Recently developed subspace-based system identification (4SID) techniq
ues have opened new routes to the identification of multi-input multi-
output systems. The 4SID techniques guarantee convergence, and run fas
ter than the statistically efficient prediction error methods without
much performance loss. The resulting computational load of the 4SID te
chniques is O(NM(2)), where N is the data length and M is the sliding
window size. However, the computational burden O(NM(2)) can become pro
hibitively large as N and M grow large. Noting that the major bottlene
ck comes from the QR factorization of an M X N data matrix and that th
e existing 4SID techniques do not exploit the structure of the matrice
s arising in the identification procedure, we propose a new implementa
tion of the existing 4SID, which reduces the computational burden to O
(NM) by exploiting the displacement and low-rank structure of the matr
ices.