摘要
在非线性、非高斯条件下进行动基座传递对准,如果采用卡尔曼滤波器误差会比较大而且可能会存在发散的问题,为了解决问题,引入了无迹卡尔曼滤波UKF(unscented Kalman filter)。使用确定性样本的方法来处理非线性的问题,使得采样点的均值和方差完全符合实际的非线性系统的均值和方差,解决了惯性导航系统动基座传递对准在正常工作时的基本条件。采用UKF和扩展卡尔曼滤波EKF(Extended Kalman Filter)的计算机仿真结果表明:UKF与EKF相比,精度提高了2倍,时间少了10秒。
Unscented Kalman Filter is generally introduced to solve the great error and even divergent problem caused by adopting Kalman Fihering in moving base transfer alignment under nonlinear and non - Gaussian condition. The paper uses a deterministic sampling approach to deal with the nonlinear systems. Thus the sample points completely capture the true mean and covariance of the nonlinear systems , and the postulate that can make moving base alignment in strapdown inertial navigation system run normally is solved. By computer simulation , the precision is multiplied 2 times with 10 seconds less comparing the UKF with the EKF.
出处
《计算机仿真》
CSCD
2008年第12期101-105,共5页
Computer Simulation
基金
国防科工委基金资助项目(J1300B004)
南京理工大学科研发展基金资助项目(XKF05031)
弹道国防科技重点实验室基金资助项目:51453060105ZS3301
关键词
捷联惯导
传递对准
动基座
无迹卡尔曼滤波
扩展卡尔曼滤波
Strapdown inertial navigaition systems (SINS)
Transfer alignment
Moving base
Unscented Kalman filter(UKF)
EKF