Abstract: The dynamic position variable of large diameter revolving part is not easy to measure precisely and cheaply, however, the accuracy of the dynamic position is of great significance to controlling the movement of the key revolving part. A novel real-time detection method is proposed for the dynamic position of large diameter revolving part rotating with constant velocity, constant acceleration or variable acceleration. In order to estimate the position variable, multisensor information fusion based on adaptive fading Kalman filter (MSIF-AFKF) algorithm, which combines traditional information fusion algorithm with an adaptive fading factor, is established to fuse multi-grating sensor information. The method based on MSIF-AFKF algorithm is simulated by Monte Carlo and also used to detect the rotation angle position in the bed head and bed tail of horizontal riveting machine, the both results show that the method can significantly improve the precision.

Key words: Adaptive fade factor, Dynamic position detection, Kalman filtering, Large diameter revolving part