一类带时滞的静态神经网络的H∞滤波问题  

The H∞Filtering for a Class of Static Delayed Neural Networks

作　　者：刘博[1] 马秀莉 LIU Bo;MAXiuli(School of Mathematical Sciences,Shanxi University,Taiyuan 030006,China;Department of Mathematics,Jinzhong College,Jinzhong 030619,China)

机构地区：[1]山西大学数学科学学院,山西太原030006 [2]晋中学院数学系,山西晋中030619

出　　处：《山西大学学报（自然科学版）》2022年第6期1447-1454,共8页Journal of Shanxi University(Natural Science Edition)

基　　金：国家自然科学基金(61973201);山西省高等学校科技创新项目(2021L492)。

摘　　要：文章研究了一类带有时滞的静态神经网络的H∞滤波问题。通过引入一个增广的Lyapunov-Krasovskii泛函,并使用Bessel-Legendre不等式估计Lyapunov-Krasovskii泛函求导后的交叉项,给出了时滞神经网络稳定性分析的新的有界实引理。基于该有界实引理,得到了滤波误差系统的H∞性能分析,并通过求解线性矩阵不等式得到滤波器。This paper addresses the H∞filtering problem for a class of static delayed neural networks.By introducing an augmented Lyapunov-Krasovskii functional,and estimating the quadratic integral terms in the derivative of the Lyapunov-Krasovskii functional through the Bessel-Legendre inequality,a new bounded real lemma is developed for stability analysis of the delayed neural networks.Based on this,the H∞performance is analyzed for the filtering error system and the corresponding H∞filter is obtained through solving a set of linear matrix inequalities.

关 键 词：静态神经网络 时滞 H∞滤波 稳定性分析 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]