具有估计误差方差约束的连续时间卡尔曼滤波新方法

在状态估计应用领域。指标要求常常表现为估计误差稳态方差的上界形式。本文考虑连续系统在动态特性及稳态误差方差约束下的卡尔曼滤波问题，即设计滤波增益，使滤波矩阵具有给定的稳定裕度，同时各状态分量的估计误差稳态方差不大于各自预先给定值。基于代表方法，文中给出了期望滤波增益的存在条件及其解析表达式，并提供了说明性的数值算例。

A New Approach to Continuous-time Kalman Filter with Estimation Error Variance Constraints

In the area of state estimation,many performance requirements are naturally described in term ofthe upper bounds on the steady-state variance ofestimation error.This paper studies the problem of Kalman filter design for continuous-time systems with dynamic behaviour and steady-state error vari-ance constraints.The purpose of the addressed problem is to design filter gain such that the filter matrix has the specified stability margin. and the steady-State variance of estimation error for each state is not greater than the individual spocified value. Based on the algebraic approach,the conditions for the exist-ence and the analytical expression of desired filter gains are given.A numerical example is applied to dem-onstrate the effectiveness of the proposed design method.