Dynamic analysis of HR-FN-HR neural network coupled by locally active hyperbolic memristors and encryption application based on knuth-Durstenfeld algorithm.

• A bistable locally active memristor model is designed, which has different hysteresis loops under different input signals. • An HR-FN-HR neural network is proposed, which consists of a 2D FN neuron and two 2D HR neurons coupled by memristors. • An encryption scheme based on the memristors-based neural network combined with Knuth-Durstenfeld algorithm is presented. A locally active hyperbolic memristor which can generate the bistability phenomenon is presented. Based on the memristor, a HR-FN-HR neural network coupled by hyperbolic memristors is composed. By theoretical calculation and mathematical simulation, the complex dynamical behavior of memristors-based neural network is discovered, and the equivalent circuit of neural network is realized. The Hamiltonian energy of the coupled neural network is calculated, which estimates the energy released during the transition between various electrical activities. In addition, a new image encryption scheme is presented, which is the Knuth-Durstenfeld algorithm combined with DNA sequence operation. The encryption results show that the scheme has better encryption effect and stronger anti-hacking ability, which can be used in the domain of secure communication. [ABSTRACT FROM AUTHOR]