[Zurück]


Vorträge und Posterpräsentationen (mit Tagungsband-Eintrag):

C. Hametner, S. Jakubek:
"Engine Model Identification Using Local Model Networks in Comparison with a Multilayer Perceptron Network";
Vortrag: International Symposium on Models and Modeling Methodologies in Science and Engineering: MMMse 2011, Orlando, Florida, USA; 27.03.2011 - 30.03.2011; in: "Proceedings of the International Symposium on Models and Modeling Methodologies in Science and Engineering in Science and Engineering (MMMSE 2011)", (2011), ISBN: 9781936338221.



Kurzfassung englisch:
This paper features a new nonlinear parameter optimisation
technique for a local model network and gives a
case based comparison to multilayer perceptron networks
in the context of a complex industrial application. The
architecture of a local model network as well as an
effective training algorithm which allows to build reliable
and interpretable models is presented. The performance
and suitability of both approaches for an automated model
generation scheme is illustrated and discussed by means
of real static and dynamic measurement data from an
internal combustion engine.

Schlagworte:
Nonlinear system identification, local model networks, multilayer perceptron network, interpretability

Erstellt aus der Publikationsdatenbank der Technischen Universitšt Wien.