The Application of a Knowledge-Based Clustering Algorithm As an Aid to Probe Selection and Inspection Process Planning

Abstract

This paper outlines the development of a clustering algorithm used for inspection planning which allows each inspection feature to be inspected at a designated cell. This is achieved by grouping (a) inspection features into feature families and (b) probe orientations into probe cells. This would result in minimal probe calibration errors and part installation errors for the relative tolerance features. This procedure would reduce the time for probe exchange and reinstallation of parts. An incidence matrix representation has been developed to represent the relationship between inspection features and their relative probe orientations. The incidence matrix which is used for grouping feature families and probe cells are similar in function to the concept of group technology (GT) as used in machine cell formation. The knowledge-based clustering algorithm possesses the flexibility for consideration of multiple constraints for grouping probe cells and feature families. The application of the developed clustering algorithm satisfies the requirement of the inspection feature grouping and provides efficiency and effectiveness in probe selection and inspection process planning for coordinate measuring machines (CMMs).