摘要
针对一类主从神经网络系统,建立了包含时变时滞、马尔可夫跳变和Lévy噪声的动态模型。为了减少数据传输的消耗以及提高系统的效率,引入事件触发控制,设计反馈增益矩阵和事件触发反馈控制器,使得主从神经网络可以快速实现指数同步。通过李雅普诺夫稳定理论和不等式分析方法,得到了误差系统同步的充分条件。之后提出一个数值算例仿真,用以验证所得结论的有效性。通过研究神经网络系统的同步,设计出更高效的控制器来改善系统的动态响应特性,使系统在外部干扰存在的情况下也能保持良好的控制效果。
A dynamic model including time-varying delay,Markov jump and Lévy noise is established for a class of master-slave neural network systems.In order to reduce the consumption of data transmission and improve the efficiency of the system,event-triggered control is introduced.The feedback gain matrix and event triggered feedback controller are designed,so that the master-slave neural network can realize exponential synchronization quickly.By Lyapunov stability theorem and inequality analysis method,the sufficient conditions for synchronization of error system are obtained.Then,a numerical example is proposed to verify the effectiveness of conclusion.By studying the synchronization of neural network systems,more efficient controllers are designed to improve the dynamic response characteristics of the system,enabling it to maintain good control performance even in the presence of external interference.
作者
张仁磊
童东兵
陈巧玉
周武能
ZHANG Renlei;TONG Dongbing;CHEN Qiaoyu;ZHOU Wuneng(School of Electronic and Electrical Engineering,Shanghai University of Engineering and Science,Shanghai 201620,China;College of Information Sciences and Technology,Donghua University,Shanghai 200051,China)
出处
《控制工程》
CSCD
北大核心
2024年第12期2184-2189,共6页
Control Engineering of China
基金
国家自然科学基金资助项目(61673257)。
关键词
神经网络
指数同步
Lévy噪声
事件触发控制
Neural networks
exponential synchronization
Lévy noise
event-triggered control
