Exponential [formula omitted] filtering for complex-valued uncertain discrete-time neural networks with time-varying delays.

The purpose of this paper is to design a compatible filter for a class of classical discrete-time neural networks (DTNNs) having uncertain complex-valued weighting parameters and time-varying delayed responses subject to the H ∞ performance measure. For this notion, the complex-valued filter scheme is designed for the proposed uncertain DTNNs with regard to the available output measurements. At first, some novel complex-valued weighted summation inequalities (WSIs) are put forth to establish a more precise linearized lower bound for the quadratic summing terms resulting from the forward difference of the assigned Lyapunov–Krasovskii functional (LKF). In what follows, an attempt has been made to propose the linear matrix inequality (LMI) based sufficient conditions for designing the robust H ∞ filter from the filtering error system attains exponential stability with the appropriate filtering gain matrices. Eventually, the theoretical conclusion is substantiated through a numerical example and the simulation outcomes reveal the applicability and efficiency of the proposed filter scheme. • The robust H ∞ exponential filter for delayed complex-valued DTNNs has been first designed. • From WSIs in Saravanakumar et al. (2019) and RCMI in Ganesan et al. (2020), we introduced the WSIs. • Our paper has sorted out the complications of the decomposition approach by WSIs. • The theoretical findings are facilitated through a numerical example with simulations. [ABSTRACT FROM AUTHOR]