Search

L2−L∞ Filtering for Constant Time-Delay Neural Networks with Quantized Output

Author(s)

Jing Han, Zi Wang, Ruoxian Wang

Affiliation(s)

School of Electronical Information Engineering, Wanjiang University of Technology, Ma’anshan, Anhui, China

Abstract

This article mainly focuses on the L2−L∞ filtering problem of a class of constant delay neural networks with quantized outputs. The main purpose of this article is to design a quantized L2−L∞ filter that makes the filtering error system asymptotically stable in the absence of external disturbances, while achieving the required L2−L∞ interference suppression index with zero initial conditions. To this end, by constructing an appropriate Lyapunov–Krasovskii functional that fully captures the delay information and quantization effects, and by employing several advanced inequality scaling techniques, a set of sufficient conditions is derived to ensure the desired L2−L∞ performance of the filtering error system. Subsequently, using systematic decoupling methods, an explicit linear matrix inequality (LMI)-based design procedure is presented for the required L2−L∞ filter under constant time delays. The filter gain matrix can be obtained efficiently by solving the derived LMIs with standard numerical solvers. Numerical examples ultimately demonstrate that the proposed filter design method is not only effective, but also has lower conservatism. In summary, this work provides a systematic and computationally tractable approach for designing quantized L2−L∞ filters for constant time-delay neural networks.

Keywords

Neural Network Model; Steady Time Delay; Quantitative Output; L2−L∞ Performan

References

[1]Zhou, J., Wang, Z., Yuan, D., & Shen, H. l2-l {infinity] Filtering for Discrete Time-Delay Markovian Jump Neural Networks. Przegląd Elektrotechniczny, 2012, 88: 96-100., 2013, 89(11), 114-118. [2]Choi, H. D., Ahn, C. K., Shi, P., Lim, M. T., & Song, M. K. L2−L∞ Filtering for Takagi–Sugeno fuzzy neural networks based on Wirtinger-type inequalities. Neurocomputing, 2015, 153: 117-125. [3]Yang, J., Ma, L., Chen, Y., & Yi, X. L2−L∞ state estimation for continuous stochastic delayed neural networks via memory event-triggering strategy. International Journal of Systems Science, 2022, 53(13): 2742-2757. [4]Yu B, Shi Y, Huang H. l2−l∞ Filtering for multirate systems based on lifted models. Circuits, Systems & Signal Processing, 2008, 27(7)-(8): 699-711. [5]Han J, Zhang Z, Zhang X, et al. Design of passive filters for time-delay neural networks with quantized output. Chinese Physics B, 2020, 29(11): 110201. [6]Chen Q, Liu X, Wang F. Improved results on L 2-L∞ state estimation for neural networks with time-varying delay. Circuits, Systems, and Signal Processing, 2022, 41(1): 122-146. [7]Seuret A, Gouaisbaut F. Hierarchy of LMI conditions for the stability analysis of time-delay systems. Systems & Control Letters, 2015, 81: 1-7. [8]Zhou K, Khargonekar P P. Robust stabilization of linear systems with norm-bounded time-varying uncertainty. Systems & Control Letters, 1988, 10(1)-(2): 17-20. [9]Boyd S, Ghaoui L E, Feron E and Balakrishnan V. Linear Matrix Inequalities in System and Control Theory. Society for Industrial and Applied Mathematics, 1994. [10]Fu M, Xie L. The sector bound approach to quantized feedback control. IEEE Transactions on Automatic control, 2005, 50(11): 1698-1711. [11]Han, J., Xia, Y., Zhang, L., Zhang, T., & Gao, H. H∞ filtering of a constant time delay neural networks with quantized output//2023 3rd International Conference on Intelligent Power and Systems (ICIPS). IEEE, 2023: 598-602.