Scholarlink

<p>Global navigation satellite system (GNSS) Doppler measurements are immune to cycle slips, providing a robust way to detect ionospheric irregularities. This study presents a novel approach to detect equatorial plasma bubble (EPB) using a support vector machine (SVM) algorithm based on the GNSS Doppler measurements. The input of the detector is the Doppler index (DI), which is extracted from the dual-frequency differential Doppler observations. Data from HKWS station located in Hong Kong during 2022 are employed to train the SVM model and validate its performance. The results show the trained SVM model achieves 96.7% validation accuracy of EPB detection. To assess the general capability of the model, EPB events throughout the entire year of 2023 are investigated at both the HKWS station and the HYDE station. The results show the performance of EPB detection by the SVM model using DI is comparable to that of by visually inspecting the total electron content time series based on GNSS carrier-phase measurements. In addition, the characteristics of EPB occurrence are also consistent to previous studies, suggesting the detection results are reliable.</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

Equatorial plasma bubble detection based on GNSS Doppler index using support vector machine algorithm

  • 期刊论文

  • 发表日期：

    2025年2月4日

查看全文
  • Fenkai Zhang,
  • Wu Chen,
  • Fan Xu,
  • Fang Zou,
  • Yuanfan Deng,
  • Long Tang

摘要

Global navigation satellite system (GNSS) Doppler measurements are immune to cycle slips, providing a robust way to detect ionospheric irregularities. This study presents a novel approach to detect equatorial plasma bubble (EPB) using a support vector machine (SVM) algorithm based on the GNSS Doppler measurements. The input of the detector is the Doppler index (DI), which is extracted from the dual-frequency differential Doppler observations. Data from HKWS station located in Hong Kong during 2022 are employed to train the SVM model and validate its performance. The results show the trained SVM model achieves 96.7% validation accuracy of EPB detection. To assess the general capability of the model, EPB events throughout the entire year of 2023 are investigated at both the HKWS station and the HYDE station. The results show the performance of EPB detection by the SVM model using DI is comparable to that of by visually inspecting the total electron content time series based on GNSS carrier-phase measurements. In addition, the characteristics of EPB occurrence are also consistent to previous studies, suggesting the detection results are reliable.