A planar minimax alogrithm for analysis of coordinate measurements

We characterize the solution of the following problem and describe an algorithm for numerically solving it. Two sets ofN points in the plane, labeled 1,...,N, are given: a fixed set of nominal points and a set of measured points. We wish to transform the messured points as a whole, by translation and rotation, so that the maximal distance between corresponding points in the two sets is minimized. This algorthm provides an accept-reject criterion that may be used together with a coordinate measuring machine to determine if two mating parts will fit, or if a part is sufficiently close to its ideal measurements. A weighted version, suitable for point-dependent tolerances, is also discussed, as is optimal joint scalling of the data.