RCLNet: an effective anomaly-based intrusion detection for securing the IoMT system

Area of Science:

• Cybersecurity in Healthcare
• Medical Informatics
• Machine Learning Applications

Background:

• The Internet of Medical Things (IoMT) offers advanced healthcare solutions like remote monitoring, but faces significant patient data security challenges from cyber threats.
• Existing machine learning and intrusion detection systems struggle with complex IoMT data patterns and identifying novel attacks, leading to data breaches and high false positives.

Purpose of the Study:

• To develop an effective Anomaly-based Intrusion Detection System (A-IDS) specifically for IoMT environments.
• To improve the security and confidentiality of sensitive patient data within IoMT systems.

Main Methods:

• Proposed RCLNet, integrating Random Forest (RF) for feature selection, Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) for pattern recognition, and a Self-Adaptive Attention Layer Mechanism (SAALM).
• Utilized focal loss (FL) to address imbalanced data distributions common in IoMT datasets.
• Evaluated performance on the WUSTL-EHMS-2020 healthcare dataset.

Main Results:

• RCLNet achieved a high accuracy of 99.78% in detecting anomalies.
• Demonstrated superior performance compared to recent state-of-the-art methods in intrusion detection for IoMT.
• Successfully managed imbalanced data and identified complex patterns effectively.

Conclusions:

• RCLNet presents a robust and effective solution for securing IoMT systems against sophisticated cyber threats.
• The proposed system significantly enhances patient data security and confidentiality in healthcare settings.
• RCLNet's advanced architecture offers a promising approach for future IoMT security research.