基于UKF的GPS/DR组合导航非线性状态估计  

NON-LINEAR STATE ESTIMATION FOR GPS-DR INTEGRATED NAVIGATION BASED ON UKF

作　　者：朱虹[1] 关桂霞[1] 田健仲[1] 关永[1] 

机构地区：[1]首都师范大学信息工程学院,北京100037

出　　处：《计算机应用与软件》2008年第12期83-84,137,共3页Computer Applications and Software

基　　金：北京市自然科学基金项目(4082009);北京市高校中青年骨干教师基金项目(075313104)

摘　　要：GPS/DR组合导航系统是一个非线性系统,以往采用的扩展卡尔曼滤波算法EKF使得函数的整体特性被局部特性所代替,加上噪声的存在使系统的性能进一步下降。为了获得更好的状态估计性能,用无迹卡尔曼滤波算法UKF(Unscented Kalman Fil-ter)实现组合导航系统的非线性状态估计,避免了EKF方法的线性化近似过程,提高了算法的收敛速度和载体的定位精度。仿真结果表明:在非线性状态估计中,UKF滤波方法优于EKF滤波方法。The model of GPS-DR integrated navigation system is non-linear, the Extended Kalman Filter （EKF） used earlier made the integral property of the function be replaced by local property, and the performance of the system further degraded due to extra noise. The nonlinear state estimation should be realized by Unscented Kalman Filter （UKF） method in order to acquire better performance. It is because the approximate process caused by linearization in EKF method is avoided, the algorithm＇ s convergence speed and the load＇ s positioning precision of UKF is far quicker than that of EKF. The simulation result shows that UKF is superior to EKF in non-linear state estimation.

关 键 词：组合导航 EKF滤波 UKF滤波 定位精度 

分 类 号：TP271[自动化与计算机技术—检测技术与自动化装置] TN967.2[自动化与计算机技术—控制科学与工程]