高斯白噪声中单频复正弦信号频率估计新方法

A New Method for Frequency Estimation of Single Complex Sinusoid in White Gaussian Noise

提出了一种高质量的正弦频率估计新方法,该方法的主要思想是:先对信号自相关矩阵进行特征值分解(Eigenvaluedecompose简称EVD),找到最大特征值,尔后按一定规律构造一个非对称的相关阵,对其进行奇异值分解(Singularvaluedecompose简称SVD),选取最小奇异值对应的奇异矢量,利用类似于Pisarenko谱估计方法的计算公式直接进行频率估计。模拟实验表明,在低信噪比(SNR)、短时数据下,利用这种方法进行单频复正弦频率估计,其分辨率和统计稳定性明显优于常规的MUSIC方法。

This paper gives a new and better method for quality frequency estimation of single complex sinusoid. The method meaning: decompose autocorrelation matrix of signal in eigen value first to find the biggest eigen value, then reorder a new non-symmetrical correlation matrix according to a certain rule,decompose it in singular value select the smallest vector of singular value and directly estimate frequency by means of the computation formula similar to the Pisarenko method. The simulations experiments show that the new method is of high resolution and high statistical stability in lower SNR and shorter data as compared with the MUSIC method.