Maximum Likelihood Estimation of the Parameters of Linear Systems

Abstract

A method is presented which estimates the parameters of Linear Systems (LS), modelled by their transfer function, using a very efficient iteration algorithm. The estimator is an error in variables method and takes into account the noise on the input and output measurements. During the estimation process, an approximation of the Cramer-Rao lower bound on the covariance matrix of the estimates is derived and the 'mean' model error is discussed.