一种基于窄带自相关的实信号频率估计算法

Frequency estimation based on autocorrelation function of the real sinusoid embedded in white Gaussian noise has been received extensive attention, because the sample autocorrelation sequence has the same frequency as the original signal, but with less noise effect. The method of the frequency estimation based on autocorrelation function can be divided into two categories: the time-domain and frequency-domain methods. The former are relatively simple while the latter have better estimation performance. In this paper, a narrow-band antocorrelation-based method for real sinusoid frequency estimation is proposed, by combining the characteristics of time domain with frequency domain antocorrelation. Firstly, spectral peak searching in frequency domain is applied to provide a coarse frequency estimate. Then, the narrow-band power spectrum of the signal is used to calculate the autocorrelation, from which the fine frequency estimate is obtained based on the simple modified covariance (MC) method. Finally, a closed-form frequency estimate is derived. Simulation results show that the performance of the proposed algorithm, when compared with several existing time-domain autocorrelation-based estimators and frequency-domain antocorrelation-based estimators, is closer to the Cramer-Rao Bound (CRB) when the signal to noise ratio(SNR) is low.