An efficient algorithm for frequency estimation from cosine-sum windowed DFT coefficients

•A numerically efficient method for frequency estimation by Interpolation on DFT coefficients.•The use of generic cosine-sum windows for interference suppression.•Performance that nearly achieves the Cramer–Rao bounds. This paper presents an algorithm for estimating frequency of a complex sinusoid from three DFT coefficients. In particular, we consider the case where the signal is multiplied by a generic cosine-sum window for interference suppression. The algorithm is an interpolator that uses the peak sample in the DFT of the data and its two neighbors. The interpolator is based on a numerically efficient procedure which guarantees convergence to a solution. Performance of the proposed interpolator is shown to be nearly optimum and comparable to the best interpolator in the literature, while the computational burden required for the proposed algorithm is much lower than the generic numerical methods available. Degradation of performance due to interference is also investigated.