Abstract This chapter reports the multi-stable patterns coexisting in a Hopfield neural network (HNN) with three neurons under the action of electromagnetic induction current. A non-ideal memristor synapse is adopted to imitate the electromagnetic induction effect induced by the membrane potential difference between two adjacent neurons. Affected by the electromagnetic induction current, coexisting chaotic and periodic patterns with different topological structures or different locations are revealed under different memristor coupling strengths, synaptic connection weights, and initial conditions. Besides, the long-term transient chaotic behaviors are uncovered in the memristive HNN, which makes partial of the simulated attraction regions mixed and riddled under normal simulation interval. Finally, an analog realization circuit is designed for the memristive HNN and its PSIM circuit simulations verify the MATLAB numerical simulations.
Paper
Full text
Multi-stable patterns coexisting in memristor synapse-coupled Hopfield neural network
Semantic Scholar · Physics · 2021
Abstract
Abstract This chapter reports the multi-stable patterns coexisting in a Hopfield neural network (HNN) with three neurons under the action of electromagnetic induction current. A non-ideal memristor synapse is adopted to imitate the electromagnetic induction effect induced by the membrane potential difference between two adjacent neurons. Affected by the electromagnetic induction current, coexisting chaotic and periodic patterns with different topological structures or different locations are revealed under different memristor coupling strengths, synaptic connection weights, and initial conditions. Besides, the long-term transient chaotic behaviors are uncovered in the memristive HNN, which makes partial of the simulated attraction regions mixed and riddled under normal simulation interval. Finally, an analog realization circuit is designed for the memristive HNN and its PSIM circuit simulations verify the MATLAB numerical simulations.