Applying artificial intelligence techniques to intrusion detection systems in serial-based industrial networks

Industrial control systems often rely on serial-based networks that lack robust cybersecurity measures, making them vulnerable to attacks as they increasingly integrate with corporate networks. While intrusion detection systems (IDSs) are extensively used in Ethernet-based networks, their adoption in serial-based networks remains limited. This chapter investigates the application of artificial intelligence (AI) techniques to enhance intrusion detection in these networks. By combining rule-based methods with supervised and unsupervised learning, AI-powered IDS can detect known and novel attacks effectively. The chapter reviews AI techniques, compares their effectiveness, and highlights their potential to extend cybersecurity measures to the most vulnerable layers of industrial networks. The findings emphasize the critical role of AI in safeguarding industrial serial-based systems and propose strategies for developing IDS, tailored to their unique requirements, addressing challenges such as legacy system constraints and the need for real-time anomaly detection.