Optimized artificial magnetic conductor design using CG-FFT and BPSO

Abstract

A method for the optimization of Artificial Magnetic Conductors (AMC) based on a planar frequency selective surface (FSS) sandwiched in multilayer dielectric media backed by a conductor is presented in this work. The optimization process is applied to tune the periods and geometry of the FSS unit cell along with the permittivities and thicknesses of the dielectric layers. The method uses a Binary version of the Particle Swarm Optimization algorithm (BPSO) to drive the optimization and a Conjugate Gradient Fast Fourier Transform (CG-FFT) solver to analyse the structures proposed by the optimizer and to obtain their reflection coefficient. The only link between the physical problem and the optimizer is given by a fitness function that forces the modulus of the reflection coefficient of the structure to one and try to keep its phase as close to zero as possible. Several fitness functions are analysed and compared in this work. In order to demonstrate the usefulness of the approach, results for a couple of AMC designs optimized using this methodology are presented.