Accurate Frequency Estimation by Using Three-Point Interpolated Discrete Fourier Transform Based on Rectangular Window

Windowed interpolated discrete Fourier transform (DFT) has been well-studied to deal with the spectral leakage and picket fence effects in frequency estimation. When the number of acquired sine-wave cycles is small, the long-range spectral leakage of the negative frequency is the main cause of low accuracy. In this article, a novel frequency estimator using three-point interpolated DFT based on rectangular window is proposed. The spectral superposition of positive and negative frequencies is adopted in the proposed interpolation algorithm to remove both short- and long-range spectral leakage. The influence of white noise on frequency estimation can be reduced because the adopted rectangular window has the narrowest mainlobe width or the smallest equivalent noise bandwidth among all windows. Simulation and experimental results confirm that the proposed method outperforms existing windowed interpolated DFT algorithms.