General Local Iterative Physical Optics CEM for Layered Dielectrics with Moderately Smooth Rough Surfaces

• Published in: IEEE journal on multiscale and multiphysics computational techniques pp. 1 - 16
• Main Authors: Liao, Shaolin, Liang, Jiong, Zhang, Chuangfeng, Li, Qun, Li, Jinxin, Soekmadji, Henry
• Format: Journal Article
• Language: English
• Published: IEEE 10.10.2025
• Subjects: Approximation algorithms; Computational Electromagnetics (CEM); Dielectric layers; Dielectrics; Electromagnetic scattering; general Local Iterative Physical Optics (gLIPO); Method of moments; Nonhomogeneous media; Optical surface waves; Rough surfaces; Sea surface; Surface impedance; Surface roughness
• ISSN: 2379-8815; 2379-8815
• DOI: 10.1109/JMMCT.2025.3620470

The accurate computation of electromagnetic scattering from electrically large dielectric layered media with rough surfaces remains a complex challenge, demanding highly efficient computational electromagnetics (CEM) algorithms. This paper presents a novel general local iterative physical optics (gLIPO) algorithm, tailored for the simulation of electromagnetic scattering in dielectric layered media characterized by relatively smooth surface irregularities. The gLIPO algorithm iteratively refines the equivalent local surface currents in accordance with the physical optics (PO) principle at dielectric interfaces, effectively mitigating electromagnetic field discontinuities across the media. Rigorous update equations are derived for both equivalent electric and magnetic surface currents. A series of comprehensive numerical simulations are conducted for four representative scenarios: 1) a single-layer ocean medium at 300 MHz; 2) a single-layer soil medium at 800 MHz; 3) a double-layer ice/ocean medium at 300 MHz; and 4) a tunnel communication scenario at 30 GHz. The results consistently demonstrate that the gLIPO algorithm converges in fewer than five iterations, reducing the maximum relative error in the equivalent surface currents to below <inline-formula><tex-math notation="LaTeX">10^{-5}</tex-math></inline-formula>, benefiting from its linear computational complexity and memory footprint of <inline-formula><tex-math notation="LaTeX">O(N)</tex-math></inline-formula>. In contrast, the method of moments (MoM) typically requires several dozen iterations, rendering gLIPO approximately an order of magnitude faster, even outperforming the multi-level fast multipole algorithm (MLFMA). Furthermore, gLIPO circumvents the need to compute and store the impedance matrix, as required by MoM, leading to substantial savings in both computational time and memory resources. The gLIPO algorithm offers significant advantages for applications such as large-scale multiple-input multiple-output (MIMO) channel state information (CSI) simulations in 5G and future wireless communication systems, making it a valuable tool for advancing electromagnetic simulation capabilities.