Adaptive Calibration for Articulated Arm Coordinate Measuring Machine

The calibration of the coordinate measuring machines is performed based on the mathematical model of those machine. The feature of the artifact is measured by CMM, the kinematic parameters are determined by comparing the calculated coordinates and the calibrated coordinates. However, it is assumed that the CMM is rigid and stable in any orientations and locations. The actual CMM is not sufficiently rigid and stable. Especially the articulated arm CMM is not. In order to consider the CMMs’ deformation and improve the accuracy of those, it is proposed that the model equation, the kinematic parameters and the estimators are replaced with the neural network. The calibration method using the neural network is called the adaptive calibration. The neural network is trained based on the relationship between the inputs and the coordinates successfully. In this article, the adaptive calibration is applied to the articulated arm coordinate measuring machine.