Research On Calibration Algorithm And Experimental Verification Of 6R Serial Robot Based On Product-Of-Exponential

Nowadays industrial robots have high repeatability accuracy but with low absolute accuracy, thus they cannot meet the requirements of off-line programming technology. Calibrating the kinematic model of a robot can promote robot absolute accuracy. Traditional calibration method based on D-H model dissatisfying integrity, continuity and non-redundancy is not suitable for robot calibration. Product of Exponential(POE) model otherwise owns these three figures and its geometric meaning is clear. It is easy to establish the model and compensate parameter.Measurement coordinate system is not consistent withrobot base coordinate system; therefore we improve the traditional calibration method based on global POE model. Transformation from measurement coordinate system to robot base coordinate system is regarded as a set of parameters to be added to the error model. These parameters are identified together with the geometric errors of a robot so that the measurement error caused by establishing the robot base coordinate system is avoided.Due to the difficulty of orientation measurement of robot end-effector, the error model of position measurement is derived based on the error model of pose measurement of local POE, which avoids the error caused by the orientation measurement.To compensate the robot error after calibration, by analyzing the properties of the two calibration algorithms, we work out how to use the parameters obtained after the identification of the parameters. For global POE calibration model without redundancy, a method to transfer POE parameters into D-H parameters is introduced.But this method does not apply to local POE which is redundant in mathematics. The numerical inverse solution based on local POE is introduced. Calculate the true joint angle by using the modified geometric parameters obtained by calibration so that the robot can reach the ideal position.Finally, the experimental verification method is designed to verify the calibration of position measurement. The experiment includes calibration principles of the laser tracker, data acquisition method and experimental results verification.The experimental results show that the absolute positioning accuracy of the robot is significantly improved.