Seybold Report ISSN: 1533-9211
1Arvind Kamble , 2 Dr.Virendra S Malemath
Vol 17, No 12 ( 2022 ) | Licensing: CC 4.0 | Pg no: 3502-3523 | Published on: 30-12-2022
Abstract
Cyber-physical systems (CPS) are becoming increasingly prevalent in modern society, from autonomous vehicles to smart homes and industrial control systems. With the proliferation of these systems, the need for robust intrusion detection methods is more pressing than ever before. Anomaly-based intrusion detection using deep recurrent neural networks (RNNs) has emerged as a promising approach for detecting cyber attacks in CPS applications. In this paper, we present a survey of the current state-of-the-art in anomaly-based intrusion detection using deep RNNs for CPS applications. We review the key challenges associated with applying deep RNNs to CPS data and discuss the various techniques that have been developed to overcome these challenges. We also provide a comprehensive overview of the datasets and evaluation metrics commonly used to benchmark intrusion detection methods in CPS. Our survey highlights the potential of deep RNNs for detecting cyber attacks in CPS applications, particularly for detecting subtle changes in system behavior that may occur over an extended period of time. We also identify several research gaps that need to be addressed to further advance the use of deep RNNs for intrusion detection in CPS, including the development of more realistic datasets and the need for more explainable intrusion detection methods. Overall, our survey provides valuable insights into the current state-of-the-art in anomaly-based intrusion detection using deep RNNs for CPS applications and highlights several areas for future research.
Keywords:
Cyber Physical Systems, Intrusion detection, Deep Recurrent Neural Network, Cyber attacks, Machine Learning