Identification of Multivariable Systems and Representation in Canonical and Pseudo-Canonical Form

摘要

The structural and parametric identification of stable, linear, time-invariant, time-discrete, and observable (or controllable) multivariable systems is treated. The structural identification is performed by methods using the determinants ratio and the singular value decomposition. The compensated least squares method, and the instrumental variable method are used to accomplish the parametric identification. The processes are identified in an Input/Output model and from there transformations to represent them in the canonical and pseudo-canonical form of the state space model are performed.