Event-Triggered Control for Switched Systems With Recurrent Neural Networks Subject to Stochastic Cyber-Attacks

This paper concentrates on event-triggered control for switched systems with recurrent neural networks. The significance of this paper is, a method of designing observers combined with event-triggered controllers for switched systems with neural networks is developed. Different from conventional swi...

Full description

Saved in:
Bibliographic Details
Published in2021 33rd Chinese Control and Decision Conference (CCDC) pp. 6672 - 6677
Main Authors Geng, Honglin, Qi, Yiwen, Jiang, Weiyu, Yu, Wenke, Zhao, Xiujuan, Xing, Ning, Zhang, Simeng
Format Conference Proceeding
LanguageEnglish
Published IEEE 22.05.2021
Subjects
Estimation error
Event-triggered control
Neural networks
Recurrent neural networks
Stability
State observers
Stochastic cyber-attacks
Stochastic processes
Sufficient conditions
Switched systems
Switches
Online AccessGet full text

Cover

More Information
Summary:This paper concentrates on event-triggered control for switched systems with recurrent neural networks. The significance of this paper is, a method of designing observers combined with event-triggered controllers for switched systems with neural networks is developed. Different from conventional switched systems, for the issue that when system state contains unknown items, this study adopts neural networks to approach the items. Then, considering the neural networks time-delay, transmission delay and stochastic cyber-attacks, closed-loop switched augmented estimation error systems are established. By utilizing multiple Lyapunov functions, sufficient conditions that guarantee the mean-square exponential stability of closed-loop switched systems are given. Subsequently, the controller and observer gains and event-triggering parameters are obtained. At last, a numerical simulation is provided to demonstrate the effectiveness of proposed method.
ISSN:1948-9447
DOI:10.1109/CCDC52312.2021.9601824