量测噪声自动加权Kalman滤波  

Measurement Noise Automatic Weight Kalman Filter

作　　者：沈云锋[1] 朱海[1] 等 

机构地区：[1]海军潜艇学院电航室，山东青岛266071

出　　处：《中国惯性技术学报》2001年第4期21-27,共7页Journal of Chinese Inertial Technology

摘　　要：从Kalman滤波技术的稳定性出发,分析了Kalman滤波算法的实质及容易发散的原因。提出在Kalman滤波中引入系统量测噪声协方差阵(R)的计算,并对其加权,从而影响滤波增益,抑制发散。推算舰位/GPS 组合导航的应用仿真表明:量测噪声自动加权Kalman滤波算法对系统模型误差和量测噪声协方差误差具有良好的自适应性。From the stability of Kalman filter, the essence of Kalman filter and the reasons of the instability of Kalman filter are analyzed. In this paper, the calculation of measurement noise covariance(R) is added to the normal Kalman filter. R is weighted and then influences the filter plus to overcome the instability of filter. The simulation of reckoning/GPS integrated navigation system shows that the measurement noise automatic weight algorithm is adaptive to the error of model and measurement noise covariance with high effectiveness and accuracy.

关 键 词：KALMAN滤波 量测噪声 组合导航 自动加权 协方差阵 模型误差 

分 类 号：TN966[电子电信—信号与信息处理] U666.12[电子电信—信息与通信工程]