My M.S. Thesis

Nature-Inspired Optimization Techniques Applied to Antennas for Wireless Communications and Radar

Nature-inspired optimization techniques were the primary theme of my research in the first two years at UCLA. I have tried my best to make the integration of these techniques as accessible and understandable as possible to a wide audience. Below I have a brief summary of my thesis just to give some flavor to those interested. I also have included some of the beginning chapters so that people can get a preview of my thesis.

Brief Motivation and Summary

Cell phones and wireless systems have become much more than just a part of one’s daily routines. For many people, the last time they checked their phone for emails, text messages, or Facebook is less than 10 minutes ago. In fact, communication and radar systems are among the leading technologies that have made a significant impact on society.  They have enabled worldwide communication, expanding one’s local community in the surrounding area to a global community.  While the system concepts may seem straightforward, designing a practical, realizable antenna solution can be quite difficult. Electromagnetic waves, electronic circuits, and antennas are all guided by the principles of Maxwell’s equations. For complex antenna designs, deriving any sort of equation to describe the performance is intractable unless simple shapes are used. With the proliferation of antenna simulation tools and computing, one can at least determine the antenna performance numerically, but the best design choice is not always clear. In light of this problem, many modern techniques have emerged to provide final antenna design solutions which can improve system performance. Among these, nature-inspired optimization techniques are one of the leading candidates in finding new designs where the classic textbook designs fail to satisfy the requirements.

In this work, two nature-inspired optimization techniques, namely Particle Swarm Optimization (PSO) and Covariance Matrix Adaptation Evolution Strategy (CMAES), are presented. Some comparisons are made between the two algorithms in different applications and different optimization problems. First, a comparison of each algorithm in resource limited problems is demonstrated using mathematical functions. Next, a comparison is shown between the two algorithms for a real-world antenna design problem for radar systems. In particular, a weather radar antenna array element is optimized using two different approaches, and some suggestions are made for antenna designers hoping to implement dual polarized antenna arrays for new use in weather radar systems. In the last half of this work, the Particle Swarm algorithm is applied to two other antenna systems. The first application of PSO investigates the use of a smooth contour septum design in circular waveguide for possible use in high power microwave systems. Similar antenna performance compared to a stepped septum in circular waveguide is demonstrated using the Sigmoid function contour. PSO is also applied on two newly proposed reconfigurable E-shaped patch antenna designs. Two reconfigurability mechanisms are introduced, namely a polarization (RHCP/LHCP) reconfigurable design and a frequency reconfigurable design. Both designs are optimized using a simple MEMS circuit model for fast optimization, and possible bias network implementations are discussed. Wideband patch designs are realized with these optimizations, and prototypes are fabricated and measured to validate these designs.