Artificial Frequency Match Neuron Implemented with Digital Logic

<p>This thesis proposes a digital artificial neuron which uses only digital logic. The purpose is to create a hardware neuron on a digital device. While conventional neurons use real valued software implemented transfer functions and threshold values to determine the output, the proposed neuron converts inputs into square waves of frequencies determined by weights, and the time that the input waves require to produce a certain pattern is used to determine the output. This neuron was tested on a software simulator and successfully implemented on a Field Programmable Gate Array (FPGA) to prove its viability.</p>

Name thesis:degree_name

Master of Science (MS) in Computer Science

Level thesis:degree_level

Thesis: EWU Only

Discipline thesis:degree_discipline

Computer Science

Year

2014

Author dc:creator

• Ellis, David J.

Contributors dc:contributor

• Dr. Kosuke Imamura, PhD
• Dr. Carol Taylor, PhD
• Doris Munson M.L.S.

• Neural networks (Computer science)
• Evolutionary computation
• Field programmable gate arrays
• Computer Sciences

Statement dc:rights

• Access perpetually restricted to EWU users with an active EWU NetID

Repository record dc:identifier

https://dc.ewu.edu/theses/1

OAI identifier oai:identifier

oai:dc.ewu.edu:theses-1000