An IoMT Power Consumption Dataset for Evaluating Machine Learning-Based Anomaly Detection Systems

The rapid adoption of connected devices in medical environments underscores the urgent need for specialized security solutions in Internet of Medical Things (IoMT) systems. In this paper, we present a comprehensive dataset designed to advance the development of energy-based intrusion detection methods tailored for healthcare IoT devices. The dataset was generated on the FIT IoT-LAB platform within a realistic simulated hospital scenario, capturing fine-grained energy consumption data from medical IoT devices. The dataset covers both normal operation and several attack conditions, including Distributed Denial of Service (DDoS), malware infections, and brute force attacks. This dataset provides a valuable resource for training and evaluating machine learning and deep learning models aimed at detecting host-based threats through power consumption analysis. To validate its effectiveness, we implemented multiple classification models that achieved high detection accuracies ranging from <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 3 \%}$</tex> to <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 9 \%}$</tex>. This dataset lays the groundwork for developing host-based IDS solutions in energy-constrained healthcare IoT contexts.