The number of seam trackers was developed for welding automation by now. Among these seam trackers, the arc sensor is prevalently used in industrial robot welding system because of its low cost and flexibility. However, the accuracy of arc sensor for measuring current would be decreased due to electrical noise and metal transfer mode.
The arc welding processes are substantially nonlinear, in addition to being highly coupled multivariable systems. Frequently, not all the variables affecting welding quality are known, nor may they be easily quantified.
In this study, two types of neural network based on the backpropagation algorithm were implemented. One of the neural networks was adopted for improvement of noise problem, and the other for selection of welding parameters, that is, current, voltage and welding speed. Handshaking between robot and computer was also developed for transmitting these selected parameters from computer to robot controller.
