Title :
Finite record effects of the Errors-In-Variables estimator for linear dynamic systems
Author :
Barbe, K. ; Pintelon, Rik ; Vandersteen, Gerd
Author_Institution :
Dept. ELEC, Vrije Univ. Brussel, Brussels, Belgium
Abstract :
The frequency domain errors-in-variables (EIV) estimator for linear dynamic systems is generally formulated as a weighted least square estimator, where the weights used are the estimated noise (co)variances. In this paper, we study the finite record influences on the uncertainty of the EIV estimator for using the estimated weights. Furthermore, we discuss when a significant influence can occur and derive a practical method to circumvent this problem without the need of introducing a parametric noise model.
Keywords :
covariance analysis; estimation theory; frequency-domain analysis; least squares approximations; linear systems; signal sampling; errors-in-variables estimator; finite record effect; frequency domain EIV estimator; linear dynamic system; noise covariance estimation; signal sampling; weighted least square estimator; Cost function; Discrete Fourier transforms; Frequency domain analysis; Frequency estimation; Frequency measurement; Gaussian noise; Least squares approximation; Least squares methods; Transfer functions; Uncertainty; Errors-In-Variables; Frequency domain; Leakage; Linear dynamic systems; Weighted Least Squares; windowing;
Conference_Titel :
Instrumentation and Measurement Technology Conference, 2009. I2MTC '09. IEEE
Conference_Location :
Singapore
Print_ISBN :
978-1-4244-3352-0
DOI :
10.1109/IMTC.2009.5168697
