Multimodal Learning for Arcing Detection in Pantograph-Catenary Systems
Hao Dong Eleni Chatzi Olga Fink, [0.5em] Chair of Structural Mechanics & Monitoring, ETH Zürich · 2026
Read Multimodal Learning for Arcing Detection in Pantograph-Catenary Systems by Hao Dong Eleni Chatzi Olga Fink, [0.5em] Chair of Structural Mechanics & Monitoring, ETH Zürich (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-08792.txt.
Description
Arcing at the pantograph-catenary interface threatens rail power delivery, wearing contacts and disrupting service, yet arcs are brief, noisy, and hard to tell from other transients. Hao Dong and two co-authors propose a multimodal framework combining high-resolution images with force measurements for robust arcing detection. Their arXiv paper builds two synchronized visual-force datasets.
One dataset comes from Swiss Federal Railways data, the other from public arcing videos plus synthetic force data mimicking real characteristics. MultiDeepSAD extends a deep anomaly-detection method across modalities with a new loss, aided by tailored pseudo-anomaly generation per data type. Tests show clear wins over baselines, even under domain shift and scarce real arcs. Readers get a practical detector for a elusive rail hazard.
Catalog record
| CREATOR | Hao Dong Eleni Chatzi Olga Fink, [0.5em] Chair of Structural Mechanics & Monitoring, ETH Zürich |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-08792