不确定时滞神经网络鲁棒采样同步控制研究  

Robust Synchronization of Uncertain Delayed Neural Networks Using Sampled-Data Control

作　　者：张赣雷 王春柱 谢立典 王嘉伟 葛超 ZHANG Gan-lei;WANG Chun-zhu;XIE Li-dian;WANG Jia-wei;GE Chao(College of Artificial Intelligence,North China University of Science and Technology,Tangshan 063200;The First Team of Trainees of Comprehensive Training Base,The PLA Rocket Force,Tangshan 064000;Shijiazhuang Haishan Industrial Development Corporation,Shijiazhuang 050200)

机构地区：[1]华北理工大学人工智能学院,唐山063200 [2]火箭军综合训练基地学兵训练一队,唐山064000 [3]石家庄海山实业发展总公司,石家庄050200

出　　处：《现代计算机》2021年第4期3-7,17,共6页Modern Computer

基　　金：国家自然基金资助项目(No.61503120)。

摘　　要：研究带有不确定性和干扰的时滞神经网络的鲁棒同步问题。通过构建带有离散时延和分布式时延的新颖Lyapu⁃nov-Krasovskii泛函(LKF),使用新的积分不等式,获得同步误差系统(SES)的鲁棒稳定准则和镇定条件。通过求解一组线性矩阵不等式(LMIs)来获得采样控制器,所得结果具有更低的保守性,最后通过数值仿真验证方法的有效性。The robust synchronization problems of uncertain delayed neural networks with external disturbance have been investigated.By constructing neoteric delay-dependent Lyapunov-Krasovskii Functional(LKF)with discrete and distributed delays and using improved integral inequality,the robust stability criteria and stabilization conditions of Synchronization Error System(SES)have been obtained.The corresponding sampled-data controller can be acquired in terms of a host of Linear Matrix Inequalities(LMIs).The obtained results are less conservative compared to existing works.The numerical example is used to illustrated the validity of the theoretical results with corresponding simulation experiments.

关 键 词：不确定性 干扰 时滞神经网络 鲁棒同步 采样控制 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]