Artificial intelligence for IoMT security: A review of intrusion detection systems, attacks, datasets and Cloud–Fog–Edge architectures
Mireya Lucia Hernandez-Jaimes, Alfonso Martínez-Cruz, Kelsey Alejandra Ramírez-Gutiérrez, Claudia Feregrino-Uribe
Abstract
Recent advances in the Internet of Medical Things (IoMT) have impacted traditional medical treatment and have evolved data communications in the Smart Healthcare scenario. Unfortunately, this also has resulted in a fertile field for attackers. As a result, traditional intrusion detection models and new detection schemes for IoMT applications have been applied. Along with it, there has been a rising trend of employing different types of artificial intelligence algorithms to improve the detection performance of attacks in medical systems communications. This paper provides a novel taxonomy of intrusion detection schemes for IoMT, which includes a comparative analysis of intrusion detection methods and an unique classification of current datasets for insights into detection performance. Additionally, we discuss the cybersecurity threats regarding the IoMT architecture and the security requirements of IoMT. Moreover, an examination of the tasks carried out in Cloud-Fog-Edge architectures, and a classification of recent literature based on the AI methods, are presented. We also discuss the legal and ethical security aspects of IoMT. Finally, we provide the challenges and novel perspectives requiring further investigation.