GENERALIZED REGRESSION NEURAL NETWORK BASED EFFICIENT MEMRISTOR MODELING

Abstract

With the recent advances in memristors as a potential building block for future hardware, it becomes an important and timely topic to study on memristor modelling. Memristor models are important for designers to exhibit memristor behavior since memristor is not yet available in market. An ideal memristor behavior has been remodel with Generalized Regression Neural Network (GRNN) and presented in this paper. Mathematical equations are used with a set of given memristor process parameters such as R-ON, R-OFF, thickness of TiO2, and instantaneous memristor behaviour is modelled. The behavior of this model is in agreement with the calculations of HP Lab's and Joglekar's SPICE model.