Recursive identification in the behavioral setting: an abstract

Summary form only given. We propose an approach to identification which is based on the assumption that an upper bound on the norm of the perturbation affecting the data is a priori known. As a consequence, an uncertain family of systems is obtained which under mild assumptions, is guaranteed to contain the system which generated the data. This family of systems is given in terms of a behavioral kernel representation. One important feature of this approach is that the family of systems can be easily updated in a recursive manner. The parametrization is suitable for applying robust control methods, for instance H_∞ control in the behavioral setting