贝叶斯理论在光栅动态误差预报中的应用

Bayes prediction for dynamic errors of grating system

针对实际测量过程中需要对数据量较少的非平稳性误差进行随机序列建模预报,提出了贝叶斯动态测量误差序列的预报方法。该方法不要求较大的数据量,对序列平稳性无限制,而且在预报过程中间隔插入标准量能有效提高预报精度。设计了一个光栅测长仪动态误差测量实验系统,利用贝叶斯预报方法对其动态误差数据进行建模预报,验证了贝叶斯预报方法的适用性和可行性。

This paper describes the Bayes prediction method for dynamic measurement errors. The method does not involve a lot of data and the data to be used have no limitation for stabilization. Furthermore, the method allows the use of standard errors to improve precision of prediction. The features a design calibration system of dynamic errors for grating-measurer and predication of the data of dynamic errors for grating-measurer, and the Bayes theory proves to be more suitable and reliable.