SPICE model of HP-memristor using PWL window function for neuromorphic system design application

Abstract

The nanoscale memristor device SPICE modeling is significant for memristive circuits and neuromorphic system design applications. In this paper, a novel modified piece-wise linear (PWL) window function proposed for the nonlinear property of the memristor SPICE model. The PWL window function used here can address the shortcoming of the parabola and Heaviside- step function based typical window functions proposed in the literature, those are established by symmetrical functions. However, in most of the typical window functions, the monotonically increasing and the asymmetrical functions have not been addressed. Also, using this PWL-window function based memristor model, a binary-weighted memristor-based artificial synaptic circuit has been designed. The performance of the device model demonstrated by the efficiency of designed artificial synapse sensitivity to the external input signal.