Aarhus University Seal / Aarhus Universitets segl

Ahmad Madary

A Bayesian Framework for Large-Scale Identification of Nonlinear Hybrid Systems

Publikation: KonferencebidragPaperForskningpeer review

  • Ahmad Madary
  • Hamid Reza Momeni, The Faculty of Electrical and Computer Engineering (ECE), Tarbiat Modares University, Iran
  • Alessandro Abate, University of Oxford, Storbritannien
  • Kim Guldstrand Larsen, Department of Mathematics and Computer Science, Aalborg University, Danmark
In this paper, a two-level Bayesian framework is proposed
for the identification of nonlinear hybrid systems from
large data sets by embedding it in a four-stage procedure.
At the first stage, feature vector selection techniques are
used to generate a reduced-size set from the given training
data set. The resulting data set then is used to identify
the hybrid system using a Bayesian method, where the
objective is to assign each data point to a corresponding
sub-mode of the hybrid model. At the next (third) stage,
this data assignment is used to train a Bayesian classifier
to separate the original data set and determine the
corresponding sub-mode for all the original data points.
Finally, once every data point is assigned to a sub-mode, a
Bayesian estimator is used to estimate a regressor for each
sub-system independently. This Bayesian approach finds a
compromise between model complexity and accuracy. The
proposed method tested on three case studies. First, the
identification based on the full data set is compared with
that based on the reduced data set, over a nonlinear hybrid
system. Then, the proposed approach is compared with an
existing method from the literature, with improved
performance over estimation error and mode assignment.
Finally, the performance of the new method is tested by
identifying a complex hybrid model.
StatusAccepteret/In press - 2021
Begivenhed7th IFAC Conference on Analysis and Design of Hybrid Systems - UClouvain, BRUSSELS, Belgien
Varighed: 7 jul. 20219 jul. 2021


Konference7th IFAC Conference on Analysis and Design of Hybrid Systems

Se relationer på Aarhus Universitet Citationsformater

ID: 211785968