研究了一类带有参数不确定性的非线性随机时滞系统的鲁棒H∞滤波问题。假设参数不确定性是范数有界的并且系统的动态方程是由伊藤微分方程所描述的。对所有容许的参数不确定性以及外界干扰,构造一个线性、无时滞、不确定性及独立的状态滤波器,使滤波误差动态系统是指数均值稳定并且独立于时滞的。针对单时滞系统和多时滞系统两种情况,基于线性矩阵不等式(LMI)方法给出了保证鲁棒H∞滤波存在的充分性条件。最后,数值仿真结果很好地说明了该方法的有效性。
A robust H∞ filtering problem for a class of nonlinear stochastic systems with time-delay and parameter uncertainty is presented. Assume the parameter uncertainty is norm-bounded and the system dynamic is modeled by Ito-type stochastic differential equations. The aim of this work is to design a linear, delayless, uncertainties independent state estimator such that for all admissible uncertainties as well as exogenous disturbances, the dynamics of the estimation error is stochastically exponential stable in mean square and independent of the time-delay. For systems with single delay and multiple delay case, the sufficient conditions are proposed respectively to guarantee the existence of desired robust H∞ filters via linear matrix inequalities(LMI). A numerical example is presented to demonstrate the effectiveness of the proposed approach.