In linear control system design we have to decide between exact (mathematical) reducibility on one side and compensation on the other side. The mathematical cancellation has to be executed, what leads to a reduced model. In other case, in reality not existing parts would remain in the model. For multi-input multi-output systems the reducibility problem is complicated, especially in case of approximated models due to measuring or rounding errors, cancellation has to be enforced. The paper introduces the new concepts of polynomial subordination and dominance, which are shown to be useful tools for such kind of problems. Moreover, they provide for a deeper insight into the structure of multi-input multi-output sampled-data systems, and this fact is demonstrated by an example.
Print ISSN: 0178-2312
Volume: 53, 09/2005
Pages: 434 - 444