Skip to main navigation Skip to search Skip to main content

Drift-Resistant Hydrogen Sensing via Domain Adaptation Graph Neural Networks

  • Haifeng Se
  • , Kai Song*
  • , Bo Wang
  • , Lu Xia
  • *Corresponding author for this work
  • School of Electrical Engineering and Automation, Harbin Institute of Technology
  • China Electronics Technology Group Corporation

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

To mitigate the impact of sensor drift in hydrogen detection exposed to complex environmental interferences, this study proposes a Domain Adaptation Graph Neural Network (DAGNN). By leveraging Graph Attention Networks (GAT) to model nonlinear relationships among sensors and employing Representation Subspace Distance (RSD) for dynamic domain alignment, DAGNN effectively compensates for drift induced by environmental disturbances. Experiments were conducted using an array of five hydrogen sensors under 32 environmental conditions, yielding 960 samples. Results demonstrate that DAGNN achieves a MRE of 4.0% and a R2 of 0.99, outperforming existing methods. These findings underscore the potential of DAGNN in enhancing the robustness and precision of hydrogen sensing, offering valuable contributions to hydrogen safety monitoring technologies.

Original languageEnglish
Title of host publicationThe Proceedings of the 20th Annual Conference of China Electrotechnical Society - Volume I
EditorsQingxin Yang, Dianguo Xu, Xuerong Ye, Qiuyue Nie, Yueshi Guan
PublisherSpringer Science and Business Media Deutschland GmbH
Pages252-259
Number of pages8
ISBN (Print)9789819590421
DOIs
StatePublished - 2026
Externally publishedYes
Event20th Annual Conference of China Electrotechnical Society, ACCES 2025 - Harbin, China
Duration: 19 Sep 202521 Sep 2025

Publication series

NameLecture Notes in Electrical Engineering
Volume1558 LNEE
ISSN (Print)1876-1100
ISSN (Electronic)1876-1119

Conference

Conference20th Annual Conference of China Electrotechnical Society, ACCES 2025
Country/TerritoryChina
CityHarbin
Period19/09/2521/09/25

Keywords

  • Domain Alignment
  • GAT
  • RSD
  • Sensor Drift

Fingerprint

Dive into the research topics of 'Drift-Resistant Hydrogen Sensing via Domain Adaptation Graph Neural Networks'. Together they form a unique fingerprint.

Cite this