Gaussian Processes with Noisy Regression Inputs for Dynamical Systems

Publikation: Beitrag in Buch/Bericht/Sammelwerk/KonferenzbandAufsatz in KonferenzbandForschungPeer-Review

Autorschaft

Organisationseinheiten

Forschungs-netzwerk anzeigen

Details

OriginalspracheEnglisch
Titel des Sammelwerks2025 American Control Conference, ACC 2025
Herausgeber (Verlag)Institute of Electrical and Electronics Engineers Inc.
Seiten160-165
Seitenumfang6
ISBN (elektronisch)9798331569372
ISBN (Print)979-8-3503-6761-4
PublikationsstatusVeröffentlicht - 8 Juli 2025
Veranstaltung2025 American Control Conference, ACC 2025 - Sheraton Denver Downtown Hotel, Denver, USA / Vereinigte Staaten
Dauer: 8 Juli 202510 Juli 2025

Publikationsreihe

NameProceedings of the American Control Conference
ISSN (Print)0743-1619

Abstract

This paper is centered around the approximation of dynamical systems by means of Gaussian processes. To this end, trajectories of such systems must be collected to be used as training data. The measurements of these trajectories are typically noisy, which implies that both the regression inputs and outputs are corrupted by noise. However, most of the literature considers only noise in the regression outputs. In this paper, we show how to account for the noise in the regression inputs in an extended Gaussian process framework to approximate scalar and multidimensional systems. We demonstrate the potential of our framework by comparing it to different state-of-the-art methods in several simulation examples.

ASJC Scopus Sachgebiete

Zitieren

Gaussian Processes with Noisy Regression Inputs for Dynamical Systems. / Wolff, Tobias M.; Lopez, Victor G.; Muller, Matthias A.
2025 American Control Conference, ACC 2025. Institute of Electrical and Electronics Engineers Inc., 2025. S. 160-165 (Proceedings of the American Control Conference).

Publikation: Beitrag in Buch/Bericht/Sammelwerk/KonferenzbandAufsatz in KonferenzbandForschungPeer-Review

Wolff, TM, Lopez, VG & Muller, MA 2025, Gaussian Processes with Noisy Regression Inputs for Dynamical Systems. in 2025 American Control Conference, ACC 2025. Proceedings of the American Control Conference, Institute of Electrical and Electronics Engineers Inc., S. 160-165, 2025 American Control Conference, ACC 2025, Denver, Colorado, USA / Vereinigte Staaten, 8 Juli 2025. https://doi.org/10.23919/ACC63710.2025.11107496, https://doi.org/10.48550/arXiv.2408.08834
Wolff, T. M., Lopez, V. G., & Muller, M. A. (2025). Gaussian Processes with Noisy Regression Inputs for Dynamical Systems. In 2025 American Control Conference, ACC 2025 (S. 160-165). (Proceedings of the American Control Conference). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.23919/ACC63710.2025.11107496, https://doi.org/10.48550/arXiv.2408.08834
Wolff TM, Lopez VG, Muller MA. Gaussian Processes with Noisy Regression Inputs for Dynamical Systems. in 2025 American Control Conference, ACC 2025. Institute of Electrical and Electronics Engineers Inc. 2025. S. 160-165. (Proceedings of the American Control Conference). doi: 10.23919/ACC63710.2025.11107496, 10.48550/arXiv.2408.08834
Wolff, Tobias M. ; Lopez, Victor G. ; Muller, Matthias A. / Gaussian Processes with Noisy Regression Inputs for Dynamical Systems. 2025 American Control Conference, ACC 2025. Institute of Electrical and Electronics Engineers Inc., 2025. S. 160-165 (Proceedings of the American Control Conference).
Download
@inproceedings{2d4d7dc8f4874612ac488eb4399be00b,
title = "Gaussian Processes with Noisy Regression Inputs for Dynamical Systems",
abstract = "This paper is centered around the approximation of dynamical systems by means of Gaussian processes. To this end, trajectories of such systems must be collected to be used as training data. The measurements of these trajectories are typically noisy, which implies that both the regression inputs and outputs are corrupted by noise. However, most of the literature considers only noise in the regression outputs. In this paper, we show how to account for the noise in the regression inputs in an extended Gaussian process framework to approximate scalar and multidimensional systems. We demonstrate the potential of our framework by comparing it to different state-of-the-art methods in several simulation examples.",
author = "Wolff, {Tobias M.} and Lopez, {Victor G.} and Muller, {Matthias A.}",
note = "Publisher Copyright: {\textcopyright} 2025 AACC.; 2025 American Control Conference, ACC 2025, ACC 2025 ; Conference date: 08-07-2025 Through 10-07-2025",
year = "2025",
month = jul,
day = "8",
doi = "10.23919/ACC63710.2025.11107496",
language = "English",
isbn = "979-8-3503-6761-4",
series = "Proceedings of the American Control Conference",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "160--165",
booktitle = "2025 American Control Conference, ACC 2025",
address = "United States",

}

Download

TY - GEN

T1 - Gaussian Processes with Noisy Regression Inputs for Dynamical Systems

AU - Wolff, Tobias M.

AU - Lopez, Victor G.

AU - Muller, Matthias A.

N1 - Publisher Copyright: © 2025 AACC.

PY - 2025/7/8

Y1 - 2025/7/8

N2 - This paper is centered around the approximation of dynamical systems by means of Gaussian processes. To this end, trajectories of such systems must be collected to be used as training data. The measurements of these trajectories are typically noisy, which implies that both the regression inputs and outputs are corrupted by noise. However, most of the literature considers only noise in the regression outputs. In this paper, we show how to account for the noise in the regression inputs in an extended Gaussian process framework to approximate scalar and multidimensional systems. We demonstrate the potential of our framework by comparing it to different state-of-the-art methods in several simulation examples.

AB - This paper is centered around the approximation of dynamical systems by means of Gaussian processes. To this end, trajectories of such systems must be collected to be used as training data. The measurements of these trajectories are typically noisy, which implies that both the regression inputs and outputs are corrupted by noise. However, most of the literature considers only noise in the regression outputs. In this paper, we show how to account for the noise in the regression inputs in an extended Gaussian process framework to approximate scalar and multidimensional systems. We demonstrate the potential of our framework by comparing it to different state-of-the-art methods in several simulation examples.

UR - http://www.scopus.com/inward/record.url?scp=105015840209&partnerID=8YFLogxK

U2 - 10.23919/ACC63710.2025.11107496

DO - 10.23919/ACC63710.2025.11107496

M3 - Conference contribution

AN - SCOPUS:105015840209

SN - 979-8-3503-6761-4

T3 - Proceedings of the American Control Conference

SP - 160

EP - 165

BT - 2025 American Control Conference, ACC 2025

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2025 American Control Conference, ACC 2025

Y2 - 8 July 2025 through 10 July 2025

ER -

Von denselben Autoren