Y-可观广义系统降阶稳态Kalman融合器  

Reduced-order Steady-state Kalman Fusers for Y-observable Descriptor Systems

作　　者：邓自立[1] 李怀敏[1] 和丽清[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》2006年第11期1457-1461,1466,共6页Science Technology and Engineering

基　　金：黑龙江大学自动控制重点实验室基金资助

摘　　要：对于带多传感器的Y-可观广义线性离散随机系统,通过状态线性变换,将其化为两个降阶的非广义多传感器子系统。应用Kalman滤波方法和白噪声估值器,提出了子系统和原系统的局部状态估值器及它们的误差互协方差公式。在线性最小方差按矩阵加权,按对角阵加权和按标量加权最优信息融合准则下,提出了原系统状态的三种稳态广义Kalman融合器,可统一处理融合滤波、平滑和预报问题,且可改善局部估计精度。For the Y-observable linear discrete-time stochastic descriptor system with multisensor, by the linear transformation of state, it can be transformed into two reduced-order non-descriptor multisensor subsystems. Using the Kalman filtering method and white noise estimators, the local state estimators and their error crosscovariance formulas are presented for subsystems and original system. Under the linear minimum variance optimal fusion criterions weighted by matrices, diagonal matrices, and scalars, the three steady-state descriptor Kalman fusers are presenced for the original system state. They can handle the fused filtering, smoothing, and prediction problems in a unified frameworke, and they can improve the accuracy of local estimators.

关 键 词：Y-可观广义系统 多传感器信息融合 加权状态融合 降阶状态估值器 Kalman融合器 Kalman滤波方法 

分 类 号：O211.64[理学—概率论与数理统计]