Abstract

In this chapter we present a control-oriented identification and (in)validation framework for a class of LPV systems. The identification step takes into account both the dependence of part of the model on time-varying parameters as well as the possible existence of a non-parametric component. The validation step (in)validates the obtained model subject to unstructured uncertainty. The main result of this chapter shows that the problems of checking consistency between the experimental data and the a priori assumptions and that of obtaining a nominal model, can be recast as Linear Matrix Inequality feasibility problems that can be efficiently solved. Moreover, the overall computational complexity is similar to that of obtaining and/or (in)validating LTI models of comparable size. These results are illustrated with a practical example arising in the context of active vision.

Citations