Frequency estimation by two- or three-point interpolated Fourier algorithms based on cosine windows

This paper investigates the frequency estimation of a complex sinusoid obtained by interpolating two or three samples of the Discrete-Time Fourier Transform (DTFT) of a signal weighted by a suitable cosine window. Versions of the algorithms based on both DTFT complex values or modules are considered and the expressions for the related frequency estimators are provided. Iterative procedures are used in order to minimize the estimator variance due to additive wideband noise. Furthermore, the accuracies of the proposed estimators are verified and compared each other and with state-of-the art algorithms by means of computer simulations when applied to noisy or noisy and harmonically distorted complex sinusoids, respectively.