Abstract: For reasons of the vibration of robot, the rough weld seam and electromagnetic disturbance of welding machine, the identifying force signals of welding seam(IFSWS) become unstable. By analyzing the model of IFSWS, the Kalman is used to filter the IFSWS in remote teaching. The next IFSWS is accurately predicted according to the equation of the state, the equation of the observation, the filtering of the state predicting, gain matrix of the filtering and the covariance matrix of the predicting state in Kalman filter. The experimental results show, by Kalman filtering, that the predicting error of the IFSWS is the smallest, the detecting precision of the IFSWS is increased, and the precision of welding seam identifying is improved in remote teaching.

Key words: Force sensing, Kalman filtering, Remote teaching, Remote welding, Welding seam identifying