Abstract
In the identification of linear systems the aim is to estimate the impulse response to within a given tolerance based on a finite number of noisy observations of the output. Whether this is possible depends upon the model set, that is, the set of impulse responses to which that of the system is assumed to belong. We give conditions on the model set which ensure that such identification is possible and also briefly review recent results concerning the complexity of identification, that is, the minimum number of required output samples.
Information