Neuromorphic behaviors in a neuron circuit based on current-controlled Chua Corsage Memristor

Abstract

Locally active memristor is considered as an ideal device for building neuron circuits. In this study, a novel current-controlled Chua Corsage Memristor (CCM), which supplements the existing CCM family, is proposed to explore its unknown neuromorphic dynamics. The non-volatility of the current-controlled CCM is verified by its power-off plot and the locally active domain is identified by its DC I-V plot. A third-order neuron circuit is developed by embedding the current-controlled CCM into a passive LC network. The edge of chaos domain of the neuron circuit is identified only by real parts of the eigenvalues of the system's Jacobi matrix. The resulting circuit is capable of sensing external current stimuli and producing neuromorphic behaviors when it is poised on or near the edge of chaos, making it more bionic. The neuromorphic behaviors in four different parameter intervals are revealed in detail, and the generation mechanisms are analyzed based on the theories of Hopf bifurcation and edge of chaos. Due to its simple structure and complex neuromorphic behaviors, it is expected that the neuron circuit may help to further study memristor-based neuron models.