Finite-time event-triggered non-fragile state estimation for discrete-time delayed neural networks with randomly occurring sensor nonlinearity and energy constraints

• Published in: Neurocomputing (Amsterdam) Vol. 384; pp. 115 - 129
• Main Authors: Wang, Yamin, Arumugam, Arunkumar, Liu, Yurong, Alsaadi, Fuad E.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 07.04.2020
• Subjects: Delayed neural networks; Energy constraint; Event-triggered scheme; Non-fragile state estimation; Sensor nonlinearity; Sensor nonlinearity; Energy constraint; Non-fragile state estimation; Event-triggered scheme; Delayed neural networks
• ISSN: 0925-2312; 1872-8286
• DOI: 10.1016/j.neucom.2019.12.038

In this paper, the problem of event-triggered non-fragile state estimator design for discrete-time delayed neural networks (DNNs) is investigated over finite-time span. In consideration of the changes of environment and high sensitivity, external disturbances and/or parameter uncertainties might be involved in estimator parameters of the concerned DNNs. Therefore, it is one of our main objectives to design a non-fragile state estimator subject to the norm bounded gain variation. The sensor nonlinearity is supposed to occur in a random way. In the meanwhile, the event-triggered scheme and energy constraints are adopted in state estimator design for the purpose of energy and resource saving. By using the Lyapunov stability theory and some analytical techniques, sufficient conditions are established to guarantee that the estimation error system is finite-time bounded and meet a prescribed mixed H∞ and passivity performance constraint. Furthermore, the estimator gains are obtained via solving a set of linear matrix inequalities (LMIs). Finally, two numerical examples are exploited to demonstrate the effectiveness of the developed technique.