Scholarlink

<p>Frequency estimation of harmonics in complex noise is a classical problem in signal processing and has many applications, including radio astronomy, wireless communications, sonar, and radar. This paper presents a frequency estimation method for harmonics in complex noise using deep learning. A one-dimensional (1D) convolutional neural network is introduced to estimate the frequency of harmonic using the original sample data. The original sample data are divided into three channels: real part channel, imaginary part channel, and phase channel to obtain more input features. By adopting an on-grid approach, the frequency estimation problem is modeled as a multi-label classification task, and the 1D convolutional neural network is trained to estimate the frequencies using the original sample data. Simulation results verified the effectiveness and robustness of the proposed algorithm and proved its superiority over the conventional cyclic correlation algorithm.</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

Frequency estimation of harmonics in complex noise based on deep learning

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月27日

查看全文
  • Shiyong Yang,
  • Bo Zhang,
  • Jinlong Jiang

摘要

Frequency estimation of harmonics in complex noise is a classical problem in signal processing and has many applications, including radio astronomy, wireless communications, sonar, and radar. This paper presents a frequency estimation method for harmonics in complex noise using deep learning. A one-dimensional (1D) convolutional neural network is introduced to estimate the frequency of harmonic using the original sample data. The original sample data are divided into three channels: real part channel, imaginary part channel, and phase channel to obtain more input features. By adopting an on-grid approach, the frequency estimation problem is modeled as a multi-label classification task, and the 1D convolutional neural network is trained to estimate the frequencies using the original sample data. Simulation results verified the effectiveness and robustness of the proposed algorithm and proved its superiority over the conventional cyclic correlation algorithm.