New H∞ controller for neural networks subject to time-varying delay by state estimation based approach

• Published in: Neurocomputing (Amsterdam) 2025-03, Vol.621, p.129300, Article 129300
• Main Authors: Liu, Haibo, Qian, Wei, Zhao, Yunji
• Format: Article
• Language: English
• Subjects: Delayed neural networks; H∞ control; Linear matrix inequality; State estimation
• ISSN: 0925-2312
• DOI: 10.1016/j.neucom.2024.129300

This study focuses on the H∞ control issue for neural networks (NNs) subject to time-varying delay and external disturbances. The main objective is to construct a new H∞ controller to ensure the asymptotical stability of the closed-loop system in the case when the neuron states are not measurable. By establishing an appropriate Lyapunov–Krasovskii functional (LKF) with single and double integral terms and utilizing some effective integral inequalities, the state estimator is presented. Then, according to the estimated state, a more effective H∞ controller is further developed, and by utilizing a decoupling technique, some less conservative results are derived which guarantees that the concerned system is asymptotically stable and achieves given H∞ performance. At the end, the advantages and less conservativeness of the presented method is shown by one well-known simulation example.