Observer-Based Synchronization of Complex Dynamical Networks Under Actuator Saturation and Probabilistic Faults

• Published in: IEEE transactions on systems, man, and cybernetics. Systems Vol. 49; no. 7; pp. 1516 - 1526
• Main Authors: Selvaraj, Palanisamy, Sakthivel, Rathinasamy, Ahn, Choon Ki
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.07.2019; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Actuator saturation; Actuators; complex dynamical networks (CDNs); Computer simulation; Controllers; Faults; Matrix methods; Observers; Optimization; Optimization techniques; Peer-to-peer computing; Probability theory; Saturation; state observer; Stochastic processes; stochastic reliable control; Symmetric matrices; Synchronism; Synchronization; Time lag
• ISSN: 2168-2216; 2168-2232
• DOI: 10.1109/TSMC.2018.2803261

This paper investigates the observer-based synchronization problem for a family of complex dynamical networks subject to time delay, external disturbance, randomly occurring actuator faults, and input saturation. A realistic actuator fault model is considered in which the actuator faults are represented by stochastic variables that are assumed to obey a certain probabilistic distribution. An <inline-formula> <tex-math notation="LaTeX">H_{\infty} </tex-math></inline-formula> performance-related criterion is obtained via the Lyapunov-Krasovskii functional approach and stochastic analysis technique to asymptotically minimize the synchronization error and observer error simultaneously. Moreover, to meet the requirement of actuator saturation, the conditions for the domain of the attraction region are determined by employing the linear matrix inequality (LMI) approach and an optimization technique. Specifically, the proposed controller for the network synchronization is very simple and easy to implement in practical systems. Furthermore, the gain values of controller and observer gains are calculated by solving a set of LMIs. Eventually, the proposed theoretical results are verified through numerical simulations.