부산대학교 도서관

The core welding of remote control is the follow-up control of the mechanical arm, and the control accuracy and time delay between the human body motion and the welding process of the mechanical arm are difficult in the problems of the follow-up control. This paper designs a follow-up control algorithm based on the fusion of human motion perception and unscented Kalman Filter to solve the problem between the human body and the follow-up of the mechanical arm, and the nonlinear prediction function are studied. Through the improvement of the a 1 bove two aspects, the problem of uncoordinated movement of the human body and the mechanical arm during the follow-up control process is solved. Compared with Kalman Filter and unscented Kalman filter, the variance and the mean value of AUKF in prediction accuracy are 6.758 and 1.791 respectively. The mean time delay point of AUKF is 1.8. Therefore, the welding quality meets the demands of the industry.