一种DFT的频率校正算法

DFT interpolation algorithm for frequency estimation

针对传统方法在工程实践中的频率估计精度不够的问题,提出了一种迭代频率估计方法.该方法具有较强的傅里叶系数的加权样本插值功能,可适用于几乎所有窗函数;对于各种窗函数的系统误差及在高斯白噪声下该算法表现出了比传统算法更优越的性能.实验结果表明该方法能显著减少错误引发的错误的谱线的位置,高斯噪声条件下进行频率估计时该方法取得了比传统方法更小的估计误差,表现出较强的估计精度与抗噪性能.

Aiming at the problem that the precision of frequency estimation is not enough in traditional engineering practice, an iterative frequency estimation method is proposed.The method has a quite effective function of weighted sample interpolation with the Fourier coefficient, and is applicable to almost all window functions.For a variety of window function system errors, this algorithm exhibited better performance than the traditional algorithm.Experimental results showed that this method was capable of significantly reducing the error caused location of the line.When the frequency estimation was performed under the conditions of Gaussian noise,this method obtained smaller estimation error than the traditional method, being impressed by higher estimation accuracy and anti-noise performance.