Applying Koopman Methods for Nonlinear Reachability Analysis

Master Thesis (2021)
Author(s)

T.H.J. Sweering (TU Delft - Mechanical Engineering)

Contributor(s)

Manuel Mazo Espinosa – Mentor (TU Delft - Team Manuel Mazo Jr)

Giannis Delimpaltadakis – Mentor (TU Delft - Team Manuel Mazo Jr)

A. Lukina – Graduation committee member (TU Delft - Algorithmics)

Faculty
Mechanical Engineering
Copyright
© 2021 Tim Sweering
More Info
expand_more
Publication Year
2021
Language
English
Copyright
© 2021 Tim Sweering
Graduation Date
05-10-2021
Awarding Institution
Delft University of Technology
Programme
['Mechanical Engineering | Systems and Control']
Faculty
Mechanical Engineering
Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Abstract

In this thesis we investigate the possibilities for applying Koopman methods for reachability analysis. Reachability analysis is a verification process used to determine that a dynamical system starting in an initial set X0 cannot reach a certain set of dangerous states D within a time interval [0,T]. Koopman methods seem promising, because they predict nonlinear behaviour using linear techniques. However they have not been widely applied to reachability analysis.

We describe three different Koopman methods: data-driven, Polyflow and Carleman. We use the Polyflow method combined with ideas from several other methods to create a new reachability tool: PolyReach. Next, we analyse the performance of PolyReach by comparing it with a state-of-the-art reachability algorithm Flow* on various nonlinear systems. Finally, we summarize the strengths and weaknesses of the PolyReach tool and discuss ideas for further improvement.

Files

License info not available