Title :
A weighted total least squares estimator for multivariable systems with nearly maximum likelihood properties
Author :
Guillaume, Patrick ; Pintelon, Rik ; Schoukens, Johan
Author_Institution :
Mech. Eng. Dept., Vrije Univ., Brussels, Belgium
fDate :
8/1/1998 12:00:00 AM
Abstract :
The aim of the present paper is to develop a parametric estimator for linear time-invariant multivariable systems with nearly maximum likelihood properties. The estimator is based on the total least squares (TLS) method. It can be seen as an “optimally” weighted iterative generalized TLS (GTLS) estimator, combining the nice asymptotic properties of the maximum likelihood (ML) method with the global minimization property of the GTLS estimator
Keywords :
iterative methods; least squares approximations; linear systems; maximum likelihood estimation; multivariable systems; singular value decomposition; asymptotic properties; global minimization property; iterative generalized TLS; linear time-invariant multivariable systems; nearly maximum likelihood properties; weighted total least squares estimator; Cost function; Covariance matrix; Equations; Frequency estimation; Least squares approximation; MIMO; Maximum likelihood estimation; Parameter estimation; Polynomials; Yield estimation;
Journal_Title :
Instrumentation and Measurement, IEEE Transactions on
