离散奇异跳变系统的网络化H_∞鲁棒滤波

Networked Robust H_∞ Filtering for Discrete Singular Jump Systems

针对不可靠网络传输下的离散奇异马尔科夫跳变系统,提出了H_∞鲁棒滤波器设计,基于网络在模态跳变以概率方式接近观测下,建立了一个全新的具有部分模态依赖的网络化滤波器存在的充分条件,其滤波器参数可以通过求解线性矩阵不等式而易于得到,且不需要对原系统矩阵分解。在系统模态跳变并以概率方式接近观测下,滤波误差系统是正则、因果和随机稳定且满足H_∞性能指标,并给出数例证明了所提出的网络化鲁棒H_∞滤波器设计是有效的。

Aimed at the discrete singular Markovian jump systems transmitted through an unreliable network, the design of robust H∞ filtering is enhanced, and based on the probability observation over network, a sufficient condition is established for a novel partially mode dependent networked filter, whose filtering parameters are available by solving the linear matrix inequality, without decomposing the original one. The corresponding filter parameters can be obtained by solving a set of linear matrix inequalities （LMIs） without decomposing the original system matrix. Under the probability observation over network, the filtering error system is to be regular, causal, stochastically stable and satisfies H∞ performance, moreover numerical examples are given to illustrate the effectiveness of the networked H∞ filtering design.