Skill Assistance with Robot for Manual Welding


by Mustafa Suphi Erden

Marie Curie Intra-European Fellowship, Project No: 297857


Host Institute:

Learning Algorithms and Systems Laboratory (LASA), École Polytechnique Fédérale de Lausanne






















































Background Research

Here are some selected results of my background research for this project


Analysis of Kinematic Data of Manual Welding

(a) (b)

Figure 1:(a) Data colllection of manual welding with motion capture system. (b) The position variation of the torch-tip in the direction of welding line (x) after passing through a high-pass filter with a cutoff frequency of 0.1Hz. Pink (light): unskilled performance data; blue (dark): skilled performance data. The two black lines represent the positive and negative thresholds of x-variation. These thresholds are used to distinguish between skilled and unskilled type movements

Applying a high-pass filter with a cutoff frequency at 0.1Hz on the position signal of the torch allowed classifying across skilled and unskilled performances (see Figure above). The threshold-based classification was not influenced by other kinematic variables, such as the curvature of the welding path or the average speed of welding.

Erden, M.S. and Tomiyama, T., (2008), “Identifying welding skills for training and assistance with a robot”. Science and Technology of Welding and Joining, 14(6): 523-532.

Assisting Manual Welding with Damping

 (a)  (b)

Figure 2:The HapticMaster used (a) for emulation of welding with air-brush-painting, (b) for actual welding with a TIG welding torch.

Manual welding is assisted with a haptic robot by applying damping at the end effector. This damping suppressed the high frequency vibrations of the human arm. Variable impedance was applied by observing the manipulation speed, in order to provide easy replacement of the welding torch in the phases that no welding was performed.

Erden, M.S. and Maric, B., 2011. “Assisting manual welding with robot”. Robotics and Computer Integrated Manufacturing, 27: 818-828.