Global exponential synchronization of discrete-time high-order switched neural networks and its application to multi-channel audio encryption

Abstract

For a class of discrete-time high-order switched neural networks with time-varying delays, the global exponential synchronization (GES) problem and its application in audio encryption are addressed. In this work, the state feedback controllers are designed to ensure the GES of response and drive systems, and the controller gains are directly computed by the network parameters. Then, a numerical example is presented to explain the correctness of the synchronization theory. Moreover, a technique to audio encryption/decryption is proposed, and it is applied to a practical example to verify the applicability of the technique. The advantages of this paper are as follows: (i) GES criterion containing a group of simple matrix inequalities is established without the construction of a Lyapunov–Krasovskii functional (LKF), which is not always obvious; (ii) the controller gains are directly represented by the parameters in the network, and then it is easy to apply them to the audio encryption.