Memristor models for synapse component

Abstract

The memristor, the fourth passive circuit element, has exhibited resistance switching mechanism, which can be used in several applications such as nonvolatile memory, digital logic circuits, and neuromorphic systems. The switching mechanism in a Ta <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> O <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">5</inf> -RRAM device is achieved by conductive filament (CF) modulation that provides a suitable analog switching for the electronic synapses. In this paper, we analyze and discuss four different memristor models to identify which of them can achieve sufficient accuracy compared to the physical Ta <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> O <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">5</inf> -RRAM device, in order to be implemented as a synapse. These examined models are the linear ion drift (HP) model, the Voltage Threshold Adaptive Memristor (VTEAM) model, the Memdiode model and the Enhanced Generalized Memristor (EGM) model. Thus, we present the simulation results of each model and we compare its switching characteristics with the experimental characteristics. This study allows us to select the most appropriate memristor model for emulating the synaptic functions.