Hierarchical constraint satisfaction for high-level dimensional inspection planning

Coordinate measuring machines (CMMs) are very precise Cartesian robots that are used for dimensional inspection. High-level inspection planning for a CMM involves spatial reasoning, to determine how to orient the part on the CMM, which probes to use, how to orient the probes, and what measurements to perform. Current planners are incomplete or only solve the problem partially. In this work, we map the inspection planning problem to a hierarchical constraint satisfaction problem (CSP). The solutions to the CSP are inspection plans of good quality. We show how to extract approximate solutions using efficient clustering methods, which do not entail search and backtracking as prevalent in other planners. We describe our implemented planner and experimental results.