多维和多重递推辅助变量辨识算法  

Multidimensional and Multiple Recursive Instrumental Variable Identification Algorithms

作　　者：徐慧勤[1] 邓自立[1] 张明波[1] 

机构地区：[1]黑龙江大学自动化系,哈尔滨150080

出　　处：《科学技术与工程》2010年第2期366-371,共6页Science Technology and Engineering

基　　金：国家自然科学基金(60874063)资助

摘　　要：对带白色观测噪声的多维自回归(AR)模型,应用相关方法,分别提出了多维和多重递推辅助变量(RIV)算法,给出了未知AR参数的强一致估计;推广了一维RIV算法。提出了模型噪声和观测噪声方差的信息融合估值器,它们具有强一致性。一个仿真例子说明其有效性。For the multidimensional autoregressive （AR） models, using the correlation method, the multidi- mensional and multiple recursive instrumental variable identification algorithms are presented, respectively, by which ,the strongly consistent estimation of the AR parameters is given. They include the one-dimension RIV algorithm as a special case. The information fusion estimators of the noise variances are also presented ,which have strong consistence.

关 键 词：多维AR模型 参数估计 递推RIV算法 信息融合噪声方差估值器 收敛性 强一致性 

分 类 号：O211.64[理学—概率论与数理统计]