Design of an all-position intelligent welding robot system

摘要

In order to reduce human intervention in the welding process and improve the intellectualization of the welding robot, a kind of all-position intelligent welding robot is designed in this paper, which improves the dynamic performance of the robot through the optimization of mechanical structure and driving system, and establishes a multi-information fusion laser welding system. The data management expert system can process the time-sharing and distribution of information and realize the comprehensive control. The laser ranging sensor is used to track the weld seam in real-time. A robust control algorithm based on joint flexibility is proposed for the nonlinear factors of friction, hysteresis and clearance in the transmission link. The image of the special-shaped pipe is acquired by a high-speed camera, and the target welded parts are identified and located by the deep learning neural network, so as to adjust the welding parameters and support the control system to track the target.