摘要
与EKF类算法相比,基于不敏变换(UT)的不敏卡尔曼滤波(UKF)算法因为不存在线性化误差从而具有更好的性能,不过存在运算量较大的问题。建立了具有线性状态方程的固定单站无源目标跟踪模型,以此为研究背景在满足L+λa=nx+λ=常数的情况下由扩维UKF算法推导出了简化UKF算法并进行了仿真分析。仿真结果表明在满足推导条件时简化UKF算法可以在降低运算量的同时保持和扩维UKF同样的定位性能,具有较强的实用性。
The unscented transform （UT） based unscented Kalman filter （UKF） algorithm has better performance compared with the EKF algorithm because it doesn＇t exist linearization error while suffering a computation problem. Considering the single non-moving observer passive location and tracking system with a linear process equation as research ground, the simplified UKF algorithm was derived from the augmented UKF under the condition of L＋λa=nx＋λ=constant and the performance analysis was made. Simulation results indicate that the simplified UKF algorithm can keep the same performance as the augmented UKF algorithm while reducing the computation.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2012年第8期1655-1659,1664,共6页
Journal of System Simulation
基金
航空科学基金(20105584004)
