Manipulator tracking technology based on FSRUKF

Aiming at the shortcoming that the traditional indus-trial manipulator using off-line programming cannot change along with the change of external environment,the key technolo-gies such as machine vision and manipulator control are studied,and a complete manipulator vision tracking system is designed.Firstly,Denavit-Hartenberg(D-H)parameters method is used to construct the model of the manipulator and analyze the forward and inverse kinematics equations of the manipulator.At the same time,a binocular camera is used to obtain the three-dimensional position of the target.Secondly,in order to make the manipulator track the target more accurately,the fuzzy adap-tive square root unscented Kalman filter(FSRUKF)is proposed to estimate the target state.Finally,the manipulator tracking sys-tem is built by using the position-based visual servo.The simula-tion experiments show that FSRUKF converges faster and with less error than the square root unscented Kalman filter(SRUKF),which meets the application requirements of the manipulator tracking system,and basically meets the application require-ments of the manipulator tracking system in the practical experi-ments.