Dynamic behavior of memristor ML neurons and its application in secure communication

Abstract

Improving neurons in a real physiological environment and studying their electrical behavior is crucial for understanding human cognitive brain functions and neural dynamics. Neuronal cells reside in a complex physiological environment, where the electromagnetic fields generated by ion transmembrane movements affect their discharge activity. Therefore, to better simulate the real conditions of biological neurons, this paper incorporated the characteristics of the memristor and constructed a four-dimensional Morris-Lecar (ML) neuron model by adding a magneto-controlled memristor into the three-dimensional ML neuron model. Through the study of time series diagrams, phase plane diagrams, inter-spike interval (ISI) bifurcation diagrams, we explored the effects of the feedback gain coefficient and the relationship coefficient between membrane potential and magnetic flux on the firing behavior of neurons in the model. It was found that variations in these two parameters can lead to complex firing patterns in neurons. We also utilized the maximum Lyapunov exponent and dissipative theory to investigate the chaotic synchronization phenomenon in the memristor-based ML neuron model. Additionally, we explored the impact of noise on neuronal synchronization behavior within the system, finding that an appropriate noise intensity can effectively accelerate the neurons’ attainment of a synchronized state. Finally, applying the chaotic synchronization system to secure Communication, the simulation results and related analysis demonstrate that the system excels in encrypting and decrypting voice signals, offering high levels of security and confidentiality.

Graphical abstract

Simulation results of speech signal encryption and decryption