Ultra-Wideband and Interleaved Polyfractal Antenna Arrays

Open Access
Petko, Joshua Stanton
Graduate Program:
Electrical Engineering
Doctor of Philosophy
Document Type:
Date of Defense:
July 17, 2008
Committee Members:
  • Claude W De Pamphilis, Committee Member
  • John Douglas Mitchell, Committee Member
  • Ram Mohan Narayanan, Committee Member
  • Douglas Henry Werner, Committee Chair
  • ultra-wideband
  • autopolyploidy
  • genetic algorithm
  • array
  • antenna
  • polyfractal
  • fractal
  • multi-beam
  • interleaved
Recently, in order to successfully combine the positive attributes of both periodic and random arrays into one design, a novel class of arrays, known as fractal-random arrays, has been introduced. In addition, global optimization techniques, such as genetic algorithms, have been applied to antenna array layouts to provide highly directive, thinned, frequency agile, and shaped-beam antenna systems. However, these methodologies have their limitations when applied to more demanding design scenarios. Global optimizations are not well equipped to handle the large number of parameters used to describe large-N arrays, and fractal-random arrays lack the recursive properties needed to reconstruct their geometries exactly from a small set of parameters. To overcome these difficulties, a new class of arrays, called polyfractal arrays, is introduced in this dissertation that possess properties well suited for the optimization of large-N arrays. These polyfractal arrays possesss underlying self-similar properties that can be exploited to exactly reconstruct the array geometries from small sets of parameters and to increase the speed of the associated array factor calculations. In addition, an autopolyploidy-based chromosome expansion native to polyfractal arrays is introduced that can dramatically accelerate the genetic algorithm optimization process. This process allows the genetic algorithm to first evolve simple designs very quickly and then add increasing levels of complexity when they are needed to continue the optimization. The entire procedure has been shown to be very effective in creating optimized large-N antenna layouts. For example a 1616-element linear array can be optimized to possess a -24.30 dB sidelobe level with a 0.0056 degree half-power beamwidth. In addition, robust Pareto optimization techniques can be applied to reduce the peak sidelobe levels at several frequencies specified over a wide bandwidth. This procedure can lead to ultra-wideband antenna array designs, with one example discussed in this dissertation maintaining a -19.34 dB peak sidelobe level with no grating lobes from a range covering 0.5 wavelength to 20.0 wavelength minimum interelement spacings, corresponding to a bandwidth of at least 40:1. These powerful array designs can be utilized as building blocks in robust, multi-beam, ultra-wideband antenna array systems.