الفهرس 
Scope and outline of the thesis
OPTIMIZATION TECHNIQUESOnly 14 pages are availabe for public view
 |  | from | 120 |  |  |
| | | from | 120 | | |
Abstract
In recent years, modern wireless communication systems and information transfer have highlighted the major need for antenna design advancements as an essential part of any wireless system. Microstrip patch antenna fulfills the desired wireless systems’ necessities. These antennas have triggered extensive research due to their significant advantages which includes compact size, low profile, light weight, low volume, relatively low manufacturing cost, ease of fabrication, and compatibility with integrated circuits. Microstrip patch antennas have led to diversified applications utilizing microwave systems such as biomedical systems, radars, mobile, satellite communications, and global positioning satellite (GPS).
The main aim of this thesis is utilizing evolutionary algorithms to design a low profile and efficient antenna which can be attained in the antenna design construction to meet the extensive emerging applications in wireless communication systems. Evolutionary algorithms have emerged as a promising solution to optimize some designs of microstrip antenna seeking an efficient performance in addition to compromising competing goals simultaneously. Two distinct design geometries have been implemented in the scope of this research.
Evolutionary algorithms are applied to cross aperture coupled circularly polarized microstrip antenna which is designed to attain a low axial ratio over a wide bandwidth and optimal impedance matching. The microstrip antenna is designed. A 50Ω microstrip feed line is used. A software simulation program (CST microwave studio) is used to compare the performance of the antenna in terms of return loss and axial ratio with evolutionary algorithms. Two distinct evolutionary algorithms; Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) have been utilized in the fitness function optimization to ensure optimal axial ratio and impedance matching. The optimal patch dimensions, the aperture length, and the microstrip feed line length are obtained from both methods and their performances are compared. Moreover, the antenna is fabricated and the network analyzer is used for measuring the return loss. A
III
comparative analysis is carried out demonstrating a high agreement between measured and simulated return loss.
Moreover, the artificial neural networks (ANNs) is employed in WiFi antenna design and modeling where ANN is used to predict the antenna characteristics to be compatible with the demand of wireless communication devices. The ANN is favored over other modeling techniques as it models non-linear relationships between the antenna input and output, needs less computation time compared to computer aided designs that use numerical methods intensively as well as reduction of mathematical computational complexties. Thus, the ANN is employed to model the antenna design parameters. This is attained by the use of ANN toolbox with the aid of MATLAB.
The ANN prior training data set is provided by CST simulator. The ANN is trained using Feed Forward Back propagation algorithm by adjusting the weights to model the particular learning task. The feed forward and back propagation is repeated until the error is minimal enough. Finally, the antenna is analyzed and verified against manifuctured antenna achieving compatiable results.
Keywords: Microstrip Antenna, Particle Swarm Optimization, Genetic Algorithm, Artificial Neural Networks, WiFi Antenna.