A hardware implementation of artificial neural networks using field programmable gate arrays

Research output: Contribution to journalArticle

15 Citations (Scopus)

Abstract

An artificial neural network algorithm is implemented using a low-cost field programmable gate array hardware. One hidden layer is used in the feed-forward neural network structure in order to discriminate one class of patterns from the other class in real time. In this work, the training of the network is performed in the off-line computing environment and the results of the training are configured to the hardware in order to minimize the latency of the neural computation. With five 8-bit input patterns, six hidden nodes, and one 8-bit output, the implemented hardware neural network makes decisions on a set of input patterns in 11 clock cycles, or less than 200 ns with a 60 MHz clock. The result from the hardware neural computation is well predictable based on the off-line computation. This implementation may be used in level 1 hardware triggers in high energy physics experiments.

Original languageEnglish
Pages (from-to)816-820
Number of pages5
JournalNuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors and Associated Equipment
Volume581
Issue number3
DOIs
Publication statusPublished - 2007 Nov 1

Fingerprint

field-programmable gate arrays
Field programmable gate arrays (FPGA)
hardware
Neural networks
Hardware
clocks
Clocks
education
High energy physics
Feedforward neural networks
actuators
cycles
physics
output
Costs
Experiments

Keywords

  • Artificial neural network
  • FPGA
  • Level 1 trigger
  • VHDL

ASJC Scopus subject areas

  • Instrumentation
  • Nuclear and High Energy Physics

Cite this

@article{27d60453f8cd40efba3745cba45af58b,
title = "A hardware implementation of artificial neural networks using field programmable gate arrays",
abstract = "An artificial neural network algorithm is implemented using a low-cost field programmable gate array hardware. One hidden layer is used in the feed-forward neural network structure in order to discriminate one class of patterns from the other class in real time. In this work, the training of the network is performed in the off-line computing environment and the results of the training are configured to the hardware in order to minimize the latency of the neural computation. With five 8-bit input patterns, six hidden nodes, and one 8-bit output, the implemented hardware neural network makes decisions on a set of input patterns in 11 clock cycles, or less than 200 ns with a 60 MHz clock. The result from the hardware neural computation is well predictable based on the off-line computation. This implementation may be used in level 1 hardware triggers in high energy physics experiments.",
keywords = "Artificial neural network, FPGA, Level 1 trigger, VHDL",
author = "Eunil Won",
year = "2007",
month = "11",
day = "1",
doi = "10.1016/j.nima.2007.08.163",
language = "English",
volume = "581",
pages = "816--820",
journal = "Nuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors and Associated Equipment",
issn = "0168-9002",
publisher = "Elsevier",
number = "3",

}

TY - JOUR

T1 - A hardware implementation of artificial neural networks using field programmable gate arrays

AU - Won, Eunil

PY - 2007/11/1

Y1 - 2007/11/1

N2 - An artificial neural network algorithm is implemented using a low-cost field programmable gate array hardware. One hidden layer is used in the feed-forward neural network structure in order to discriminate one class of patterns from the other class in real time. In this work, the training of the network is performed in the off-line computing environment and the results of the training are configured to the hardware in order to minimize the latency of the neural computation. With five 8-bit input patterns, six hidden nodes, and one 8-bit output, the implemented hardware neural network makes decisions on a set of input patterns in 11 clock cycles, or less than 200 ns with a 60 MHz clock. The result from the hardware neural computation is well predictable based on the off-line computation. This implementation may be used in level 1 hardware triggers in high energy physics experiments.

AB - An artificial neural network algorithm is implemented using a low-cost field programmable gate array hardware. One hidden layer is used in the feed-forward neural network structure in order to discriminate one class of patterns from the other class in real time. In this work, the training of the network is performed in the off-line computing environment and the results of the training are configured to the hardware in order to minimize the latency of the neural computation. With five 8-bit input patterns, six hidden nodes, and one 8-bit output, the implemented hardware neural network makes decisions on a set of input patterns in 11 clock cycles, or less than 200 ns with a 60 MHz clock. The result from the hardware neural computation is well predictable based on the off-line computation. This implementation may be used in level 1 hardware triggers in high energy physics experiments.

KW - Artificial neural network

KW - FPGA

KW - Level 1 trigger

KW - VHDL

UR - http://www.scopus.com/inward/record.url?scp=35348832307&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=35348832307&partnerID=8YFLogxK

U2 - 10.1016/j.nima.2007.08.163

DO - 10.1016/j.nima.2007.08.163

M3 - Article

AN - SCOPUS:35348832307

VL - 581

SP - 816

EP - 820

JO - Nuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors and Associated Equipment

JF - Nuclear Instruments and Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors and Associated Equipment

SN - 0168-9002

IS - 3

ER -