利用SRAR-MUSIC的动态系统误差分离方法研究  

System error separation mechanism based on SRAR-MUSIC

作　　者：关秋菊[1] 刘桂雄[2] 

机构地区：[1]仲恺农业工程学院机电工程学院,广州510225 [2]华南理工大学机械与汽车工程学院,广州510640

出　　处：《现代制造工程》2009年第11期47-50,共4页Modern Manufacturing Engineering

摘　　要：通过分析动态数据的组成结构,针对动态数据中的系统误差项,提出一种分离系统误差的新算法。该方法运用基于逐步舍选回归和求根MUSIC谱分析相结合的建模机理,实现了动态数据中系统误差的有效分离。仿真实例的系统误差分离结果显示,系统误差的估计值与模拟值相近,系统误差中非周期项的系数、周期项的频率和幅度估计值的相对误差分别为0.9%、5%和1.5%。Through analyzing the structure of dynamic data, a new algorithm is present for separating dynamic system error from original data. The method is based on root-MUSIC modeling principle combining Stepwise Rejection-Accept Regression（SRAR）. The performance of the algorithm is illustrated according the results obtained from a numerical simulation study. The estimated system errors are close to the simulated ones, the relative error of the estimated value of the coefficient of the trend component, the frequency and magnitude of harmonic component is 0.9% ,5% and 1.5% respectively.

关 键 词：动态数据 系统误差 逐步舍选回归分析 求根MUSIC谱分析 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]