NASA Logo

NTRS

NTRS - NASA Technical Reports Server

Back to Results
robust neural classifier circuits using asynchronous designAerospace neural circuits must be adaptive, offer a practical size-performance ratio, and be environmentally robust. Our approach to building such circuits combines asynchronous design with a new fuzzy/neural classifier model. Asynchronous circuits offer many design advantages for neural hardware and our hybrid fuzzy/neural model, using mainly min and max operators, promises a low circuit complexity. The general approach is described and a description of a use of rule-induction to further reduce circuit complexity is described.
Document ID
19940016638
Document Type
Conference Paper
Authors
Hurdle, John F.
(Utah Univ. Salt Lake City, UT, United States)
Conwell, Peter R.
(Utah Univ. Salt Lake City, UT, United States)
Brunvand, Erik L.
(Utah Univ. Salt Lake City, UT, United States)
Date Acquired
September 6, 2013
Publication Date
January 1, 1993
Publication Information
Publication: New Mexico Univ., The Fifth NASA Symposium on VLSI Design
Subject Category
ELECTRONICS AND ELECTRICAL ENGINEERING
Distribution Limits
Public
Copyright
Work of the US Gov. Public Use Permitted.

Related Records

IDRelationTitle19940016163Analytic PrimaryTwenty-Fourth Lunar and Planetary Science Conference. Part 3: N-Z