Institutional Repository of Vilnius University
Anomalijų aptikimas palydovų posistemėse /
Abstract
dc:descriptionThis paper explores the possibility of using unsupervised machine learning techniques for anomaly detection in satellite subsystem telemetry. Several comprehensive studies comparing a wide range of anomaly detection models on a number of datasets are analyzed, after which an educated choice is made when choosing models to be tested. The testing of anomaly detection models and their ensembles on both simulated (CATS) and real-world satellite telemetry data revealed several key findings. A KNN model ensemble showed the best performance on the CATS dataset, suggesting its strong potential for anomaly detection in satellite telemetry data. Combining KNN and OCSVM models improved detection capabilities highlighting the potential of the combination of different techniques. Finally, the best performing model is trained on actual satellite data and its detection capabilities are analyzed.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Leipus, Joris,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD210646568&prefLang=en_US