基于分段谱平均的高精度频率估计方法

提出了一种适用于高斯白噪声背景下的正弦波频率估计的新算法。通过将采样信号分成多段,搜索其平均幅度谱最大位置对应的频率从而完成粗略估计。而累积谱的方法有效地提高了信号的信噪比,使该算法的信噪比阈值降低。采用加权的方法提取出FFT相位中包含的信号频率与最大谱线位置之间的相对偏差,进行二次精估计。性能仿真和比较结果验证了该算法的可行性。

High-Accuracy Frequency Estimator Based on Spectrum Segmentation

A new frequency estimation algorithm of single-tone in the presence of white Gaussian noise background is proposed.The coarse estimate is implemented by segmenting the signal into several parts and searching for the maximum of mean amplitude spectrum which is corresponding to the expected frequency.The average of the cumulative spectrum improves the SNR,thus lowering the algorithm's SNR threshold.Thereafter a precise estimation is prosecuted using phase weighting,where the phase of FFT reflects the offset of signal frequency from the maximum spectrum line.Performance simulation and comparison results prove the feasibility of the algorithm.