Estimation of the Instantaneous Periodic Parameters in the Frequency Domain

The paper presents effective algorithms for fast measurement and estimation of the unknown changing frequency and amplitude. The method for emulating coherent sampling with changing signal parameters by the interpolation of the DFT to estimate the frequency and amplitude of the particular component is described. An analysis of errors of the DFT coefficients caused by the frequency variation shows that the errors have two systematic contributions: the bias part due to the window spectrum main lobe and the long-range contribution. They can be both reduced with the interpolation. The optimum for reducing the time of measurement to two periods of the fundamental component and simultaneously reducing the influences of the harmonic components as possible can be the estimation with the two-point interpolation and the Manning window