Title :
Generalized Arma Model for MIMO System Identification
Author :
Bingulac, S.P. ; Krtolica, R.
Author_Institution :
Electrical Engineering Department, Virginia Polytechnic Institute & State University, Blacksburg, VA 24061; University of Belgrade, Department of Organizational Sciences, Yugoslavia
Abstract :
It is shown that in deterministic least square identification conventional ARMA models, when characterized by a redundant number of parameters, lead to ill-conditioned or even singular matrices. A generalized ARMA model eliminating redundant parameter characterization and the subsequent ill-conditioning, is derived.
Keywords :
Art; Inspection; Least squares methods; MIMO; Mathematical model; Observability; Parameter estimation; System identification; ARMA models; Identification; least-square methods; multivariable systems; parametrization;
Conference_Titel :
American Control Conference, 1985
Conference_Location :
Boston, MA, USA
