The validation of model sets on the basis of closed-loop feedback system generated data

Model validation provides a useful means of assessing the ability of a model to account for a specific experimental observation, and has application to modeling, identification and fault detection. In the robust control context this amounts to determining whether or not there exists an element of the model set that can account for the observed experimental datum. Efficient algorithms have been developed for model validation with frequency domain, discrete-time, and sampled-data frameworks. The focus of this paper is on the application of this theory to feedback systems. We outline these algorithms and discuss how they form part of a computationally based set of integrated tools for modeling, identification, and control system design.