicable to large sized equipment.