SJTU Intelligentized Robotic Welding Technology Laboratory

Founded in Mar. 2000, Intelligentized Robotic Welding Technology Laboratory is abundance in technologies by several years' research works on intelligentized welding robots and robotic welding intelligentized technologies, such as using artificial intelligent technologies to simulate the observation, judgement and operation of human welders; automatically recognizing the welding surrounding based on visual sensing technologies, guiding and tracking the weld beam and intelligently controlling the weld pool dynamics in a local surrounding. Dynamical process control of pulsed TIG welding in some typical process conditions, which is a difficult problem in welding society, is successfully tackled by using such intelligentized technologies as computer vision, knowledge-based modeling, fuzzy-artificial neural network and intelligent control. A series of research works are accomplished, such as single visual-based the welding surrounding recognition, 2D and 3D shape calculation of dynamical welding pool, fuzzy-artificial neural network controller of weld dynamical process, artificial neural network compensation and self learning control of time-delay uncertain objects and the knowledge-based modeling for welding process using fuzzy identification and rough set theory. The researchers also devoted themselves to the key technologies of locally autonomous intelligent welding robot systems. A locally autonomous intelligent welding robot (LAIWR) system is developed, which can sense the welding surrounding, autonomously recognize and guide the initial welding position, recognize the weld bead and autonomously plan, modify and track the trajectories, visual sense and extract the characteristics of weld pool dynamics, systematically guide by experts, intelligently control weld penetration and weld bead formation. It also has a unit which can simulate the basic intelligent operation of human welders in the working space. Many innovative accomplishments have been achieved in the visual sensing based welding robot’s key technology on autonomous and intelligent control and system realization. 

The discrete event Petri Net theory is used to design the multi-robots welding flexible manufacturing unit/system. The research on the Petri Net modeling and coordinated control of multi welding robots has been conducted and the timed color Petri net (TCPN) of welding flexible manufacturing unit with the sensing and information intercrossing, running time adjustment and intelligent decision of welding process of intelligentized welding robot is established. The fuzzy TCPN modeling information optimization algorithm (FPIOA) is designed according to the multi robots welding processing information flow. The process experiment of highly-difficult suspension fillet welding coordinative operation based on multi robots welding system Petri Net modeling of aluminium alloy AC pulsed GTAW is accomplished, which realizes the coordinative control of multi welding robots complex system and extend the application of Petri Net theory in the robot welding field.

In view of multi agents, the coordinative control of intelligentized multi robots welding flexible manufacturing unit and system is researched. A hybrid resource agent structure (HRAS) of multi robots welding flexible manufacturing system, which has a recognition structure and assists with a reflection structure, is established. A multi-robots welding flexible manufacturing system, which coordinate and cooperate to accomplish the welding task by several independent operate agents, is established. Utilizing the JADE multi agent system, a multi-agent control system, which is composed by spot welding robots, TIG welding robots, MAG welding robots, transport welding robots and multi freedom positioner, is established. This system has the preliminary research basis for the control system simulation and process experiment of the multi welding robot agents.  
  
Works on the above research fields have supported by more than 30 foundations, such as National Science Foundation of China (NSFC), National Defense Pre-Research Foundation of China, National Defense Foundation of China, and Science Foundation of Ministry of Education of China and Foundation of Shanghai Science and Technology committee. “Key technology of the local-surroundings autonomous intelligent welding robot (LAIWR)” is applied in the welding manufacturing of aerospace craft, ships and automotives, which promotes the automatic welding manufacturing technology in the equipment manufacturing industry of China. The mainly engaged persons have been rewarded by two 2nd grade awards of National Science and Technology Progress granted by State Department of P.R.C., one 1st grade award and one 2nd grade award of National Defense Science and Technology granted by National Defense Science and Technology Committee of P.R.C., one 1st grade award of Science and Technology of Chinese Mechanical Industry granted by Chinese Mechanical Industry Association and Chinese Mechanical Engineering Society, one 1st grade award of Science and Technology of Ministry of Education granted by Ministry of Education of P.R.C., one 2nd grade award of National Science and Technology Progress granted by State Department of P.R.C., and one 2nd grade award of Shanghai Science and Technology Progress granted by Science and Technology Committee of Shanghai Government of P.R.C. and so on. 10 national invention patents and 5 software copyrights are successfully applied for.