| With the rapid development of shipbuilding industry, higher requirements of the welding quality are put forward. By a robot based on rotating arc sensor, the tracking on space curved fillet weld can be well realized. It is also conducive to promote the welding automation of trellis seam in the cabins.On the basis of previous work, the mechanical and hardware parts have been improved, which makes the robot more reliable in performance and more like a product.In this paper, the signal filtering of rotating arc sensor, the algorithm on deviation identification of the welding torch and control algorithm are studied. In terms of the filtering of welding current, this paper does research on median filter, mean filter, threshold processing filter, linear filtering, soft threshold wavelet filter and the fitting filter of cubic curve, and two methods of combination filter are also introduced with one based on soft threshold wavelet filter and the other the process of threshold. As for the algorithm of deviation recognition, the least square method, absolute difference method and minimum distance classifier method are researched. When it comes to the control algorithm, fuzzy control algorithm and multi-mode perceptions are included.With regard to the problem that the weld seam in the factory is long while the distance that the horizontal slider moves is short, research are done on the tracking of curved weld by the robot based on rotating arc sensor, including the tracking on weld seam via the stretch and the indent of the horizontal slider, differential movement of two wheels, coordinated movement between two wheels and the horizontal slider as well as that among the wheels, the horizontal and vertical sliders. It is proved by the experimental results that the algorithms designed are correct.Trellis weld are made up of right-angle fillet weld. In order to implement the automatic welding of the trellis weld by robot, tracking on the right-angle fillet weld without drain holes in the horizontal plane is conducted in this paper, The key techniques as listed as follows: the recognition of the intersection of right-angle fillet weld and path planning. What’s more, this paper also discusses the detection of the endpoint of weld seam, tracking on vertical welding seam and overhead weld. The experimental results show the correctness of the algorithm designed.Finally, in order to improve the welding robot’s ability to adapt to the environment of the factory, This paper also does research on the passing through the welding bead, the existing of butt weld and assembled weld and fillet weld with gap, the recognizing of the initial welding position as well as the multi-pass welding. In addition, corresponding measures are also taken to maintain the accuracy of the tracking on the welding seam so as to improve the intellectualization of the robot, It is showed that by the algorithms designed, the anti-interference ability of robots is obviously improved. |
