Modified Training System for Manual Arc Welding by Using Mixed Reality and Investigation of Its Effectiveness

Abstract

While many of welding tasks are automatized nowadays, there is still much demand for skillful manual welding tasks. Therefore, the inheritance of manual welding skill and, safe and efficient training of unskilled welders are important. As a solution for them, a simulator by means of mixed reality technique and a training system with it have been proposed by now. In the simulator and training system, however, there were problems such as registration error of virtual objects, measurement error of gap between the electrode tip and the workpiece, slow refresh rate of image displayed to trainees, obscure teaching function of skill improvement, etc. In this research, first, the problems in the simulator have been solved by changing the camera lens, gap sensor setup and so on. As the result, reality of the simulator has been improved considerably. Secondly, in the teaching function, teaching items have been refined from the viewpoint of novice welder training. Applying the modified system to the training of ten unskilled subjects, it has been confirmed that the system is effective in the training.