MEMRISTOR SPICE MODEL COMPARISON FOR NEUROMORPHIC COMPUTING

Abstract

Memristors, also known as memory resistors, are a novel kind of electrical component that set them apart from more conventional passive circuit components. This abstract provides a concise overview of memristor technology, covering their fundamental properties, diverse applications, and the promising future they hold in the realm of electronics. Memristors exhibit non-volatile memory behavior, retaining resistance states even when power is turned off. Their V-I (voltage-current) characteristics display non-linear, hysteresis-driven responses, allowing for precise control and manipulation of resistance. Because of this, memristors are appropriate for a variety of uses, including neuromorphic computing and non-volatile memory storage. The essential properties of memristors, including as their resistance-switching ability, bistability, and dual-mode operation. They are use in developing fields like brain-inspired computing and non volatile memory. Because memristors provide faster, more energy-efficient solutions, they have the potential to completely transform memory technologies. Memristor device modeling is required for memristor-based circuit and system design. This study reviews the state-of-the-art memristor modeling methods and offers simulations that compare a number of models with memristor characterization data that has been published. The simulations have been completed in LTspice, to compare the output of the various models to the relationships between current and voltage in real devices. Throughout the simulations, sine and triangle pulse inputs were used to assess each model's capabilities.