Free space material characterization using genetic algorithms

Abstract

This work demonstrates the use of genetic algorithms (GA) to transform the traditional free space characterization process into an optimization process. The basis of the process is to compare the measured reflection coefficient to the reflection coefficient computed by the GA. The GA computes the reflection coefficient using candidate solutions of the permittivity and permeability and the Fresnel reflection coefficient equations. The permittivity and permeability that yield the reflection coefficient closest to the measured reflection coefficient is determined to be the best solution. Results using synthetic reflection coefficient data using a Plexiglas sample show accurate extraction of the permittivity with both transverse electric and transverse magnetic polarized plane waves.