Autoregressive spectral estimation in additive noise

Abstract

The estimation of the spectral density of a discrete-time stationary Gaussian autoregressive process AR (p) from a finite set of noise observations is considered. A modified spectral estimator based on the high-order Yule-Walker equations is considered. Joint asymptotic normality of this spectral estimator is established; a precise asymptotic expression for the covariance matrix of the limiting distribution is obtained. The special case of AR(1) plus noise is considered in some detail.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>