Design of a Wideband Rectangular Dielectric Resonator Antenna Using Deep Learning

  • Xijiao Yang
  • , Yuanchao Shi
  • , Bin Li
  • , Liangliang Cui
  • , Chen Yang
  • , Weidong Hu

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

Abstract

A method using deep learning (DL) and genetic algorithm (GA) is proposed for the design of a wideband dielectric resonator antenna (DRA). The three-step approach involves data preparation for network training, construction of the neural network architecture, and optimization of parameters using the genetic algorithm. The trained DL network is capable of fast and accurate predicting the DRA geometry and performance. The DL network was used to design DRAs within a predefined frequency band, achieving an -10dB impedance bandwidth of over 1.7GHz. The trained neural network can predict both the impedance bandwidth and the dimensions of the DRA. Compared to traditional optimization methods, this approach significantly improves the efficiency of DRA design.

Original languageEnglish
Title of host publication2025 IEEE International Workshop on Electromagnetics
Subtitle of host publicationApplications and Student Innovation Competition, iWEM 2025
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages104-106
Number of pages3
ISBN (Electronic)9798331522407
DOIs
Publication statusPublished - 2025
Externally publishedYes
Event2025 IEEE International Workshop on Electromagnetics: Applications and Student Innovation Competition, iWEM 2025 - Hong Kong, Hong Kong
Duration: 4 Aug 20256 Aug 2025

Publication series

Name2025 IEEE International Workshop on Electromagnetics: Applications and Student Innovation Competition, iWEM 2025

Conference

Conference2025 IEEE International Workshop on Electromagnetics: Applications and Student Innovation Competition, iWEM 2025
Country/TerritoryHong Kong
CityHong Kong
Period4/08/256/08/25

Keywords

  • Deep Learning
  • Dielectric Resonator Antennas
  • Genetic Algorithm
  • Wideband

Fingerprint

Dive into the research topics of 'Design of a Wideband Rectangular Dielectric Resonator Antenna Using Deep Learning'. Together they form a unique fingerprint.

Cite this