Scholarlink

<p>Frequency estimation of complex exponential signals in low signal to noise ratio environments has attracted considerable attention in many fields. This paper proposes an asymptotically unbiased frequency estimation algorithm for accurate and robust extraction of sinusoidal frequency parameters under low SNR conditions. A frequency fitting model is constructed by using the magnitudes of the spectral peak and its adjacent interpolation points. By combining segmentation averaging with iterative refinement, the proposed method reduces the influence of fractional frequency offset and improves both estimation accuracy and noise robustness. Both theoretical analysis and experimental results demonstrate that the estimation variance of the proposed method can asymptotically approach the Cramer Rao lower bound, and the minimum ratio of the mean square error (MSE) to the Cramer Rao lower bound reaches 1.0073. Compared with existing interpolation based frequency estimation methods, the proposed method achieves higher estimation accuracy and stronger noise robustness while maintaining stable performance over different fractional frequency offset values, which makes it well suited for practical engineering applications.</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 Frequency-Independent Iterative DFT Interpolation Method for Robust Sinusoidal Frequency Estimation

  • 期刊论文

  • 发表日期：

    2026年5月24日

查看全文
  • Fei-Fei Feng,
  • Hai-Bo Tang,
  • Qiao Meng,
  • Cong-Yan Chen

摘要

Frequency estimation of complex exponential signals in low signal to noise ratio environments has attracted considerable attention in many fields. This paper proposes an asymptotically unbiased frequency estimation algorithm for accurate and robust extraction of sinusoidal frequency parameters under low SNR conditions. A frequency fitting model is constructed by using the magnitudes of the spectral peak and its adjacent interpolation points. By combining segmentation averaging with iterative refinement, the proposed method reduces the influence of fractional frequency offset and improves both estimation accuracy and noise robustness. Both theoretical analysis and experimental results demonstrate that the estimation variance of the proposed method can asymptotically approach the Cramer Rao lower bound, and the minimum ratio of the mean square error (MSE) to the Cramer Rao lower bound reaches 1.0073. Compared with existing interpolation based frequency estimation methods, the proposed method achieves higher estimation accuracy and stronger noise robustness while maintaining stable performance over different fractional frequency offset values, which makes it well suited for practical engineering applications.