ABSTRACT
We have demonstrated a fast method by which opti mal probe placements can be obtained for any known polygonal object. More importantly, we can guaran tee that the number of probes resulting from our tech nique is within a constant of the actual optimal num ber of probes necessary. These probing strategies re fine the position of an object whose pose is approx imately known. Furthermore, we claim that this is a real problem often encountered in industrial man ufacturing. Our paper also shows that the problem of optimal probe placement is dual to the well studied push-pull grasping problem of positioning frictionless fingers on an object.