Michael A. Saville, Ph.D., P.E. (Advisor); Josh Ash, Ph.D. (Committee Member); Yan Zhuang, Ph.D. (Committee Member); Hirsch Chizever, Ph.D. (Committee Member)
Master of Science in Electrical Engineering (MSEE)
A method for the beam forming control of an array of reconfigurable antennas is presented. The method consists of using two parallel convolutional neural networks (CNNs) to analyze a desired radiation pattern image, or mask, and provide a suggestion for the reconfigurable element state, array shape, and steering weights necessary to obtain the radiation pattern. This research compares beam forming systems designed for three distinct element types: a patch antenna, a reconfigurable square spiral antenna restricted to a single reconfigurable state, and the fully reconfigurable square spiral. The parametric sweeps for the design of the CNNs are presented along with several examples of activation maps and mask-suggestion pairs. The beam forming error for each element type in both isolated and embedded cases is calculated over a set of 100 random masks. The network performance is reported as a root mean square error in steering direction and beam widths in both azimuth and elevation.
Department or Program
Department of Electrical Engineering
Year Degree Awarded
Copyright 2022, all rights reserved. My ETD will be available under the "Fair Use" terms of copyright law.