Attracting Similar Shapes Towards Each Other

This paper draws an analogy between the way insects use vision to move themselves with respect to local landmarks, and the problem of moving objects relative to each other in vision-guided robotic assembly. In particular, an algorithm is presented for attracting similar shapes together which was directly inspired by a model of navigation in honeybees, and which shares the same characteristics of robustness and immunity to noise. In particular, the algorithm can rotate, translate and scale one 2D shape to align it with another despite the presence of significant distortion and missing or extraneous features.