Thesis Defense Presentation By John Kassab Committee Members: Dr. Santosh Nagaraj (chair)

Thesis Defense Presentation


By John Kassab


Committee Members:

Dr. Santosh Nagaraj (chair)

Dr. fred harris

Dr. Carl Eckberg


Time: 2:00
-
2:50 pm

Location: E423

Date: 3/18


Abstract:

Orthogonal Frequency Division Multiplexing, OFDM, is a powerful
communic
ations scheme that has grown much popularity because of its ability to support
high data rates, and at the same time, cope with severe multi
-
path fading channels that
cripple a data link. OFDM exploits the concepts of orthogonality and diversity, which
wh
en utilized effectively, can support increased data rates and improve performance over
traditional means. OFDM is a multi
-
carrier modulation scheme which transmits multiple
symbols in parallel. Techniques have been developed in OFDM that make it possible

for
the channel to be estimated in which an adaptive modulation of the N subcarriers is
possible. Adaptive modulation is a means which the power or modulation order of the
subcarriers can be varied so that the performance of the system can be improved wh
en a
fading channel is present.



Artificial Neural Networks, ANNs, are mathematical models of the networks of
biological neurons. Neural networks have been used to infer a function from a set of
observations which normally are too complex to model. A ne
ural network is trained by
some learning paradigm which is then used to do a variety of tasks such as data mining,
classification, regression and much more.


This thesis proposes that the performance of an OFDM communication system
can be improved throug
h adaptive modulation

by means of artificial neural network. In
order to evaluate the effectiveness of this, simulations are run and tested against this
hypothesis, where the figure of merit used is symbol error probability.