Master–slave synchronization of neural networks subject to mixed-type communication attacks

This paper concerns the master–slave synchronization issue of neural networks subject to mixed-type communication attacks. The synchronization strategy is based on static output feedback controller followed by an event-triggered scheme. The communication network is assumed to be under various types...

Full description

Saved in:
Bibliographic Details
Published inInformation Sciences Vol. 560; pp. 20 - 34
Main Authors Kazemy, Ali, Saravanakumar, Ramasamy, Lam, James
Format Journal Article
LanguageEnglish
Japanese
Published Elsevier Inc 01.06.2021
Elsevier BV
Subjects
Cyber-attack
Deception attack
DoS attack
Neural network
Replay attack
Synchronization
Neural network
Replay attack
Synchronization
Deception attack
Cyber-attack
DoS attack
Online AccessGet full text

Cover

More Information
Summary:This paper concerns the master–slave synchronization issue of neural networks subject to mixed-type communication attacks. The synchronization strategy is based on static output feedback controller followed by an event-triggered scheme. The communication network is assumed to be under various types of cyber-attacks, namely, deception, replay, and denial-of-service attacks. All these attacks are investigated in a unified Markovian jump framework. Using the Lyapunov–Krasovskii theory and stochastic analysis techniques, some design criteria are derived and formulated in terms of matrix inequalities. A convex optimization algorithm is proposed to design the static output feedback controller. Finally, two chaotic examples are presented to demonstrate the effectiveness of the event-triggered static output feedback controller.
ISSN:0020-0255
1872-6291
DOI:10.1016/j.ins.2021.01.063