Spectral domain interpolation of antenna radiation patterns using Model-Based Parameter Estimation and genetic algorithms

Model-Based Parameter Estimation (MBPE) techniques that use a rational function fitting model with complex coefficients have been shown to yield approximations which accurately describe the frequency dependence of antenna characteristics such as input impedance or admittance. In this paper we apply MPBE to the radiated electric field of a 0.5 m dipole and use the rational function fitting model to interpolate the electric field spectrum of this antenna over the interval 150 to 950 MHz. A method is presented for using these electric field interpolations in conjunction with the associated impedance interpolations in order to reconstruct gain patterns at any frequency within this model range. Conventional Pade approximations require a matrix inversion operation in order to estimate the parameters of the rational function fitting model. In this paper we introduce a Genetic Algorithm (GA) approach for obtaining the required fitting model parameters. This new approach to MBPE has the advantages that it is very general and avoids the need for any kind of matrix inversion.