Anomaly Detection on Network Traffic for the Healthcare Internet of Things

The Internet of Things (IoT) has revolutionized technologies in society, including in households, offices, factories, and health centers. Among these, the Healthcare Internet of Things (HIoT) significantly transforms medical assistance for patients. By using wearable devices with remote network connections, caregivers monitor patients’ physiological data to gain valuable insights into their health conditions. Despite the many benefits of the HIoT, several security vulnerabilities still exist. Hackers can exploit the internet connection to steal or modify credential information regarding patients, violating the integrity and confidentiality of the security policy. Moreover, they can launch cyberattacks on hospitals or critical life-support systems, further endangering patients’ lives. Consequently, it is crucial to implement robust cybersecurity measures to enhance the security of healthcare services. Therefore, we proposed an anomaly detection method based on network traffic for the HIoT, adopting Markov models. Owing to their simplicity, interpretability, and well-developed theory, the Markov models have been applied to network traffic prediction and modeling, serving as a viable approach to cater to our needs. We evaluated the proposed method using the public dataset ToN_IoT and analyzed the results.