<p>The diffraction of H-polarized electromagnetic (EM) waves by a finite-length plate in an anisotropic medium is rigorously analyzed using a hybrid analytical-machine learning framework that combines the Wiener-Hopf technique and artificial neural networks (ANNs). Analytical solutions for the diffracted field are derived under impedance boundary conditions, explicitly isolating the separated and interaction field components through asymptotic stationary phase analysis. These solutions reveal anisotropic suppression of field oscillations, which is critical for mitigating signal distortion in ionospheric communication. The analytical results are later validated against an artificial neural network (ANN) trained on parameterized plasma conditions for the observational angle. The ANN architecture consists of a simple composition of 1 hidden layer with Rectified Linear Unit (ReLU) activation that effectively achieved a mean squared error (MSE) of <InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="11082_2025_8322_Article_IEq1.gif" Format="GIF" Height="17" Rendition="HTML" Resolution="72" Type="Linedraw" Width="74" /> </InlineMediaObject> <EquationSource Format="TEX">$3.2 \times 10^{-3}$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mn>3.2</mn> <mo>×</mo> <msup> <mn>10</mn> <mrow> <mo>-</mo> <mn>3</mn> </mrow> </msup> </mrow> </math></EquationSource> </InlineEquation> for anisotropic cases and <InlineEquation ID="IEq2"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="11082_2025_8322_Article_IEq2.gif" Format="GIF" Height="17" Rendition="HTML" Resolution="72" Type="Linedraw" Width="82" /> </InlineMediaObject> <EquationSource Format="TEX">$4.63\times 10^{-2}$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mn>4.63</mn> <mo>×</mo> <msup> <mn>10</mn> <mrow> <mo>-</mo> <mn>2</mn> </mrow> </msup> </mrow> </math></EquationSource> </InlineEquation> for isotropic cases with <InlineEquation ID="IEq3"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="11082_2025_8322_Article_IEq3.gif" Format="GIF" Height="16" Rendition="HTML" Resolution="72" Type="Linedraw" Width="74" /> </InlineMediaObject> <EquationSource Format="TEX">$R \ge 0.999$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>R</mi> <mo>≥</mo> <mn>0.999</mn> </mrow> </math></EquationSource> </InlineEquation> and accurately predicted anisotropic suppression of field oscillations compared to isotropic cases. Computational efficiency is enhanced by three orders of magnitude: analytical solutions that require hours are replaced by ANN predictions in a few seconds. This work establishes the first integration of Wiener-Hopf asymptotics with machine learning for plasma-EM interactions, offering a scalable tool for real-time optimization of ionospheric communication systems under dynamic plasma environments.</p>

Weiner-Hopf assisted artificial neural network for H-polarized wave diffraction in anisotropic media

期刊论文
发表日期：

2025年7月11日

查看全文

Sajjad Hussain,
Aysha Bibi,
Maham Zubair,
Sabih Haider

摘要

The diffraction of H-polarized electromagnetic (EM) waves by a finite-length plate in an anisotropic medium is rigorously analyzed using a hybrid analytical-machine learning framework that combines the Wiener-Hopf technique and artificial neural networks (ANNs). Analytical solutions for the diffracted field are derived under impedance boundary conditions, explicitly isolating the separated and interaction field components through asymptotic stationary phase analysis. These solutions reveal anisotropic suppression of field oscillations, which is critical for mitigating signal distortion in ionospheric communication. The analytical results are later validated against an artificial neural network (ANN) trained on parameterized plasma conditions for the observational angle. The ANN architecture consists of a simple composition of 1 hidden layer with Rectified Linear Unit (ReLU) activation that effectively achieved a mean squared error (MSE) of $3.2 \times 10^{-3}$ $3.2 \times 10^{- 3}$ for anisotropic cases and $4.63\times 10^{-2}$ $4.63 \times 10^{- 2}$ for isotropic cases with $R \ge 0.999$ $R \geq 0.999$ and accurately predicted anisotropic suppression of field oscillations compared to isotropic cases. Computational efficiency is enhanced by three orders of magnitude: analytical solutions that require hours are replaced by ANN predictions in a few seconds. This work establishes the first integration of Wiener-Hopf asymptotics with machine learning for plasma-EM interactions, offering a scalable tool for real-time optimization of ionospheric communication systems under dynamic plasma environments.