Anomaly Detection using System Identification Techniques
Gheorghe Sebestyen, Anca Hangan
2018
Abstract
As cyber-physical systems are becoming more human independent any anomaly or system failure should be detected and solved in an autonomous way. In the last decade significant research was performed to find more intelligent and accurate anomaly detection methods. Most of these methods are analyzing only the output(s) of a system hoping to find some inconsistencies in the data stream. Our attempt is to consider the system’s model as well and develop an anomaly detection methodology that tries to identify slight changes in the behavior of the system, detectable through model changes. The key part of our detection method is the system identification step through which we compute the system’s model considered as a differential equation between input and output signals or as an autoregression formula. We demonstrate the feasibility of the proposed method through a simulated and a real-life example.
DownloadPaper Citation
in Harvard Style
Sebestyen G. and Hangan A. (2018). Anomaly Detection using System Identification Techniques.In Proceedings of the 15th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO, ISBN 978-989-758-321-6, pages 482-487. DOI: 10.5220/0006888604820487
in Bibtex Style
@conference{icinco18,
author={Gheorghe Sebestyen and Anca Hangan},
title={Anomaly Detection using System Identification Techniques},
booktitle={Proceedings of the 15th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO,},
year={2018},
pages={482-487},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0006888604820487},
isbn={978-989-758-321-6},
}
in EndNote Style
TY - CONF
JO - Proceedings of the 15th International Conference on Informatics in Control, Automation and Robotics - Volume 1: ICINCO,
TI - Anomaly Detection using System Identification Techniques
SN - 978-989-758-321-6
AU - Sebestyen G.
AU - Hangan A.
PY - 2018
SP - 482
EP - 487
DO - 10.5220/0006888604820487