Cyber-Physical Systems and Reliability Issues
- DOI
- 10.2991/aisr.k.201029.026How to use a DOI?
- Keywords
- cyber-physical systems, internet of things, classification, time-series, machine learning
- Abstract
Anomaly detection is a well researched concept used in many areas, including engineering systems design, where it helps detect errors and prevent failures. Traditional anomaly detection methods, based either on comparing the behavior of the real system with its model or on different signal processing methods, have been successfully applied for Fault Detection and Isolation (FDI) in mechatronic systems. Cyber-Physical Systems (CPS) are complex in both structural and behavioral terms. They consist of numerous heterogeneous components that generate large volumes of data, exchange information and form extremely complex patterns of behaviour. This makes it almost impossible to effectively set up and apply classical reliability assessment methods. The article discusses the basic models of machine learning and the possibility of their application to solve the problem of CPS reliability. It is proposed to use neural networks with long short-term memory (LSTM) to detect anomalies in the CPS. The neural network architecture with 3 hidden, input and output layers was designed. The experiment with testing data (Tennessee Eastman Process dataset) was conducted and analysis of the results was carried out.
- Copyright
- © 2020, the Authors. Published by Atlantis Press.
- Open Access
- This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
Cite this article
TY - CONF AU - Nafisa Yusupova AU - Dmitry Rizvanov AU - Dmitry Andrushko PY - 2020 DA - 2020/11/10 TI - Cyber-Physical Systems and Reliability Issues BT - Proceedings of the 8th Scientific Conference on Information Technologies for Intelligent Decision Making Support (ITIDS 2020) PB - Atlantis Press SP - 133 EP - 137 SN - 1951-6851 UR - https://doi.org/10.2991/aisr.k.201029.026 DO - 10.2991/aisr.k.201029.026 ID - Yusupova2020 ER -