DEVELOPMENT OF RECONFIGURABLE NONLINEAR CIRCUITS FOR NEURAL NETWORKS

دانشگاه آزاد اسلامی اصفهان(خوراسگان)

منو

رکورد قبلی

رکورد بعدی

" DEVELOPMENT OF RECONFIGURABLE NONLINEAR CIRCUITS FOR NEURAL NETWORKS "

Document Type	:	Latin Dissertation
Language of Document	:	English
Record Number	:	152026
Doc. No	:	ET23818
Main Entry	:	Abdullah B. Shwehneh
Title Proper	:	DEVELOPMENT OF RECONFIGURABLE NONLINEAR CIRCUITS FOR NEURAL NETWORKS
Note	:	This document is digital این مدرک بصورت الکترونیکی می باشد
Abstract	:	In this thesis, a new generic CMOS circuit for realizing different nonlinear functionsfrom the same topology is presented. This circuit can be very useful in NeuralNetworks applications as it implements the main nonlinearities required by manytypes of Neural Networks. With transistors operating in strong inversion the circuitcan be digitally configured to realize any of the following four functions: Gaussian(Radial Basis), Sigmoid and two piecewise linear functions - Triangular and Satlin.The circuit can approximate these functions with RRMS error less than 1 . It isshown that the center, width, peak amplitude and slope of the dc transfer curve can beindependently controlled. Simulation results using 0.18 CMOS process modelparameters of TSMC technology are included.Keywords: Analog Circuits, Programmable Neural Networks.,.............-....,....-...,..tested for theQ1 PC1 bus cardBoth these projects mere sofixare des elopment efforts tonards contributing to dlfferentaspects of Roboucs and lZ1echatronics projects m the Controls and Roboucs Group..
Subject	:	Electericl tess
	:	برق
electronic file name	:	TL47078.pdf
Title and statement of responsibility and	:	DEVELOPMENT OF RECONFIGURABLE NONLINEAR CIRCUITS FOR NEURAL NETWORKS [Thesis]