Positioning error compensation for machining center by support vector regression

Abstract

To improve the positioning accuracy of a certain type of bridge gantry machining center, a positioning error compensation method based on support vector regression (SVR) is presented. After analyzing the influence of training data sets distribution, and pointing out that SVR may have a substandard performance when the training data are distributed sparsely at neighborhood of outliers or distributed in a small range, a training data sets constructing criterion is proposed. SVR is employed in position error compensation of a bridge gantry machining center to establish the model of the positioning error, actual positioning error compensation effects shows that the positioning accuracy and compensation operating efficiency are improved effectively.