基于幅度比值和子空间的频率估计算法

Frequency estimation algorithm based on amplitude ratio and subspace

摘要:

为降低单频正弦复信号的频率估计误差，针对传统Rife算法在较低信噪比条件下幅值最大和次大谱线判断错误问题，利用RootMUSIC算法结果代替快速傅里叶变换(FFT)结果，确定幅值最大和次大谱线序号；同时针对Rife算法在部分频率区间估计误差较其他区间偏差大的问题，采用频移重估思路，进一步降低了整个频率估计区间的估计误差。仿真表明，在较低信噪比条件下，所提RM-Rife算法较Rife算法、Root MUSIC算法和M-Rife算法，具有更优的估计性能，可用于相关工程实践中。

Abstract:

In order to effectively reduce the frequency estimation error of single complex sinusoid, aiming at the estimation error of the traditional Rife algorithm caused by the judgment error of the maximum amplitude and sub maximum amplitude spectral line under the condition of low signal-to-noise ratio, replacing the Fast Fourier Transform(FFT) result with frequency estimation result of Root Multiple Signal Classification(RootMUSIC) algorithm is adopted to determine the maximum amplitude and sub maximum amplitude spectral line. In view of the problem that the estimation error of Rife algorithm in some frequency intervals is larger than that in other intervals, frequency shift and re-estimation method is employed to further reduce the estimation error in the whole frequency estimation range. Simulation results show that the proposed RootMUSIC-Modified-Rife(RM-Rife) algorithm has better estimation performance than Rife algorithm, RootMUSIC algorithm and Modified-Rife(M-Rife) algorithm under low SNR conditions, and can be applied to related engineering practice.