Mittag-Leffler synchronization of fractional order disturbed chaotic neural networks with varying time-delay using hybrid controller and its application to biometric image encryption

The sufficient conditions for the Mittag-Leffler (M-L) synchronization of fractional ordered chaotic neural networks, with varying time-delay, is established in this work. The situation where the system is affected by state-dependent exogenous disturbances is carefully considered and studied here. The synchronization is attained by introducing a hybrid controller, which involves a feedback controller and an event-triggered impulsive controller. The main results are proved by utilizing the concepts of Lyapunov theory and Linear Matrix Inequalities (LMIs). Further, the Zeno behavior is avoided by carefully designing the settings for the event-triggering conditions. The dynamic behavior of the networks, which are stabilized using the hybrid controller, during initial time, was effectively utilized in biometric image encryption application. In addition to that, Josephus scrambling technique was introduced to enhance the security. The qualitative and quantitative results proved that the method effectively retrieve the original image after decryption. Further, comparative results proving the efficiency of the proposed methods with the existing literature is also provided.