摘要
在具有控制输入和动态噪声与观测噪声相关的情况下,给出线性随机系统的集值滤波方程;利用矩阵分解和系统变换的技巧,得到广义随机系统的集值滤波方程.这种状态估计方法适用于初始状态均值位于一个凸集之中的随机系统.与传统K a lm an滤波产生单个条件分布不同,这里的集值滤波给出一个条件分布的凸集.
In case of the existance of control inputs or correlation between the inputs noise and the measurement noise, the set-valued filtering equation for stochastic linear systems is given. By using the techniques of the matrix decomposition and the system transformation, the set-valued filtering for stochastic singular linear systems is also designed. This method is fit to estimate the state of a stochastic system that the initial state is a random vector with mean value lying in a convex set. Rather than propagating a single conditional distribution as does conventional Kalman filtering, a convex set of conditional distributions is provided by the set-valued filtering.
出处
《控制与决策》
EI
CSCD
北大核心
2005年第12期1438-1440,共3页
Control and Decision
基金
国家自然科学基金项目(60574077)
江苏省教育厅自然科学基金项目(04KJD110168)
苏州科技学院科研基金重点项目
关键词
随机系统
广义系统
集值滤波
Stochastic system
Singular system
Set-valued filtering
