摘要
针对带控制项的自回归模型,提出了一种适用于在线建模的,可以自动改变模型阶数的最小方差递推算法(VORLS),并给出一种新的阶数估算标准.该算法可以减轻计算负担,提高模型精度.基于VORLS和新的阶数估算标准设计了一种在线辨识器.应用结果表明该辨识器精度较高,相对误差小于2%.
Aimed at the automatic regression models with controlled input,a variable-order recursive least squares algorithm and a novel criterion of order estimation are derived in the paper,which can be used in on-line modeling.The algorithm can lessen calculative burden and improve the accuracy of the models.An on-line identifier is formulated on the basis of the methods.The results of the experiments indicate that the identifier is highly precise,its relative error is less than 2%.
出处
《大庆石油学院学报》
CAS
北大核心
2006年第6期83-85,111,共4页
Journal of Daqing Petroleum Institute
关键词
在线辨识器
参数估计
线性系统
变结构
on-line identifier
parameter estimation
linear system
variable structure
