基于FFT幅度和相位插值的频率估计改进算法

摘要

There is a inversion of the interpolation direction on the interpolated FFT algorithm in low SNR or the relative deviation be-tween the actual frequency and the estimated frequency which is small. Aimed at the characteristic,an conjoint algorithm is proposed which combined the advantages of the segmented-FFT phase difference method and the Rife method. The algorithm can solve the problem that the Rife algorithm has a large variance wrren estimate deviation δ approaches zero and the segmented-FFT phase difference algorithm has a large variance when δ is close to 0. 5 in low SNR. The results of simulation show that the improved algorithmin have a better accuracy and stability then the Rife algorithm and the segmented-FFT phase difference algorithm in low SNR conditions, So it has some practical value.%针对基于FFT的插值算法在信噪比较低或信号的实际频率与估计频率之间的相对偏差较小时,存在插值方向错误的情况,通过结合其中分段相位差法与Rife算法的优点提出了一种改进算法,该算法可以解决在低信噪比情况下当估计偏差δ接近零时,Rife算法估计误差较大、而当δ接近0.5时相位差法误差大的问题.算法仿真实验结果表明:改进算法在低信噪比条件下的频率估计精度高于Rife算法和分段相位差法,稳定性也比较高,具有很好的应用价值.