ExGAME: An Explainable Game Theoretic and Adaptive Intrusion Detection Framework for Human-Centric Medical IoT
TL;DR
The proposed framework combines machine-learning-based anomaly detection with explainable AI and a game-theoretic defense strategy to improve both detection performance and decision-making clarity and shows that packet rate and traffic flow characteristics are very important for distinguishing between normal and abnormal network activities.
Abstract
: The rapid deployment of Internet of Medical Things (IoMT) devices in current healthcare systems has made it much easier to maintain patient monitoring, make diagnoses, and provide long-distance medical treatment. The interconnection of these devices also creates significant cybersecurity problems, including distributed denial-of-service attacks, data breaches, and network intrusions. High detection accuracy and interpretability are essential for a trustworthy intrusion detection system. This study presents ExGAME, an Explainable Game-Theoretic Artificial Intelligence framework intended for intrusion detection in human-centric IoT networks. The proposed framework combines machine-learning-based anomaly detection with explainable AI and a game-theoretic defense strategy to improve both detection performance and decision-making clarity. A game-theoretic perspective is used to explain the interaction of the attackers and defenders. Experiments have been performed using IoT-23, Bot-IoT, CICIDS2017, UNSW-NB15, WUSTL-EHMS-2020, and MedBIoT datasets. Results achieved an accuracy range of upto 98% in different attack scenarios. The results show that packet rate and traffic flow characteristics are very important for distinguishing between normal and abnormal network activities. The proposed ExGAME architecture makes detection more transparent while enabling effective intrusion localization. This research aids in the creation of secure, comprehensible, and flexible protection mechanisms for future healthcare IoT systems.