Study On Initial Welding Point Positioning Technology Of Mobile Welding Robot Operating In Small Space

Mobile welding robot is primarily used to work in cramped cabins,which is depend on manual work when locating the initial position of welding.However,it is sometimes hard for workers to work in a narrow vertical cubicle.In addition,lifting the robot needs to overcome the gravity of the robot and the magnetic attraction force of its pedestal.Based on the welding environment and the mechanical characteristics of the robot,this article designed a control system that can help the robot find the initial welding position automatically in an unknown circumstance and its main introduction is focused on the implementation process of the control system.The article first introduces the composition of the robot system.The control system,developed by the PC104 bus,adopts certain boards such as motion control and data collection to achieve the function of control.Besides,monocular vision sensors and ultrasonic sensors are applied as the component to capture external environmental messages,installed in a specific location for synchronization with the mechanical control.The image processing part was developed by using the OpenCV computer open source visual library and upgraded by lens distortion correction with combination of the working characteristics of the welding robot.With the operating characteristics of the ultrasonic sensor,the program rotates the robot to scan the wall to obtain the effective position information,and use the least squares method to fit the point-in-time when it is perpendicular to the wall during the whole scanning process.When the time has been calculated,the robot rotates in reverse at a constant speed to adjust the angle.The mathematical model of robot margin adjustment is established in this paper,which can adjust the margins while keeping the angle of the robot and the target steel plate unchanged.At last,the center line of the laser line is extracted by the improved median filtering and the Hilditch thinning algorithm from the collected images.The position of the centerline vertices can be calculated by examining the images in sequence to get the image coordinates of the endpoint and acquiring the start and end positions of two line segments by fast Hough transform.The coordinatesof the feature points when the robot is in different poses are obtained by measurement,and the results are applied in the practical deviation adjustment control of the welding gun.The robot has been adjusted into the ideal welding position each time after many experiments,which shows the control system designed in this paper has good adaptability to the complex working environment in cabin.