Fast subspace-based system identification: An instrumental variable approach (Q1898499)

A new fast implementation for a subspace-based system identification technique of \textit{M. Viberg}, \textit{B. Ottersten}, \textit{B. Wahlberg} and \textit{L. Ljung} [Performance of subspace based state-space model identification methods, in Proc. 12th IFAC World Congress (1993), Vol. 5, 63-66 (1994)] that requires substantially less flops and storage space is presented. The time and storage savings are achieved by the Hankel structure of the data matrix and the low-rank structure of the covariance matrix.