Study On Kinematic Calibration Of The6-dof Parallel Adjustment Device

The calibration device of wind tunnel balance is an essential equipment in modernwind tunnel air dynamic experiment. Its accuracy will directly affect the results of windtunnel test. The adjustment mechanism is an important part of the platform, it is used toadjust the balance to the initial position or the designated position. The reset accuracydirectly related to the accuracy of the calibration. So it is an important research directionto raise the reset precision of adjustment mechanism in wind tunnel balance calibration.Parallel mechanism has the advantages of high static stiffness, small accumulated errorand higher precision. In order to improve the accuracy of reduction and adjustment, thispaper will study about the Stewart platform and make the kinematic calibration of theStewart platform. The kinematic calibration includes the following four steps: toestablish the error model of parallel mechanism, parameter measurement, parameteridentification and error compensation.Based on robot mathematics, a method of the inverse kinematics and kinematics ofthe6-DOF PKM is according to the structure characteristics of6-DOF parallelmechanism fixed platform and moving platform. This method provides the theory basisfor the following research work.From the structure of the parallel mechanism, the main error sources are analyzedfrom the all error sources that may effect on the final reduction and accuracy. In view ofthese error sources, the error model will be established by using differential method andnumerical method. The differential method is very complex with the repeateddifferential item. The numerical method is better because of difference instead ofdifferential.A complete pose measurement scheme is designed based on the error model andthe parameter identification model. The identification model based on the least squaresmethod. With Matlab software, the identification simulation based on Guass-Newtonalgorithm. The numerical method to identify the model is better and faster than thedifferential method.Finally, the end-effector motion accuracy is improved by error compensation with the mechanism errors, which are obtained by identification of structural parametersby measuring the parameters. The measurement accuracy, selection of control precisionand pose measurement are all have a great influence on the terminal pose compensationaccuracy. This method provides a theoretical basis for selecting measuring instrumentsin measurement experiment. From the calibration simulation based on the Stewart resetadjustment mechanism, using the full pose relative measurement, the error modelingmethod is correct and the parameter identification and error compensation arereasonable and efficient.