Pseudo-canonical forms, identifiable parametrizations and simple parameter estimation for linear multivariable systems: Parameter estimation (Q797522)

In a previous paper of the authors [ibid. 20, 429-442 (1984)] a parametrization of a matrix fraction description (MFD) of multivariable systems was derived from the p.c.o. parametrization of state space models. In this paper some estimators are constructed for the parameters in this MFD parametrization. They are modifications of three equation-error type methods: linear least squares, ''compensated'' estimator and instrumental variables. The last two estimators are shown to be consistent. Finally, two results are presented on the construction of parametrization- independent instrumental variables. For some of the proofs reference is made to the Ph.D. dissertation of \textit{G. O. Corrêa} [Cambridge University U.K. (1983)].