Generalized synchronization theorems for a kind of Neural Network with application in data encryption

Two constructive generalized synchronization (GS) theorems for a kind of neural network are introduced, which are described by discrete-time array equation systems (DTAEs). Based on the theorems, one can design a GS driven DTAE via a driving chaotic DTAE and an inverse function of H. As an application, a generalized Henon cellular neural network (CNN) with three state variables is introduced. Using the GS theorems and the generalized Henon CNN constructs a coupled GS DTAE with 2646 cells. The hyper chaotic GS phenomena of the GS DTAE have been simulated. The numerical simulation results display complex behaviors of the GS DTAE. Using the DTAE designs a encryption scheme with ldquoone-time padrdquo function. This scheme is able successfully to encrypt and decrypt original information without any loss. The scheme is sensitive to the perturbations of the initial conditions and some system parameters of the DTAE. The key space is huge.