Koopman-based lifting techniques for nonlinear systems identification

Alexandre Mauroy, Jorge Goncalves

Research output: Contribution to journalArticle

Abstract

We develop a novel lifting technique for nonlinear system identification based on the framework of the Koopman operator. The key idea is to identify the linear (infinite-dimensional) Koopman operator in the lifted space of observables, instead of identifying the nonlinear system in the state space, a process which results in a linear method for nonlinear systems identification. The proposed lifting technique is an indirect method that does not require to compute time derivatives and is therefore well-suited to low-sampling rate datasets.
Considering different finite-dimensional subspaces to approximate and identify the Koopman operator, we propose two numerical schemes: a main method and a dual method. The main method is a parametric identification technique that can accurately reconstruct the vector field of a broad class of systems. The dual method provides estimates of the vector field at the data points and is well-suited to identify high-dimensional systems with small datasets. The present paper describes the two methods, provides theoretical convergence results, and illustrates the lifting techniques with several examples.
LanguageEnglish
Number of pages16
JournalIEEE Transactions on Automatic Control
Publication statusSubmitted - 2018

Fingerprint

Nonlinear systems
Identification (control systems)
Sampling
Derivatives

Cite this

Mauroy, A., & Goncalves, J. (2018). Koopman-based lifting techniques for nonlinear systems identification. Manuscript submitted for publication.
@article{42b7ad16f7644b0f8024601f1c94c67e,
title = "Koopman-based lifting techniques for nonlinear systems identification",
abstract = "We develop a novel lifting technique for nonlinear system identification based on the framework of the Koopman operator. The key idea is to identify the linear (infinite-dimensional) Koopman operator in the lifted space of observables, instead of identifying the nonlinear system in the state space, a process which results in a linear method for nonlinear systems identification. The proposed lifting technique is an indirect method that does not require to compute time derivatives and is therefore well-suited to low-sampling rate datasets.Considering different finite-dimensional subspaces to approximate and identify the Koopman operator, we propose two numerical schemes: a main method and a dual method. The main method is a parametric identification technique that can accurately reconstruct the vector field of a broad class of systems. The dual method provides estimates of the vector field at the data points and is well-suited to identify high-dimensional systems with small datasets. The present paper describes the two methods, provides theoretical convergence results, and illustrates the lifting techniques with several examples.",
author = "Alexandre Mauroy and Jorge Goncalves",
year = "2018",
language = "English",
journal = "IEEE Transactions on Automatic Control",
issn = "0018-9286",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

Koopman-based lifting techniques for nonlinear systems identification. / Mauroy, Alexandre; Goncalves, Jorge.

In: IEEE Transactions on Automatic Control, 2018.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Koopman-based lifting techniques for nonlinear systems identification

AU - Mauroy, Alexandre

AU - Goncalves, Jorge

PY - 2018

Y1 - 2018

N2 - We develop a novel lifting technique for nonlinear system identification based on the framework of the Koopman operator. The key idea is to identify the linear (infinite-dimensional) Koopman operator in the lifted space of observables, instead of identifying the nonlinear system in the state space, a process which results in a linear method for nonlinear systems identification. The proposed lifting technique is an indirect method that does not require to compute time derivatives and is therefore well-suited to low-sampling rate datasets.Considering different finite-dimensional subspaces to approximate and identify the Koopman operator, we propose two numerical schemes: a main method and a dual method. The main method is a parametric identification technique that can accurately reconstruct the vector field of a broad class of systems. The dual method provides estimates of the vector field at the data points and is well-suited to identify high-dimensional systems with small datasets. The present paper describes the two methods, provides theoretical convergence results, and illustrates the lifting techniques with several examples.

AB - We develop a novel lifting technique for nonlinear system identification based on the framework of the Koopman operator. The key idea is to identify the linear (infinite-dimensional) Koopman operator in the lifted space of observables, instead of identifying the nonlinear system in the state space, a process which results in a linear method for nonlinear systems identification. The proposed lifting technique is an indirect method that does not require to compute time derivatives and is therefore well-suited to low-sampling rate datasets.Considering different finite-dimensional subspaces to approximate and identify the Koopman operator, we propose two numerical schemes: a main method and a dual method. The main method is a parametric identification technique that can accurately reconstruct the vector field of a broad class of systems. The dual method provides estimates of the vector field at the data points and is well-suited to identify high-dimensional systems with small datasets. The present paper describes the two methods, provides theoretical convergence results, and illustrates the lifting techniques with several examples.

UR - https://arxiv.org/abs/1709.02003

M3 - Article

JO - IEEE Transactions on Automatic Control

T2 - IEEE Transactions on Automatic Control

JF - IEEE Transactions on Automatic Control

SN - 0018-9286

ER -