摘要
卡尔曼滤波器的稳定性并不能保证实际的滤波器算法具有收敛性,而传统的关于滤波发散的判定方法是基于量测不能出现野值这一严格的假设条件,具有一定局限性;并且通常关于滤波发散只是一种定性的描述,具有一定模糊性。为此,考虑到量测可能出现野值的情况,从滤波发散的具体数学形式出发,提出了卡尔曼滤波过程的稳定性概念,在此基础上给出了滤波发散的判定定理,既消除了通常关于滤波发散描述的模糊性,又降低了通常判定方法的要求。
The stability of Kalman filter can not assure the convergence of Kalman filtering algorithm. And the traditional methods for detecting divergence of the Kalman, filter also have some disadvantages, with a strict assumption that single measurement must be correct without measurement outlier. In addition, the traditional qualitative description of Kalman filter divergence is vague to some extent. Considering the probability of false measurements, a strict definition of the stability of filtering process by the mathematic limit form is first presented, according to the filtering effect. So the vagueness of the description of divergence is avoided. Finally, several detecting theorems are put forward, with fewer constraints than the traditional ways.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2005年第1期33-35,共3页
Systems Engineering and Electronics
关键词
卡尔曼滤波
发散
滤波过程稳定性
收敛
野值
Kalman filter
divergence
stability of filtering process
convergence
outlier
