Curve fitting optimization based on a mixed model for assembly applications

Gao Xiang,Wang Hua,Chen Guanlong
DOI: https://doi.org/10.1007/s10845-013-0843-x
IF: 8.3
2013-01-01
Journal of Intelligent Manufacturing
Abstract:During the final assembly process, finding a best fitting resolution for gap and flush between assembled parts is the main object. This paper divides these fitting problems into two optimization stages, which are attitude fitting and position fitting, and proposes a mixed model which combines the Gaussian Imaging method (Enriched Complex Extended Gaussian Imaging) and even fitting algorithm (Hausdorff distance based criteria). It is illustrated that Gaussian Imaging method can be used to orient two similar curve objects and therefore accelerate the optimization process by the even fitting method. We also demonstrate the model by fitting a 2D classic case, an automobile door fitting into a frame, and generalize the application of the model into 3D curve fitting by fitting a simple automobile lamp into an auto-body. The best rotation and translation is found by using two curve objects’ similarity in attitude and minimum width in gap and flush as criteria.
What problem does this paper attempt to address?