Frequency Estimation of Discrete-Time Sinusoidal Signals With Time-Varying Amplitude

This brief focuses on the frequency estimation problem of a non-stationary discrete-time sinusoidal signal, where the amplitude is supposed to be a measurable function multiplied by an unavailable constant. This problem is successfully solved by the following two steps. The first step is to derive a linear regression model of the relative parameter, and the second step is to construct an estimation algorithm based on the standard gradient approach. It is shown that the proposed algorithm assures global exponential convergence of the frequency estimation and enjoys certain robustness. Moreover, it is shown that the proposed estimator is also suitable for incomplete measurement case. The efficiency of the proposed approach is demonstrated through several numerical simulations.