Application of iterative techniques for electromagnetic wave scattering from dielectric random rough surfaces

Mobile radio planning requires accurate prediction of electromagnetic field strengths over large terrain profiles. However the conventional method of moments (MoM) becomes unsuitable for electrically large rough dielectric surfaces, because of the O(N³) computational cost due to the large number of surface unknowns N. Iterative Methods are beneficial methods for faster electromagnetic problem solutions. By using such methods, very accurate results can be achieved, causing a computational cost of O(N²). In this work, among the stationary iterative methods; Forward-Backward Method (FBM), and among the nonstationary iterative ones; Conjugate Gradient Squared (CGS), BiConjugate Gradient Stabilized (Bi-CGSTAB) and Quasi Minimal Residual (QMR) Methods are presented to investigate the electromagnetic wave scattering from dielectric random rough surfaces. These techniques are compared to each other over various kinds of surface models that reflect the real terrains to find out the best solution methodologies. Furthermore, efficiency of the methods are assessed by comparing the obtained scattering results, normalized radar cross sections (NRCS) of the surfaces considered, with the numerically exact ones computed by employing the MoM.