The Design Of Welding Robot System And The Study On Its Intelligent Control

Industrial robot technology is one of the high-new products that fixed people's eyes on our times and a lot of countries regard the robot as the strategic objective of the high-tech field in the national development plan. Welding is one of important industrial field for robot to be applied; at present, adopting robot to weld is the modernized and intelligent main sign in welding automatic technology.This paper combining the development plan project" Research and industrialization of intelligent steel welding net equipment "of Qingdao scientific and technological office, introduced one high quality innovative and intelligent welding robot system(welding production line),which consulting the procedure of the import equipment. The paper also put forward the design idea and control scheme of the intelligent welding robot.The moving and single torch welding robot was designed in the welding robot system, it adopts resistance welding technique. The system adopts PC technique to form two level control scheme with web function which can be used anywhere without the limitation of the power supply and place. According to the idea of mechanical and electric integration, the system design does its best to make itself flexible and industrial. The all- purpose and industrial intelligent robot is to be designed for this purpose.The robot is a strong coupling, non-linear controlled target with multi-inputs and multi-outputs, therefore it is very difficult to set up accurate mathematics model. This paper combines ILC with neural network to put forward ILC intelligent control method based on NN model identification. The ILC (Iterative Learning Control) algorithm is used in controller, this kind of algorithm is carried out simply and needs not accurate dynamics model. The neural network has the ability to approaches the non-linear model, it can study and adapt to the unknown and uncertain systematic dynamic characteristic, which making the system has stronger adaptability and robust. So we can utilize historical date of the uncertain system to estimate the stable state characteristic of the system through the neural network. The neural network can...