Precise pose and assembly detection of generic tubular joints based on partial scan data

Intelligent and accurate determination of the position and orientation, or pose, of a workpiece which is manually placed is essential for automating fabrication tasks such as welding. In this paper, a novel algorithm based on minimizing the area of a boundary enclosing partial scan data points is proposed for detecting both the pose and assembly of tubular joints with the aid of reference ideal models. The proposed algorithm can also be applied to tubular joints with non-cylindrical cross sections. The fit-up information obtained can be used to determine whether realignment is required or combined with the pose information to re-plan paths for subsequent tasks. The focus of existing state-of-the-art is on objects with features, and the localization of featureless objects such as generic tubular joints using partial and sparse scan data remains a challenge. The proposed algorithm is applied to an actual robotic welding system to locate a tubular workpiece. Experiment results using the scan data as ground truth show that root mean square error is less than 1% of the pipe diameters, considering both brace and chord components with diameters greater than 200 mm.