Dynamics of stimuli-based fractional-order memristor-coupled tabu learning two-neuron model and its engineering applications

Abstract

External stimulus has an impact on the functional behavior of the biological nervous system, and appropriate stimulus helps the organism to maintain neural function. Inspired by this, the effects of different external stimuli on the dynamical behaviors of neuron model are studied in this paper. Firstly, a fractional-order (FO) memristor-coupled tabu learning two-neuron model whose equilibrium points are symmetric about the origin and unstable is constructed. The dynamical behaviors of this neuron model under different stimuli are further discussed, which are no external stimulus, external forced current stimulus, and electromagnetic radiation (EMR), respectively. The neuron model without external stimulus has periodic attractors and transient chaos, but applying external stimulus to one of the neurons can produce abundant two-scroll chaotic attractors and multistability; especially, when the neuron model is stimulated by EMR, it can generate hyperchaotic attractors that have not been observed in the tabu learning neuron model before. Besides, the transient transition behaviors of the model under different stimuli are also studied. Then, a pseudo-random number generator is designed and its random performance is tested with NIST suite. Finally, it is applied to voice encryption, and the result shows that it has good encryption effect. Therefore, it can be said that the FO memristor-coupled tabu learning two-neuron model has superior randomness and is suitable for chaotic-based engineering applications.