Recursive regulator: a deep-learning and real-time model adaptation strategy for nonlinear systems

Abstract Adaptive modeling is imperative for analyzing nonlinear systems deployed in natural dynamic environments. It facilitates filtering, prediction, and automatic control of the target object in real time to respond to unpredictable and non-repetitive sudden physical impairment caused by ambient...

Full description

Saved in:
Bibliographic Details
Main Authors: Jinming Sun, Yanqiu Huang, Wanli Yu, Alberto Garcia-Ortiz
Format: Article
Language:English
Published: Nature Portfolio 2025-08-01
Series:Communications Engineering
Online Access:https://doi.org/10.1038/s44172-025-00477-4
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849342977376256000
author Jinming Sun
Yanqiu Huang
Wanli Yu
Alberto Garcia-Ortiz
author_facet Jinming Sun
Yanqiu Huang
Wanli Yu
Alberto Garcia-Ortiz
author_sort Jinming Sun
collection DOAJ
description Abstract Adaptive modeling is imperative for analyzing nonlinear systems deployed in natural dynamic environments. It facilitates filtering, prediction, and automatic control of the target object in real time to respond to unpredictable and non-repetitive sudden physical impairment caused by ambient impacts, such as corrosion, thermal drift, interference, etc. Existing nonlinear modeling approaches, however, are too complex for online training or fall short in rapid model recalibration under such conditions. To address this challenge, here we present a strategy that applies a regulator to the Koopman operator, enabling real-time model adaptation for nonlinear systems. In our approach, the regulator is directly implemented in nonlinear state-space without disrupting the pre-trained black-box predictor. The proposed technique demonstrates efficacy in capturing a broad spectrum of nonlinear dynamics and exhibits rapid adaptability to system changes without requiring offline retraining. Furthermore, its lightweight implementation and high-speed performance make it well-suited for embedded systems and applications demanding fast model recalibration and robustness.
format Article
id doaj-art-404f03447bec4b679f4a4c9cf77bca9b
institution Kabale University
issn 2731-3395
language English
publishDate 2025-08-01
publisher Nature Portfolio
record_format Article
series Communications Engineering
spelling doaj-art-404f03447bec4b679f4a4c9cf77bca9b2025-08-20T03:43:11ZengNature PortfolioCommunications Engineering2731-33952025-08-014111310.1038/s44172-025-00477-4Recursive regulator: a deep-learning and real-time model adaptation strategy for nonlinear systemsJinming Sun0Yanqiu Huang1Wanli Yu2Alberto Garcia-Ortiz3Institute of Electrodynamics and Microelectronics, University of BremenFaculty of Electrical Engineering, Mathematics and Computer Science, University of TwenteInstitute of Electrodynamics and Microelectronics, University of BremenInstitute of Electrodynamics and Microelectronics, University of BremenAbstract Adaptive modeling is imperative for analyzing nonlinear systems deployed in natural dynamic environments. It facilitates filtering, prediction, and automatic control of the target object in real time to respond to unpredictable and non-repetitive sudden physical impairment caused by ambient impacts, such as corrosion, thermal drift, interference, etc. Existing nonlinear modeling approaches, however, are too complex for online training or fall short in rapid model recalibration under such conditions. To address this challenge, here we present a strategy that applies a regulator to the Koopman operator, enabling real-time model adaptation for nonlinear systems. In our approach, the regulator is directly implemented in nonlinear state-space without disrupting the pre-trained black-box predictor. The proposed technique demonstrates efficacy in capturing a broad spectrum of nonlinear dynamics and exhibits rapid adaptability to system changes without requiring offline retraining. Furthermore, its lightweight implementation and high-speed performance make it well-suited for embedded systems and applications demanding fast model recalibration and robustness.https://doi.org/10.1038/s44172-025-00477-4
spellingShingle Jinming Sun
Yanqiu Huang
Wanli Yu
Alberto Garcia-Ortiz
Recursive regulator: a deep-learning and real-time model adaptation strategy for nonlinear systems
Communications Engineering
title Recursive regulator: a deep-learning and real-time model adaptation strategy for nonlinear systems
title_full Recursive regulator: a deep-learning and real-time model adaptation strategy for nonlinear systems
title_fullStr Recursive regulator: a deep-learning and real-time model adaptation strategy for nonlinear systems
title_full_unstemmed Recursive regulator: a deep-learning and real-time model adaptation strategy for nonlinear systems
title_short Recursive regulator: a deep-learning and real-time model adaptation strategy for nonlinear systems
title_sort recursive regulator a deep learning and real time model adaptation strategy for nonlinear systems
url https://doi.org/10.1038/s44172-025-00477-4
work_keys_str_mv AT jinmingsun recursiveregulatoradeeplearningandrealtimemodeladaptationstrategyfornonlinearsystems
AT yanqiuhuang recursiveregulatoradeeplearningandrealtimemodeladaptationstrategyfornonlinearsystems
AT wanliyu recursiveregulatoradeeplearningandrealtimemodeladaptationstrategyfornonlinearsystems
AT albertogarciaortiz recursiveregulatoradeeplearningandrealtimemodeladaptationstrategyfornonlinearsystems