Pre Compensation Method for Continuous Rotation Angle Error of Robot Joints Under Real-time Prediction
Due to the nonlinearity,coupling effects and positioning errors between robot joint linkages,current robot joint error compensation methods are all post positional,supplemented by the idea of error calibration after errorgeneration.Once the error is too large,it will cause actual losses.A real-time prediction based compensation method is proposed for continuous rotation angle error of robot joints.Using the modified complete parameter continuity(MCPC)method,an inertial coordinate system is established to ensure that the joint motion of the robot meets the constraints in the time series.Introducing Kalman filtering algorithm,integrating multi-source kinematic data,predicting and updating the rotational status of robot joints in joint space,and based on the updated results of the final rotation angle and the comparison with the expected rotation angle,the real-time angle error judgment is achieved,without relying on post position error results.Based on the adaptive variation relation between the domain of fuzzy controller and angle error,a fuzzy rule library is established as compensation historical data to achieve active proactive compensation for joint continuous rotation angle error.Experiments have shown that the proposed method can accurately identify the current rotation angle in robot joint rotation,and the error compensation effect is obvious,which has certain significance for improving the accuracy of robot operations.
NETL
NSTL
万方数据
