摘要
该文在分析了伪线性扩展卡尔曼滤波中针对非线性观测方程线性化误差的产生原因以及对滤波影响的基础上,提出用n时刻的状态估计值代替一步预测状态值。由于n时刻的状态值中隐含着n时刻的测量值,因此比一步预测状态值更接近此时刻的真实状态矢量,从而将伪线性的观测方程在更接近真实状态值处泰勒展开将产生更小的线性化误差,通过这种处理从而提高了状态定位精度。最后通过计算机仿真证明了改进算法的有效性。
In the puppet linear Extended Kalman Filter process, based on the analysis of the effect on the linearization error about the measurement non-linear equation, this paper uses n time state estimation instead of first step principium forecast. Because the n time state estimation contains n time measure, it is close to the real state of this time so as to reduce the linearization error. The precision of state location is improved by means of this method. The simulation results verify the efficiency of the new method.
出处
《电路与系统学报》
CSCD
北大核心
2008年第3期140-144,共5页
Journal of Circuits and Systems
关键词
无源定位与跟踪
扩展卡尔曼滤波
非线性
passive location and tracking
Extended Kalman Filter （EKF）
non-linear
