Input-output identifiability of continuous-time linear systems (Q1599201)

The identifiability, discussed within the framework of deterministic identification, of continuous-time systems is described by the input-output equation \(y= h^* u+\eta\), where the disturbance term \(\eta\) is sufficiently small and the unknown model \(h\) is a priori assumed as belonging to a prescribed model set. The study uses both the time-domain \(L^p\) norms and the frequency-domain \(H^p\) norms, \(1\leq p\leq\infty\). The main result of the paper consists in a necessary and sufficient condition for a model set to be identifiable.