A Low Computational Complexity Frequency Estimation Method with High Precision of Sinusoid Based on DFT

In this paper, a new synthetic frequency estimator of sinusoid signal based on discrete Fourier transform (DFT) amplitudes is proposed. Compared with the existing methods, the proposed method has the advantages of low computational complexity and high precision. The method has two main stages. Firstly, N-point sampled data in the time domain DFT is performed and the coarse estimation is completed by searching the location of the spectrum line with maximum amplitude. Secondly, the synthetic fine estimation is performed by utilizing both the two samples values of Discrete-Time Fourier Transform (DTFT) of the original signal and the maximum amplitude spectrum line and its right or left neighbor. To verify the proposed method, simulation is conducted. The result shows that the root mean square error (RMSE) of frequency estimation of proposed synthetic method is close to the Cramer-Rao Lower Bound, and it outperforms the Candan method. This method is also comparable to Fang method and RCTSL method with zero-padding when conducted under medium and high SNR condition. Moreover, without zeropadding, the proposed method has much lower computational complexity than other known estimators with zero-padding.