Representations and parameter identification of multi-output linear systems

This paper examines the implications of the hypothesis that a physical process can be modeled by a system of the form (C,A + KC,B) where (C,A) is an observable pair and K and B are parameters to be identified. The hypothesis leads directly to a linear equation relating K,B and a model´s state to quantities computable from measured data. The findings of [2-4] are extended by showing that K,B and a model´s state can be asymptotically estimated using a dynamic identifier. By relaxing the preceding hypothesis it is shown that these results also apply to more general classes of models.