Study On Identification And Tracking Control Of Space Surfaces Weld Seam Of Autonomous Mobile Robot

In the steel structure welded parts of large ships,there are many problems such as complicated structure of the truss structure and many kinds of box type,different welding position and shape,small welding working area,low welding efficiency and so on.Industrial robotic arm welding robot,because the operating base can not move and the elongation of the arm is limited,resulting in welding operation area and space constraints,is difficult to adapt to large-scale steel structure of the adaptive dynamic seam tracking requirements.This paper through the ship’s welding process to do in-depth analysis,the use of welding machines,industrial computers,wheeled mobile intelligent welding robot and rotary arc dynamic real-time tracking technology combined to compensate for the processing or installation of the error and welding thermal deformation of the workpiece,which provides a reference for solving the real-time dynamic tracking problem of large-scale structural workpiece in the long-distance,large-scale and complex space trajectory.(1)Real-time dynamic tracking of space surface welds consisting of four-wheel differential travel platform and five-degree-of-freedom mechanical adjustment arm based on Pioneer 3: Four-wheel differential mobile platform is responsible for steering and walking coarse adjustment;five-degree-of-freedom mechanical adjustment arm Which is responsible for the precise adjustment of the vertical,horizontal and inclination of the gun’s attitude.The end-adjusting slider controls the rotating arc torch to perform linear periodic scanning motion to realize the deviation characteristic signal of the space weld.(2)The joint motion model of the wheeled mobile car and the mechanical adjustment arm is established,and the dynamic adjustment matrix of the attitude of the gun is deduced,which provides the mathematical basis for the welding gun attitude coordination control of the welding robot.(3)The linear periodic scan path of the rotating arc is optimized for the shortcomings of the scanning signal of a single sensor,and the extraction area of the original characteristic signal is optimized.A method based on the limit,rotation period mean and Bayesian transform and the IVT dimensionality reduction algorithm is used to reduce the dimension of the characteristic signal,and then the on-lineAdaboost classification is carried out by affine projection and Mahalanobis distance,which improves the efficiency and accuracy of the attitude measurement of the gun under the limited welding parameters.(4)Based on the analysis of the dynamic characteristics of the wheeled moving trolley and the torch adjusting mechanism,the self-adjusting dynamic fuzzy rule and the hierarchical threshold coordination controller are adopted according to the deviation characteristic signals of the weld around the height,the height and the inclination angle obtained by the rotary scanning arc sensor to improve the coordination of the tracking system.In this paper,a new type of real-time dynamic space surface seam tracking technology is proposed.After a lot of tests,it is shown that the automatic welding error of the space surface weld is within the allowable range,and the expected accuracy and effect are achieved.