Е-Публикации
Технически университет - София

Abstract: Integrating machine learning (ML) into industrial Internet of Things (IIoT) environments offers promising opportunities for improving cybersecurity in critical infrastructure, manufacturing, and industrial control systems. However, IIoT systems face unique challenges, including legacy equipment, real-time constraints, and highquality zero-day vulnerabilities, which make conventional security methods inadequate. This paper reviews current machine learning approaches for IIoT security, identifies limitations in detecting sophisticated threats, and proposes a hybrid architecture that combines traditional encryption methods and machine learning tools to enhance security while keeping hardware capabilities in mind. An adaptive framework for distributed and scalable security is presented. Key research directions are outlined to provide explainable, lightweight, and ethical machine learning solutions tailored to the unique requirements of HoT systems

References

1. [Langner, R. (2011). Stuxnet: Dissecting a Cyberwarfare Weapon. IEEE Security & Privacy, 9(3), 49–51.
2. FireEye. (2017). Attackers deploy new ICS attack framework “TRITON”. https://www.fireeye.com.
3. Knowles, W. et al. (2015). A study of cybersecurity management in industrial control systems. Int. J. Critical Infrastructure Protection, 9, 52–80.
4. Stouffer, K. et al. (2015). Industrial Control Systems Security Guide. NIST SP 800-82, revised edition 2.
5. Carcano, A. et al. (2011). Multivariate critical state analysis for SCADA intrusion detection. IEEE Trans. Ind. Informatics, 7(2), 179–186.
6. Goh, J. et al. (2017). Anomaly Detection in Cyber-Physical Systems Using Recurrent Neural Networks (RNNs). In IEEE HASE.
7. Li, T. et al. (2020). Federated learning: Challenges, methods, and future directions. IEEE Signal Process. Mag., 37(3), 50–60.
8. Worden, P. and Situnayake, D. (2019). TinyML. O'Reilly Media.
9. Goodfellow, I. et al. (2015). Explanation and use of examples of competitive relationships. arXiv:1412.6572.
10. Doshi-Veles, F. and Kim, B. (2017). Towards a rigorous science of interpretable machine learning. ArXiv:1702.08608.
11. Holzinger, A. et al. (2019). Explaining Artificial Intelligence in Medicine. WIREs Data Mining Knowl Discoveries.
12. Hristov A. Simulation of Cookie Poisoning network attacks, Proceedings of the 59th International Scientific Conference on Information, Communication and Energy Systems and Technologies – ICEST 2024, IEEE Conference, Rec # 62335, July 1-3, 2024, Sozopol, Bulgaria, DOI: 10.1109/ICEST62335.2024.10639772
13. IEC 62443 Series. (2018). Industrial Communication Networks – Security.
14. NIST. (2018). Framework for Improving Critical Infrastructure Cybersecurity.
15. Alcaraz, K. (2019). Security in ICS and IoT. Future Internet, 11(5), 114.
16. Humayed, A. et al. (2017). Security of Cyber-Physical Systems – A Survey. IEEE IoT Journal, 4(6), 1802–1831.
17. Hristov A. Using Python for development of an application for building and experimenting with GPSS simulation models, Proceedings of the 31st scientific conference "TELECOM 2023", IEEE Conference, Rec # 59629, 16 – 17 November 2023, Sofia, Bulgaria, DOI: 10.1109/TELECOM59629.2023.104096 96
18. N. Nikolov, O. Nakov and D. Gotseva, "Design and Research of Smart IoT Control System for Electrical Appliances," 2021 29th National Conference with International Participation (TELECOM), Sofia, Bulgaria, 2021, pp. 39-42, doi: 10.1109/TELECOM53156.2021.9659650.
19. N. Nikolov and D. Gotseva, "A Secure Firmware Update over the Air of ESP32 using MQTT Protocol from Cloud," 2023 XXXII International Scientific Conference Electronics (ET), Sozopol, Bulgaria, 2023, pp. 1-4, doi: 10.1109/ET59121.2023.10278597.

Issue