Scholarlink

<p>The rise of wireless technologies, IoT devices, and next-generation communication systems has created a demand for efficient, wideband antennas supporting diverse applications. Traditional antenna design methods are time-intensive, requiring significant computational resources and expertise. To address these challenges, this work integrates machine learning (ML) techniques to optimize antenna design, enhancing performance and reducing development time. This study focuses on developing a UWB antenna and using ML models to predict its reflection coefficient (S<sub>11</sub>). Eight ML models, including Decision Tree, Random Forest, Artificial Neural Networks (ANN), K-Nearest Neighbors (KNN), Extra Trees, CatBoost, Gradient Boost, and XGBoost, were tested, with CatBoost achieving the highest accuracy of 85%. The proposed fractal-based antenna achieves a bandwidth of 11.87&#xa0;GHz, spanning 2.53&#xa0;GHz to 14.4&#xa0;GHz, demonstrating gains of 3.4 dBi, 2.6 dBi, 3.7 dBi, and 1.26 dBi at resonant frequencies of 3.6&#xa0;GHz, 6.32&#xa0;GHz, 7.32&#xa0;GHz, and 12.54&#xa0;GHz, respectively, with corresponding S<sub>11</sub> of −22.5&#xa0;dB, −23.14&#xa0;dB, −29.52&#xa0;dB, and −17.30&#xa0;dB. These frequencies align with critical 5G NR bands such as n7 (2.62–2.69&#xa0;GHz), n78 (3.3–3.8&#xa0;GHz), n79 (4.4–5.0&#xa0;GHz), n96 (5.925–7.125&#xa0;GHz), and n104 (13.0–14.0&#xa0;GHz), making the antenna suitable for 5G, Wi-Fi, IoT, radar, and satellite communication applications. This work demonstrates the potential of ML in antenna engineering, simplifying design and ensuring high performance for next-generation communication systems. The proposed ML-assisted UWB antenna offers a cost-effective, efficient solution in a rapidly evolving technological world.</p> Graphical abstract <p></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

Design and analysis of UWB antenna using machine learning for next-generation communications

  • 期刊论文

  • 发表日期：

    2025年4月28日

查看全文
  • Rachit Jain,
  • Vandana Vikas Thakare,
  • P. K. Singhal

摘要

The rise of wireless technologies, IoT devices, and next-generation communication systems has created a demand for efficient, wideband antennas supporting diverse applications. Traditional antenna design methods are time-intensive, requiring significant computational resources and expertise. To address these challenges, this work integrates machine learning (ML) techniques to optimize antenna design, enhancing performance and reducing development time. This study focuses on developing a UWB antenna and using ML models to predict its reflection coefficient (S11). Eight ML models, including Decision Tree, Random Forest, Artificial Neural Networks (ANN), K-Nearest Neighbors (KNN), Extra Trees, CatBoost, Gradient Boost, and XGBoost, were tested, with CatBoost achieving the highest accuracy of 85%. The proposed fractal-based antenna achieves a bandwidth of 11.87 GHz, spanning 2.53 GHz to 14.4 GHz, demonstrating gains of 3.4 dBi, 2.6 dBi, 3.7 dBi, and 1.26 dBi at resonant frequencies of 3.6 GHz, 6.32 GHz, 7.32 GHz, and 12.54 GHz, respectively, with corresponding S11 of −22.5 dB, −23.14 dB, −29.52 dB, and −17.30 dB. These frequencies align with critical 5G NR bands such as n7 (2.62–2.69 GHz), n78 (3.3–3.8 GHz), n79 (4.4–5.0 GHz), n96 (5.925–7.125 GHz), and n104 (13.0–14.0 GHz), making the antenna suitable for 5G, Wi-Fi, IoT, radar, and satellite communication applications. This work demonstrates the potential of ML in antenna engineering, simplifying design and ensuring high performance for next-generation communication systems. The proposed ML-assisted UWB antenna offers a cost-effective, efficient solution in a rapidly evolving technological world.

Graphical abstract