Development of Novel Nuero-Fuzzy Techniques

By: Krishna, T VamsiContributor(s): Satapath, J K [Supervisor] | Department of Electrical EngineeringMaterial type: TextTextLanguage: English Publisher: 2006Description: 75 pSubject(s): Engineering and Technology | Electronics and Communication Engineering | Fuzzy Systems | Adaptive SystemsOnline resources: Click here to access online Dissertation note: Thesis (M.Tech (R)- National Institute of Technology, Rourkela Summary: N ovel approaches for de signing adaptive sche mes based on neuro-fuzzy platform have been developed. Two kinds of adaptiv e sc hemes na mel y, adaptive equalization and system identification are im plemented using th e dev eloped pro posed techniques. The Radial basis function (RBF) equalizer is chosen as a ca se study for adaptive equalization of the digital communication ch annels. An efficien t method for reducing the cente rs of a RBF equalizer based on eigenvalue analy sis is presen ted. The efficien cy of t he m ethod is furt her verified for RBF equaliz ers with deci sion fee dback for tackling channels with overlappi ng channel state s. A comparative study between the p roposed center reduction technique and other center reduction techniques for the RBF equa lizer is discussed. In another breakthroug h a parallel int erpretation of the ANFIS (adap tive network based fuzzy inference sy ste ms) architecture i s proposed. This approach helps to investigate the role of the fuzzy inferenc e part and the sub-filter part of the ANFIS separ ately . The parallel interpretation of the ANFIS redefines the opinion reserved for the fuzzy in fere nce sy ste m, thereby allowing it to be considered as a fuzzy weighted sub-filter netw ork, with the weighting functions and the sub- filter units arranged parallely . T his approach motivated in developing m any novel schem es for designing adaptive sy ste ms with application to sy stem identification problems. Finally, the lim itations of the ANFIS architecture are dis cuss ed. These li mitations are exploited to develop neur o-fuzzy m odels sim ilar to the ANF IS with the objective of reducing the num ber of param eters in co mpariso n to the ANF IS. Th e developed neuro-f uzzy m odels are co mpared to the ANFIS in term s of the ti me required fo r learning and n umber of param eters to b e adapted.
Tags from this library: No tags from this library for this title. Log in to add tags.
    Average rating: 0.0 (0 votes)
Item type Current location Collection Call number Status Date due Barcode
Thesis (Ph.D/M.Tech R) Thesis (Ph.D/M.Tech R) BP Central Library
Thesis Section
Reference Not for loan T19

Thesis (M.Tech (R)- National Institute of Technology, Rourkela

N
ovel approaches for de
signing adaptive
sche
mes
based on neuro-fuzzy
platform
have been developed. Two kinds of adaptiv
e sc
hemes na
mel
y, adaptive equalization and
system
identification are im
plemented using th
e dev
eloped pro
posed techniques. The Radial
basis function (RBF) equalizer is
chosen as
a ca
se study
for adaptive equalization of the
digital communication ch
annels. An efficien
t method for
reducing the cente
rs of a RBF
equalizer based on eigenvalue analy
sis is presen
ted. The efficien
cy of t
he m
ethod is furt
her
verified for RBF equaliz
ers with deci
sion fee
dback for tackling channels with overlappi
ng
channel state
s. A
comparative study
between
the p
roposed center reduction
technique and
other center reduction techniques for the RBF
equa
lizer is discussed. In another breakthroug
h
a parallel int
erpretation of the ANFIS
(adap
tive network based fuzzy
inference sy
ste
ms)
architecture i
s proposed.
This approach helps to
investigate the role of the fuzzy
inferenc
e
part and the sub-filter part of the ANFIS separ
ately
. The parallel interpretation of the ANFIS
redefines the
opinion reserved for the fuzzy
in
fere
nce sy
ste
m, thereby
allowing it to be
considered as a fuzzy
weighted sub-filter netw
ork, with the weighting functions and the sub-
filter units arranged parallely
. T
his approach
motivated in developing m
any
novel schem
es
for designing
adaptive sy
ste
ms with application to
sy
stem
identification problems. Finally, the
lim
itations of the ANFIS
architecture
are dis
cuss
ed. These li
mitations are
exploited to
develop neur
o-fuzzy m
odels sim
ilar to the ANF
IS with the
objective of reducing the num
ber
of param
eters in co
mpariso
n to the ANF
IS. Th
e developed neuro-f
uzzy
m
odels
are co
mpared
to the ANFIS in term
s of the ti
me required fo
r learning and n
umber of param
eters to b
e
adapted.

There are no comments on this title.

to post a comment.

Implemented and Maintained by Biju Patnaik Central Library.
For any Suggestions/Query Contact to library or Email: library@nitrkl.ac.in OR bpcl-cir@nitrkl.ac.in. Ph:91+6612462103
Website/OPAC best viewed in Mozilla Browser in 1366X768 Resolution.

Powered by Koha