Development Of SCARA Robot Oil Injection Control System Based On Machine Vision

With the development of sintering industry in China,China’s sintering equipment manufacturing has made significant progress.During the operation of sintering machine,the wheel bearing is suffering from high temperature and heavy load,so it is necessary to inject lubricating grease into the wheel to improve its service life.However,in the aspect of wheel oil injection of sintering machine,it is mainly manual oil injection at present.In the process of oiling,workers need to adjust the handle of the long rod type oiling head to the movement direction of the sintering machine at the designated position.The operation process is complex,and improper oiling will not only reduce the service life of the bearing,but also increase the sintering energy consumption.In this paper,it is of great significance to design a machine vision guided robot tracking and oiling system.Specific work is summarized in the following aspects:1.According to the site environment,the layout of the general scheme was planned,and the design of the general scheme was divided into three modules:cleaning,vision and oil injection.The scheme of the control system and the selection of the main hardware were put forward,and the software of the control system was introduced.2.Introduced the image preprocessing,recognition and location.A three-point calibration method was used to solve the position relation matrix between pixel coordinates and the world coordinate system,and the position relation among the sintering machine wheel,camera and robot coordinate system was calibrated.3.According to the structural characteristics of SCARA robot,the forward and inverse kinematics of SCARA robot were solved.The solution was applied to the workspace of SCARA robot,and the rationality of the trajectory design was verified.the cubic polynomial interpolation algorithm was used to plan the path back to the waiting point after tracking.4.The RBF neural network control and tracking problem of robot was studied.The structure of side mounted SCARA robot was simplified and its dynamic model was established.Due to the uncertainty of the system,a RBF neural network controller with simplified dynamic model was designed.Aiming at the problem that the initial weights of RBF neural network were irregular and had a great impact on the initial accuracy,this paper proposed an improved flying mouse optimization algorithm,optimized the initial weights of RBF neural network,and substituted the optimized initial weights into the RBF neural network controller,and verified that the controller meets the design requirements through MATLAB.5.The experimental platform were built,and the tracking detection and oil injection detection tests of different sintering machine wheels were completed,which verified the feasibility of side mounted SCARA robot tracking and oil injection in this paper.