Overhead weld tracking by robots based on rotating arc sensor

In order to improve the quality and efficiency of welding, implement automatic tracking overhead weld by robots is demcomel require. The method of composite filter is used to filter out noises. The least square method is used to fit sampling points from the 16th to the 48th, and the slope of the fitting line is regarded as a deviation. The deviation and the sum of three consecutive deviations are regarded as the input of the fuzzy controller, and the pulse number corresponding to the horizontal slider movement speed is regarded as the output of the fuzzy controller. It is the endpoint of weld seam if the absolute value of three consecutive deviations are all greater than 0.9. Last, experiments of overhead weld tracking, experiments of overhead weld endpoint detection and passing drain holes have been done. The results show that the robot can track overhead weld with high accuracy, good reliability, accurately identify the endpoint of overhead weld and pass drain holes by using the designed algorithm.