Abstract |
: |
Two recent approaches 4, 14 in subspace identification problems require the computation of the R factor of the QR factorization of a blockHankel matrix H,which, in general has a huge number of rows. Since the data are perturbed by noise, the involved matrix H is, in general, full rank. It is well known that, from a theoretical point of view, the R, factorization of is equivalent to the Cholesky factor of the correlation , apart from a multiplication by a sign ma- trix. In 12 a fast Cholesky factorization of the correla- tion matrix, exploiting the blockHankel structure of is described. In this paper we consider a fast algorithm to compute the factor based on the generalized Schur algorithm. The proposed algorithm allows to handle the rankdeficient case. |