This Is AuburnElectronic Theses and Dissertations

A CMOS Implementation of a Biological Neuron with Memristor as Synapse

Date

2017-04-26

Author

Li, Zhao

Type of Degree

Master's Thesis

Department

Electrical and Computer Engineering

Restriction Status

EMBARGOED

Restriction Type

Auburn University Users

Date Available

04-27-2022

Abstract

A CMOS neuron with 130-nm SiGe technology is presented in this paper to simulate the biological activities of synapses and neurons in human brains. Biological neuron behaviors are introduced in the beginning then comes with its model. A device with name memristor is introduced to simulate synapse of the neuron. The circuit design can be separated into 3 stages: Dendrite input stage, LIF (leaky integrate and fire) stage, synapse stage. The first stage is achieved by a V-I converter and an Axon-Hillock circuit is implemented to simulate LIF stage, the synapse stage is modeled by resistor-memristor voltage divider based on STDP rule. The features of the whole neuron are based on the studies from neuroscience and mimicked by utilizing digital and analog circuits. The purpose of this paper is to give building blocks for future neural network applications.