基于噪声随机模型的加权观测融合方法  

A New Approach for Weighted Measurement Fusion by Noises Random Model

作　　者：黄贤源[1] 隋立芬[1] 范玉茹[1] 

机构地区：[1]信息工程大学测绘学院,河南郑州450052

出　　处：《测绘科学技术学报》2009年第1期52-55,共4页Journal of Geomatics Science and Technology

基　　金：国家自然科学基金资助(40474007)

摘　　要：信息融合技术中,在各局部传感器的有色观测噪声为一阶AR模型的情况下,可以利用观测扩增方法消除有色噪声的影响,得到最优加权观测融合方程,从而实现状态的最优滤波解。对于有色观测噪声为MA或ARMA模型的情况,观测扩增方法不再适用。提出了基于有色观测噪声随机模型级数展开的方法,求解出各局部传感器有色观测噪声的方差,并利用该方差对加权观测融合滤波器进行了构造。通过计算实例证明,该方法不仅适用于观测噪声为AR模型,同时适用于噪声MA或ARMA模型。In the process of information fusion, when the colored observation noises are models, observation-expand filter can eliminate the influence of colored observation noises and the optimal weighted measurement fusion formula is obtained then the optimal results are accounted. When the colored observation noises are MA or ARMA model, obser- vation-expand filter cannot be used. A new approach was presented in the article by polynomial-quotient which trans- lated colored observation noises into infinite series, then calculated the variance of sensors dummy white observation noises and constructed the weighted measurement fusion filter. The simulation results showed that the method was not only fit for the observation noises model AR but also for MA or ARMA.

关 键 词：加权观测融合 有色观测噪声 现代时间序列分析方法 ARMA模型 观测扩增方法 

分 类 号：P207[天文地球—测绘科学与技术]