Frequency, amplitude, and phase tracking of nonsinusoidal signal in noise with extended Kalman filter

Abstract

This thesis applies extended Kalman filtering to the problem of estimating frequency, amplitude, and phase of a nonsinusoidal periodic signal contaminated by additive white, Gaussian noise. Parameters will be estimated up to mth significant harmonic component. It also gives an approach for the case of less than mth significant harmonic components. The estimator will track the signal's fundamental frequency, amplitudes, and phases while these parameters are changing slowly over time. The amplitudes are estimated as if the fundamental frequency estimate is correct; the frequency and the phases of the signal are estimated as if the amplitude estimation is correct. This thesis also contains tracking and the capture behavior of the filter.