基于Kalman滤波的自回归滑动平均信号信息融合Wiener滤波器  

Kalman filtering-based information fusion Wiener filter of autoregressive moving average signals

作　　者：邓自立[1] 高媛[1] 

机构地区：[1]黑龙江大学自动化系,黑龙江哈尔滨150080

出　　处：《控制理论与应用》2005年第4期641-644,共4页Control Theory & Applications

基　　金：国家自然科学基金资助项目(60374026);黑龙江大学自动控制重点实验室资助项目(04-01).

摘　　要：应用Kalman滤波方法,在按矩阵加权线性最小方差最优信息融合规则下,提出了带白色观测噪声的多通道ARMA信号的多传感器信息融合Wiener滤波器.它可统一处理信息融合滤波、平滑和预报问题.为了计算最优加权阵,提出了计算局部滤波误差互协方差阵的公式.同单传感器情形相比,可提高估计精度.一个带三传感器的目标跟踪系统的仿真例子说明了其有效性.By using the Kalman filtering method and the linear minimum variance optimal fusion rule weighted by matrices, a multisensor information fusion Wiener filter is presented for the multichannel autoregressive moving average（ARMA） signals with white observation noise. It can handle the information fusion filtering, smoothing and prediction problems in a unified framework. In order to compute the optimal weighting matrices, the formula of computing the cross - covariance matrices among local filtering errors, is presented. Compared with the single sensor case, the estimation accuracy is improved. A simulation example for a target tracking system with three-sensor shows its effectiveness.

关 键 词：多通道ARMA信号 多传感器信息融合 按矩阵加权最优融合规则 WIENER滤波器 KALMAN滤波 方法 

分 类 号：TN713[电子电信—电路与系统]