论文部分内容阅读
现代互谱估计是抑制有色噪声的一个有效方法。本文首次对正弦信号的扩阶互相关函数矩阵的特征结构进行了深入的分析，并进而提出了有色噪声背景下正弦信号频率估计的互谱Ｐｉｓａｒｅｎｋｏ方法和ＭＵＳＩＣ方法。这两种方法的突出特点是几乎不需要任何关于有色噪声的先验信息，并可在很低信噪比条件下工作，因此特别适合于工程应用。仿真结果表明，在有色噪声和很低信噪比条件下，该方法的谱分辨率和谐估计的稳定性大大优于自谱ＳＶＤ方法。
Modern cross-spectrum estimation is an effective way to suppress colored noise. For the first time, this paper analyzes the eigen-structure of the extended cross-correlation matrix of sinusoidal signals, and then proposes a cross-spectrum Pisarenko method and a MUSIC method of sinusoidal signal frequency estimation in the background of colored noise. The salient features of both methods are that they require little prior information about colored noise and work at very low signal-to-noise conditions and are therefore particularly well-suited for engineering applications. The simulation results show that under the condition of colored noise and very low signal-to-noise ratio, the stability of the spectral resolution harmonic estimation of the proposed method is much better than the SVD method.