Scholarlink

<p>In this article, an innovative optimization design method for ultra-wideband (UWB) antennas that leverages machine learning models and intelligent optimization algorithms is presented. The method aims to improve the efficiency and reduce the cost of antenna design. By incorporating U-shaped slots and rectangular slot resonators, the UWB antenna achieves dual-notch frequency bands. This method is based on a machine learning model to establish the relationship between structural parameters and performance parameters, and then uses a method to create a dataset based on prior knowledge. Comparing the performance of three machine learning models, Gaussian process regression, support vector machine regression, and BP neural network, the (BP) neural network model and genetic algorithm are finally adopted for the optimization of the geometric structure of ultra-wideband antennas. The optimized antenna demonstrates dual-band suppression capabilities at center frequencies of 3.9&#xa0;GHz and 5.4&#xa0;GHz, effectively mitigating interference in the satellite communication C-band and the WLAN band, respectively. The effectiveness of the optimization strategy combining machine learning with a genetic algorithm was validated through full-wave simulations.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A machine learning and genetic algorithm-based notch band design method for ultra-wideband antennas

  • 期刊论文

  • 发表日期：

    2025年4月30日

查看全文
  • Chengjie Li,
  • Ruxin Zheng,
  • Yaogen Li,
  • Yixing Gu,
  • Mingjie Sheng,
  • Shiping Tang

摘要

In this article, an innovative optimization design method for ultra-wideband (UWB) antennas that leverages machine learning models and intelligent optimization algorithms is presented. The method aims to improve the efficiency and reduce the cost of antenna design. By incorporating U-shaped slots and rectangular slot resonators, the UWB antenna achieves dual-notch frequency bands. This method is based on a machine learning model to establish the relationship between structural parameters and performance parameters, and then uses a method to create a dataset based on prior knowledge. Comparing the performance of three machine learning models, Gaussian process regression, support vector machine regression, and BP neural network, the (BP) neural network model and genetic algorithm are finally adopted for the optimization of the geometric structure of ultra-wideband antennas. The optimized antenna demonstrates dual-band suppression capabilities at center frequencies of 3.9 GHz and 5.4 GHz, effectively mitigating interference in the satellite communication C-band and the WLAN band, respectively. The effectiveness of the optimization strategy combining machine learning with a genetic algorithm was validated through full-wave simulations.