Title :
Algorithm for extended least-squares model reduction
Author_Institution :
Centro de Pesquisa e Desenvolvimento em Engenharia Eletrica, Univ. Federal de Minas Gerais, Belo Horizonte, Brazil
fDate :
10/26/1995 12:00:00 AM
Abstract :
A new algorithm is presented to determine the numerator of a reduced-order model by means of least-squares (LS) techniques. The only requirement is that the simplified denominator should be previously determined. The new algorithm seems quite relevant in the light of a recent Letter by Smith and Lucas (see ibid., vol.31, no. 11, p. 929-930, 1995) in which it was shown that the LS estimation of numerator coefficients is not achieved using other LS procedures
Keywords :
estimation theory; least squares approximations; reduced order systems; LS estimation; least-squares model reduction; numerator coefficients; numerator determination algorithm; reduced-order model;
Journal_Title :
Electronics Letters
DOI :
10.1049/el:19951311
