EOS
> Zum Inhalt
2016-11-07 [

R. Kucera

]

Gastvortrag: Professor Paul van den Hof

Wir laden herzlichst alle Interessenten zum Gastvortrag von Herrn Professor Paul van den Hof ein.

Datum/Uhrzeit: 21.11.2016, 11:00 - 12:00 Uhr
Ort: Freihaus, DA grün 06A
Vortragender: Paul van den Hof, TU Eindhoven (Niederlande)
https://www.tue.nl/en/university/departments/electrical-engineering/department/staff/detail/ep/e/d/ep-uid/19750719/

Vortrag: Identifiability of dynamic networks with noisy and noise-free nodes

Abstract:

Dynamic networks are structured interconnections of dynamical systems (modules) driven by external excitation and disturbance signals. In order to identify their dynamical properties and/or their topology consistently from measured data, we need to make sure that the network model set is identifiable. We introduce the notion of network identifiability, as a property of a parameterized model set, that ensures that module dynamics are uniquely related to the network objects that can be uniquely estimated from data. In the classical prediction error framework, these objects are the predictor filters that constitute the one-step ahead output predictors. We generalize this situation to include the option of having noise-free node signals.

 

The results can be used to specify which presence of excitation signals will result in a unique representation of the network dynamics in a particular network model parametrization. We combine aspects of the classical notion of system identifiability with a uniqueness-oriented parametrization concept, and extend this to the situation of highly structured model sets. All node signals in the network are treated in a symmetric way as both inputs and outputs. The presented concepts and theory allow for the incorporation of particular structural prior knowledge of the network structure.