[Paper Review] IPO: Iterative Physical Optics Image Approximation

Shaolin Liao, Lu Ou|arXiv (Cornell University)|Jul 3, 2020

Electromagnetic Scattering and Analysis20 references4 citations

TL;DR

This paper proposes an Iterative Physical Optics (IPO) method that enhances the accuracy of conventional Physical Optics (PO) surface current approximations for Perfect Electric Conductor (PEC) surfaces with high curvature, such as parabolic dish antennas and PEC spheres. By iteratively correcting surface currents using local plane wave approximations to satisfy the electric field boundary condition, IPO achieves over two orders of magnitude improvement in surface current accuracy compared to standard PO.

ABSTRACT

An improved Iterative Physical Optics (IPO) image approximation method has been presented to dramatically increase the accuracy of the approximation and extend its applicability to PEC surfaces with smaller radii or larger curvatures. Starting from the first-order conventional PO image approximation, the IPO image approximation method iteratively correct the surface current to compensate the deviation of the electric field boundary condition on the PEC surfaces, making use of the local plane wave approximation. Numerical validations with two popular PEC surfaces, i.e., the parabolic dish antennas and the PEC spheres, are carried out and the results show that the IPO approximation method increases the surface current accuracy by more than two orders of magnitude, compared to the conventional PO image approximation method.

Motivation & Objective

• Address the limited accuracy of conventional Physical Optics (PO) for PEC surfaces with small radii or high curvature.
• Overcome the fundamental limitation of PO, which is exact only for planar surfaces and approximate for curved ones.
• Develop a computationally efficient method that extends PO applicability to electrically large, highly curved PEC structures.
• Achieve high-accuracy surface current approximation without resorting to full-wave rigorous methods like Method of Moments (MoM).

Proposed method

• Start with the first-order PO surface current approximation: $\overline{J}^{PO} = 2\hat{n} \times \overline{H}^i$.
• Compute the scattering electric field $\overline{E}^s$ using the 2D convolution of the surface current with the dyadic Green's function $\overline{\overline{G}}_e$.
• Calculate the boundary condition deviation $\hat{n} \times \delta\overline{E}^s$ between the actual and approximated electric fields.
• Apply local plane wave approximation to relate field deviations to magnetic field deviations: $\delta\overline{E}^s = \eta \delta\overline{H}^s \times \hat{n}$.
• Update the surface current iteratively using $\overline{J}^{IPO}_{k+1} = \overline{J}^{IPO}_k + \Delta\overline{J}_k$, where $\Delta\overline{J}_k$ corrects the boundary condition error.
• Repeat the process until convergence is achieved, ensuring the electric field boundary condition is progressively satisfied.

Experimental results

Research questions

• RQ1 Can the accuracy of conventional PO be significantly improved for highly curved PEC surfaces such as parabolic dishes and spheres?
• RQ2 To what extent can iterative correction of surface currents reduce boundary condition errors on curved PEC surfaces?
• RQ3 How does the IPO method compare quantitatively to conventional PO in terms of surface current accuracy for varying curvature?
• RQ4 Does the IPO method maintain computational efficiency while achieving high accuracy on electrically large, curved PEC structures?
• RQ5 What is the convergence behavior of the IPO algorithm across different surface curvatures and geometries?

Key findings

• The IPO method improves surface current accuracy by more than two orders of magnitude compared to conventional PO for parabolic dish antennas with focus lengths ranging from $50\lambda$ to $150\lambda$.
• For PEC spheres of radius $R=60\lambda$, the IPO method reduces surface current deviation by over 200 times compared to conventional PO.
• Convergence of the IPO algorithm is observed in numerical results, with surface current deviation decreasing significantly across iterations for both parabolic dishes and PEC spheres.
• The surface current deviation maps (Figs. 3–7) show that IPO effectively suppresses errors in all Cartesian components ($J_x$, $J_y$, $J_z$) across the surface.
• The method maintains computational efficiency by avoiding full MoM solves while achieving near-rigorous accuracy on curved PEC surfaces.
• The improvement is consistent across different curvature radii, demonstrating robustness for surfaces with small radii or high curvature.