A sliding iterative scheme for frequency estimation of complex exponential sinusoids

The estimation of the frequency of a complex exponential sinusoid that is corrupted by random noise is fundamental in the domain of signal processing, as it has a variety of application in many areas. In this study, we presented a frequency estimator in the frequency domain of a complex signal. The estimator is composed of two steps: a standard maximum bin coarse search and a fine search. The fine search aims at searching for the frequency offset and is implemented by adopting a sliding iterative scheme, where the frequency offset is estimated based on three spectral points that are dynamically determined in the iterative procedure. Our approach requires only one calculation of Fourier coefficient in each iteration. The simulation results show the superiority of our algorithm, especially in the aspect of time complexity. It can serve as a promising alternative to other methods in the case of real-time application.