A Model-Based Framework for Optimal Measurements in Machine Tool Calibration

Calibration is the procedure of quantifying mechanical deficiencies of machines and compensating them by appropriate adjustment. This paper introduces a model-based measurement framework for improving calibration procedures of machine tools. The goal is to precisely estimate the mechanical deficiencies based on a minimal number of measurements. For that purpose, the mechanical deficiencies of linear and rotary joints are modeled using splines. Uncertainties of the deficiency model are formulated stochastically, which allows incorporation of imprecise measurement data and prediction of optimal measurement parameters. We derive a method for optimally estimating a set of splines, i.e., joint errors, based on a set of measurements and for predicting the optimal joint configuration for new measurements.