基于递推并行Kalman时序模型的随机误差的分离方法  

Random error separation method based on time series model of recursive Kalman running in parallel

作　　者：关秋菊[1] 刘桂雄[1] 金军[1] 

机构地区：[1]华南理工大学机械工程学院,广州510640

出　　处：《现代制造工程》2008年第2期116-118,共3页Modern Manufacturing Engineering

摘　　要：通过分析动态数据的组成结构,针对动态数据中的随机性成分项,提出一种分离动态随机误差的新算法。该方法基于递推辅助变量法与Kalman滤波相结合的时序建模机理,通过具有强大数据处理功能的MATLAB语言编程,实现了动态数据中随机误差分离。通过对仿真实例计算,结果表明使用递推并行Kalman(RIV-Kalman)时序建模方法可分离99.53%的随机误差。Through analyzing the structure of dynamic data, a new algorithm is presented for separating dynamic random error from dynamic data. The method is based on time series modeling principle combining recursive instrumental variables and Kalman filter. The separation of the random error is completed using MATLAB. The performances of the algorithm is illustrated according the results obtained from a numerical simulation study,99.53% of the random errors are separated.

关 键 词：递推辅助变量法 KALMAN滤波 动态数据 随机误差 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]