基于nu-SVM的三坐标测量机空间相关动态误差建模

为了建立高精度的三坐标测量机动态误差修正模型,分析了三坐标测量机空间动态误差相关性的特点,介绍了所采用的建模方法支持向量机（nu-SVM）的基本原理和算法。利用双频激光干涉仪建立了MC850移动桥式三坐标测量机的单向动态误差分离装置,分离了不同空间位置下的单向动态误差。分别应用BP神经网络、nu-SVM建模方法和误差数据建立了测量机的空间三维动态误差模型,进行了验证和比较,nu-SVM的建模精度比BP神经网络高。

Dynamic Correlation Error Modelling of CMM Based on nu-SVM

In order to build high precision dynamic error correcting model,the spatial dynamic error correlation of CMM was analyzed.The principle and algorithm of the adopted new support vector method（nu-SVM） was proposed.The experimental facility was designed by using the laser interferometer,which was used to separate the dynamic error in single direction of the moving bridge CMM.The dynamic errors in the different spatial positions were separated,which were used to build the three dimension error models by adopting the BP neural network and nu-SVM.The dynamic error models were verified and compared,whose result indicates that the modelling precision of nu-SVM is higher than that of BP neural network.