Abstract
In coordinate metrology, a feature (Gaussian associated feature) is calculated from a measured data set of CMM (Coordinate Measuring Machine) using a least squares method. This data processing flow is called as "feature based metrology". In the feature based metrology, it is a key technique to estimate the uncertainty of measurement in the specific measuring strategy. The estimation method for uncertainties of measured parameters has been already proposed when the only random errors are put in the consideration. In this paper, the effects of systematic errors are theoretically analyzed to estimate the uncertainties in feature based metrology. The center position error and the diameter error of the ball probe are occurred from the random errors of probing in calibration process. These errors propagate as unknown systematic errors to the uncertainties of measured parameters such as the center position and the diameter of a measured circle. The method to calculate the error matrix was derived when the center position and the diameter of the circle are measured. Using this method, the uncertainties of the measured parameters can also be calculated in the complex measuring strategy. The series of simulations for this method in statistical way directly implies that the concept and the basic data processing method in this paper are useful to the feature based metrology.
