Enhancing Medical Cybersecurity: A Machine Learning Solution for Intrusion Detection in IoMT Devices

In the past few decades, the healthcare sector has undergone major technological advancements, which have included the adoption of Internet of Medical Things (IoMT), and networked medical devices. These advancements have changed the manner in which patient care is provisioned, enabling for more accurate diagnosis, personalized treatment plans, among other benefits. However, as healthcare providers are growing more reliant on interconnected systems, they have also been exposed to critical cybersecurity vulnerabilities. Medical devices have now become targets for cyberattacks, threatening patient data and its security and endangering patient's health and safety. This thesis examines the affects that inadequate cybersecurity measures, regulations, and frameworks have on the safety and functionality of medical devices within the healthcare sector. Furthermore, we provide a comprehensive review on current cybersecurity controls, examine their limitations, and highlight potential areas of improvement. Additionally, we also explore the function of machine learning (ML) methodologies in addressing these challenges, with a particular focus on intrusion detection systems (IDS) for Internet of Medical Things devices. Finally, we showcase a ML-based IDS solution that will detect cyber threats on IOMT devices and assess the model's performance and address the limitations faced by the solution. This thesis aims to support to the current efforts in strengthening the cybersecurity of healthcare systems by recommending practical solutions for healthcare providers, regulatory bodies, and cybersecurity professionals.