基于多段差频正弦信号频谱相关的频率估计算法

A Frequency Estimation Algorithm Based on Spectrum Correlation of Multi-Section Sinusoids with Known Frequency-Difference

针对多段差频正弦信号提出一种基于频谱相关的频率估计算法,用以提高低信噪比条件下短时正弦信号的频率估计精度,扩展多段信号融合法的适用范围。首先,设计差频修正矩阵消除各段信号频率不等对频谱融合的影响,使同频化处理后的多段差频正弦信号频谱等同于多段同频正弦信号频谱;其次,构造加权因子消除各段信号相位不连续对频谱分析的影响,使得到的最优加权积累频谱近似于与多段差频正弦信号长度相等的相位连续信号频谱;然后,通过对最优加权积累频谱和多段差频正弦信号累加频谱的相关处理,抑制虚假谱峰和噪声干扰。最后,峰值搜索频谱相关谱,实现频率的精确估计。实验表明与现有算法相比,本文算法估计精度高、抗噪性好、普适性好,特别在低信噪比、短时时宽下性能优良。

Based on spectrum correlation of Multi-section sinusoids with Known Frequency-Difference （M-sinusoids- KFD）, a frequency estimation algorithm is proposed to improve frequency estimation accuracy of the short sinusoid in low SNR and to extend the applicable range of the muhi-section signal fusion method. To deal with the spectrum problems caused by the known frequency-difference and to make the spectrum of M-sinusoids-KFD as same as the multi-section co-frequency sinusoids, the frequency-difference correction matrix is created. Then, through weighted-accumulation spectrum of M- Sinusoids-KFD by the designed weighted factor, the gained optimization weighted-accumulation spectra become approximately the same as a coherent phase sinusoid. To suppress the false spectral peak and noises, the correlation spectrum is constructed by correlating the optimization weighted-accumulation spectrum and the accumulated spectrum of M-Sinusoids-KFD. Finally, accurate-frequency estimation can be obtained through spectral peak searching of the correlation spectrum. Simulation results demonstrate that the proposed algorithm works better in accuracy, noise immunity and universality, especially has a superior performance for the short sinusoid in low SNR compared with some present algorithms.