Efficient Anomaly Detection for Smart Hospital IoT Systems

Area of Science:

• Internet of Things (IoT) security
• Healthcare systems engineering
• Network intrusion detection

Background:

• False alerts in critical IoT applications like healthcare and defense cause significant disruptions and resource waste.
• Internet of Things (IoT) network intrusions compromise the accuracy of event detection systems.
• Smart hospital environments require reliable monitoring for patient well-being and operational efficiency.

Purpose of the Study:

• To propose an Anomaly Detection System (ADS) for smart hospital IoT environments.
• To simultaneously detect patient health events and network intrusions.
• To enhance the reliability and accuracy of e-health monitoring and network supervision.

Main Methods:

• Developed an Anomaly Detection System (ADS) deployed at the network edge for low-latency processing.
• Integrated network infrastructure supervision with e-health monitoring capabilities.
• Evaluated the ADS using the Contiki Cooja simulator and realistic e-health datasets.

Main Results:

• The proposed ADS demonstrated high detection accuracy for both e-health events and IoT network intrusions.
• Edge deployment ensured low-latency processing, crucial for real-time healthcare applications.
• A unified system for monitoring and intrusion detection optimized resource utilization and system reliability.

Conclusions:

• The integrated ADS effectively improves the accuracy of patient care decisions and environmental adaptations in smart hospitals.
• Combining health event detection with network intrusion detection offers a more robust and efficient IoT security solution.
• The edge-based ADS is a viable approach for enhancing the security and reliability of critical IoT healthcare systems.