频率估计的多段差频正弦信号加权融合算法

针对多段差频正弦信号,提出一种基于加权融合的频率估计算法,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。为消除各段信号频率不等对频谱分析的影响,根据各段信号间频率差生成差频修正矩阵,对多段差频正弦信号频谱进行同频化处理;为消除各段信号相位不连续对频谱分析的影响,构造具有相位连续特性和降噪特性的加权因子,对同频化的多段差频正弦信号频谱进行加权融合,得到最优加权融合频谱;最后,谱峰搜索最优加权融合频谱,实现高精度频率估计.仿真实验表明,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.

A frequency estimation algorithm based on weight-fusion of multi-section sinusoids with known frequency difference

Based on weight-fusion of multi-section sinusoids with the known frequency-difference(MS-sinusoids-KFD),a frequency estimation algorithm was proposed to improve frequency estimation of the short sinusoid in low signal-to-noise ratio(SNR),and to extend the scope of application of the multi-section signal fusion method.Firstly,the frequency-difference modified matrix was created based on the known frequency-difference.To eliminate the influence caused by the known frequency-difference,spectra of MS-sinusoids-KFD were made the same as those of multi-section co-frequency-sinusoids by this matrix.Secondly,the weighted factor which can make phases continuous and decrease noise was constructed,and the optimization weighted-accumulation spectrum(OW-A spectrum) was gained through weighted-accumulating spectra of MS-Sinusoids-KFD by this weighted factor.Consequently,precise frequency estimation can be obtained through spectral peak searching of the OW-A spectrum.Simulation results show that compared with the current algorithms,the proposed one works better in precision,noise immunity and universality,and is particularly superior for short sinusoids in low SNR.