Memory-based event-triggered synchronization of dynamic memristor delayed cellular neural networks for image encryption

Abstract

This paper focuses on synchronization issue of dynamic memristor-delayed cellular neural networks (DM-DCNNs) for the first time. Different from the traditional memristor-based NNs (MNNs) that are modeled by discontinuous switched systems, DM-DCNNs where the memristor has flux-controlled and continuous-time nonlinear relation have been paid widespread attention. In order to reduce network burden, a novel memory-based event-triggered mechanism (METM) is proposed to synchronize drive–response systems of DM-DCNNs. With the help of some inequality techniques and Lyapunov stability theory, the sufficient conditions for synchronization are given by some linear matrix inequalities (LMIs). Unlike previous researches, all synchronization results were conducted in the flux-charge domain, which may be a potential advantage for information processing. Then, a numerical example is employed for supporting correctness of these synchronization criteria. Furthermore, the synchronization of DM-DCNNs under METM is further designed as a new type of encryption and decryption algorithms for image protection. Finally, the experimental performances are also provided to verify the high security of designed algorithm with anti-attack capability.