A behavioral approach to the controlled system identification problem

Abstract

Considers a behavioral approach to the problem of identifying a dynamical system that is known to be a component of a controlled system configuration. For suitable notions of complexity and accuracy, we characterize the class of plants of bounded complexity which, after interconnection with a given controller, achieve a guaranteed accuracy of the controlled system behavior. It is shown that the controlled systems that can be obtained in this way are parametrized in terms of the most powerful unfalsified model of reduced controlled system data. The key consequence of this result is that it translates the approximate controlled system identification problem to a realization problem.