Discrete memristive chaotic neuron model with attractor growth and its applications

Abstract Due to their unique nonlinearity, memory capability, and plasticity, memristors have emerged as ideal devices for realizing synaptic coupling between neurons. However, the application of discrete memristor-coupled neuron models in image encryption has not yet received widespread attention. In this study, a novel one-dimensional neuron model is proposed, which is coupled with a memristor featuring sinusoidal memductance to construct a new two-dimensional discrete memristive chaotic neuron model (DMCN). Experimental results demonstrate that the mapping exhibits not only high complexity across various parameter planes but also intriguing phenomena such as attractor growth and enhanced sensitivity to initial conditions. Moreover, a hardware circuit based on a DSP platform was constructed, successfully capturing the hyperchaotic attractors of the DMCN mapping, thereby verifying its physical realizability. Finally, a novel encryption algorithm was designed based on the DMCN mapping, and comparative studies confirmed that it outperforms existing encryption schemes.