Skip to main navigation Skip to search Skip to main content

A Data-driven approach for Solving 2D Combined-field Integral Equations Based on WaveKAN

  • Lei Wu
  • , Di Wu
  • , Xin Zhang
  • , Tao Shan*
  • *Corresponding author for this work
  • Beihang University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In this work, a data-driven model is proposed to solve two-dimensional combined-field integral equations (CFIE) for perfectly electrically conducting (PEC) objects by leveraging Wavelet Kolmogorov-Arnold Networks (WavKAN). WavKAN employs mother wavelet functions as kernel functions, utilizing translation and scaling operations to adaptively capture features. This capability enables the simultaneous extraction of temporal and frequency characteristics at multiple resolutions, making WavKAN particularly well suited for electromagnetic modeling. Numerical experiments are conducted to validate the effectiveness of the proposed WavKAN framework.

Original languageEnglish
Title of host publication2025 International Applied Computational Electromagnetics Society Symposium, ACES-China 2025 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781733467711
DOIs
StatePublished - 2025
Event2025 International Applied Computational Electromagnetics Society Symposium, ACES-China 2025 - Huangshan, China
Duration: 8 Aug 202511 Aug 2025

Publication series

Name2025 International Applied Computational Electromagnetics Society Symposium, ACES-China 2025 - Proceedings

Conference

Conference2025 International Applied Computational Electromagnetics Society Symposium, ACES-China 2025
Country/TerritoryChina
CityHuangshan
Period8/08/2511/08/25

Keywords

  • PEC objects
  • Wavelet Kolmogorov-Arnold Networks
  • combined-field integral equation
  • electromagnetic modeling

Fingerprint

Dive into the research topics of 'A Data-driven approach for Solving 2D Combined-field Integral Equations Based on WaveKAN'. Together they form a unique fingerprint.

Cite this