GNSS Interference Monitoring: Resilience of Machine Learning Methods on Public Real-World Datasets

Heublein L, Raichur NL, Feigl T, Brieger T, Heuer F, Asbach L, Ruegamer A, Ott F (2024)

Publication Language: English

Publication Type: Journal article, Original article

Publication year: 2024

Journal

Navigation, Journal of the Institute of Navigation Institute of Navigation

Book Volume: 7

Pages Range: 1-20

Journal Issue: 5

Abstract

The accuracy of vehicle localization is critical for applications like self-driving cars, toll systems, and

digital tachographs. Vehicles rely on GNSS receivers for positioning, but GNSS signals can be disrupted

by interference, necessitating its identification, classification, purpose determination, and localization

to mitigate effects. Machine learning (ML) methods have shown promise in interference monitoring,

though their real-world applicability remains uncertain. These methods require realistic, labeled datasets

that reflect operational noise and multipath effects, but creating such datasets is challenging due to

legal restrictions on causing GNSS interference. To address this, we conducted large-scale measurement

campaigns in real-world settings (highways in Germany and the Seetal Alps in Austria) and controlled

indoor environments. We evaluated supervised ML methods, explored pseudo-labeling for unsupervised

learning, and investigated challenges like dataset discrepancies, outlier detection, domain adaptation, and

data augmentation. This study highlights the potential of ML models to adapt to varying datasets and

reports on their performance in practical scenarios.

Authors with CRIS profile

Tobias Feigl Lehrstuhl für Informatik 2 (Programmiersysteme)

Involved external institutions

Fraunhofer-Institut für Integrierte Schaltungen (IIS) DE Germany (DE) Deutsches Zentrum für Luft- und Raumfahrt e.V. (DLR) / German Aerospace Center DE Germany (DE)

How to cite

APA:

Heublein, L., Raichur, N.L., Feigl, T., Brieger, T., Heuer, F., Asbach, L.,... Ott, F. (2024). GNSS Interference Monitoring: Resilience of Machine Learning Methods on Public Real-World Datasets. Navigation, Journal of the Institute of Navigation, 7(5), 1-20.

MLA:

Heublein, Lucas, et al. "GNSS Interference Monitoring: Resilience of Machine Learning Methods on Public Real-World Datasets." Navigation, Journal of the Institute of Navigation 7.5 (2024): 1-20.

BibTeX: Download