Smart ICU Security Framework Using IoT and Deep Learning-Based Anomaly Detection

Abstract

The rapid growth of Internet of Things (IoT) technologies in healthcare has improved patient monitoring and automated medical services, especially in Smart Intensive Care Units (ICUs). However, the increasing number of connected medical devices also exposes healthcare systems to cyber threats and security vulnerabilities. This study proposes an IoT-enabled Smart ICU framework for real-time anomaly detection using advanced machine learning and deep neural network techniques. The BoT-IoT dataset is utilized to train and evaluate the proposed system. Data preprocessing methods such as feature selection, label encoding, SMOTE-based class balancing, and feature standardization are applied to improve model performance. Several machine learning algorithms, including K-Nearest Neighbors (KNN), Random Forest, and XGBoost, are implemented along with deep learning models such as Feedforward Neural Network (FFNN), Deep Neural Network (DNN), Convolutional Neural Network (CNN), and CNN-BiLSTM. In addition, ensemble learning approaches like Voting and Stacking classifiers are used to enhance prediction accuracy and robustness. Experimental results show that the Stacking Classifier achieves the best performance with 99.99% accuracy. Explainable AI techniques such as LIME and SHAP are incorporated for model interpretability. A Flask-based web application is also developed for real-time monitoring and prediction in Smart ICU environments.