Mixed-Mode Artificial Neuron for CMOS Integration

Author

Zatorre, Guillermo ; Medrano, Nicolás ; Sanz, M. Teresa ; Martínez, Pedro A. ; Celma, Santiago

Author_Institution

Grupo de Diseno Electron., Universidad de Zaragoza

fYear

2006

fDate

16-19 May 2006

Firstpage

381

Lastpage

384

Abstract

A new approach to designing artificial neurons in CMOS technology is proposed in this paper. Design and simulation results of the basic building blocks are presented. Programmable weights are obtained using a current-mode mixed-signal four-quadrant multiplier, whereas the non-linear output function is implemented with a specifically designed class AB current conveyor. Starting from circuit simulation results, the behaviour of the proposed neuron was modeled. A multilayer perceptron network implemented with the new artificial neuron structure was trained to tackle linearization of a giant magneto-resistive sensor. Simulation results show the efficiency of the new implementation

Keywords

CMOS integrated circuits; analogue-digital conversion; current-mode circuits; magnetoresistive devices; mixed analogue-digital integrated circuits; multilayer perceptrons; multiplying circuits; sensors; CMOS integration; circuit simulation; class AB current conveyor; current-mode mixed-signal four-quadrant multiplier; giant magneto-resistive sensor; mixed-mode artificial neuron; multilayer perceptron network; nonlinear output function; programmable weights; Analog-digital conversion; Artificial neural networks; CMOS technology; Circuit simulation; Energy consumption; Inverters; Magnetic sensors; Multilayer perceptrons; Neurons; Voltage;

fLanguage

English

Publisher

ieee

Conference_Titel

Electrotechnical Conference, 2006. MELECON 2006. IEEE Mediterranean

Conference_Location

Malaga

Print_ISBN

1-4244-0087-2

Type

conf

DOI

10.1109/MELCON.2006.1653118

Filename

1653118