Spectral estimation of quasi-periodic data

Abstract

A computationally efficient algorithm for estimating the autocorrelation function and/or the power spectrum of periodic/quasi-periodic time series, when information about its period is available, is described. The autocorrelation matrix obtained from the sample covariance matrix using this method is a good approximation of the maximum-likelihood estimate of the autocorrelation matrix. Simulation results involving sinusoids in noise and speech data are presented. >