Investigations on Optimization of Antenna Parameters using the Adaptive Neuro-Fuzzy Inference System by Aarti Kishor Gehani

Guided by: Dr. Dhaval Pujara With Synopsis and CD 11EXTPHDE51

ABSTRACT:
An antenna is at the heart of all wireless communication systems. In fact, wireless
communication is impossible without antennas. Antennas play a critical role in deciding
the overall performance of the wireless system. Hence, designing an antenna
with required specifications is crucial, as it involves trade-offs between various design
and performance parameters. Over the last several decades, many conventional
electromagnetic techniques such as the finite element method, method of moments,
differential equation and others have been employed to design and optimize antennas.
These techniques involve relatively complex analytical models and lengthy calculations,
which tends to make them complicated and time consuming at times. Additionally,
they require fairly high computational resources including a fast processor
and substantial memory to arrive at a final solution.
During the recent years, researchers have explored various soft computing techniques
to design and analyze different types of antennas. Many antenna problems
have been successfully solved by different soft computing techniques, including genetic
algorithms, artificial neural networks, particle swarm optimization, ant colony
optimization and bacteria foraging optimization algorithm, to name a few. A comprehensive
study has been carried out to understand the operation of these algorithms,
their strengths, limitations and applications towards optimization of antenna parameters,
and the summary of the same is presented in this thesis.
From the literature survey, it was observed that the adaptive neuro-fuzzy inference
system (ANFIS), which is a combination of artificial neural networks and fuzzy
inference system, is relatively less explored in the field of antenna engineering. The
present work aims to investigate the potential of the adaptive neuro-fuzzy inference
system in the field of antenna optimization. The objective here is not to compare the
performance of the adaptive neuro-fuzzy inference system with other soft computing techniques, but rather prove that it can also be used for the analysis and synthesis of
a wide variety of antennas. In order to understand the development of the adaptive
neuro-fuzzy inference system based models, examples are presented for the analysis
and synthesis of different antenna structures like the circularly polarized elliptical
patch antenna, multi-band hexa-band planar inverted-F antenna, Sierpinski carpet
fractal antenna and Sierpinski gasket fractal antenna. The problem of diagnosing the
location of faulty elements in a planar array is also solved using the adaptive neurofuzzy
inference system. The results obtained through the proposed model in each case
are validated either by evaluating the statistical parameters or by comparing them
with the measured results. Based on the present work, it can be concluded that the
ANFIS gives accurate results compared to the ANNs, requires less number of training
dataset and is fast. Thus, it can be a potential candidate for solving many antenna
design problems.