A new class of uncertain linear systems stabilizable via output feedback

We consider physical systems represented by mathematical models affected by time-varying uncertainties. No statistical properties of these uncertainties are known, but bounds on their maximum excursions are assumed to be given. We study the problem of guaranteed stabilization via output feedback. It is shown that the inclusion of reduced-order observers in the feedback loop expands the class of stabilizable systems. The Constrained Lyapunov Problem and the Gain Selection Lemma are instrumental to our design methods.