PARAMETER AND ORDER ESTIMATION IN A CLASS OF MULTIVARIATE STOCHASTIC SYSTEMS.

The problem of estimating the parameters and order of a class of multivariate systems is treated. The systems considered are described by a stochastic time invariant linear difference equation. The so-called canonical form III is introduced as a possible unique representation of the system. It is shown that by using this canonical form, the computational effort compared with other canonical forms can be reduced. Further, it is shown that pole-zero cancellation, which is one of the methods used in identifying the order of single input-single output systems, can be extended to multivariate systems in canonical form III.