摘要
利用卡尔曼滤波算法进行估计需要给定初始状态估计值和初始误差方差阵。通常，初始状态估计值和初始误差方差阵由经验给定，而初始状态估计值和初始误差方差阵的选取影响着卡尔曼滤波的估计精度。文中提出了加权最小二乘-卡尔曼滤波算法，并运用到惯导系统动基座初始对准中，进行了仿真。仿真结果表明，利用加权最小二乘算法可得到更加精确的卡尔曼滤波的初始状态估计值和初始误差方差阵，提高卡尔曼滤波的估计精度，进而提高了初始对准的精度。
Estimating using Kalman filtering algorithm requires predicted initial values of states and covariance matrix. Usually, they are chosen according to experience and influence the estimating precision of Kalman filtering algorithm. In this paper, a weighted least square Kalman filtering algorithm was developed and applied to initial alignment of inertial navigation system on dynamic base. The result of simulation shows that this algorithm can get more accurate initial values of states and covariance matrix that means better estimation of Kalman filtering, that is, the precision of initial alignment is improved
出处
《海军航空工程学院学报》
2008年第4期395-397,400,共4页
Journal of Naval Aeronautical and Astronautical University
关键词
加权最小二乘
卡尔曼滤波
动基座
weighted least square
Kalman filtering
dynamic base
