Intrusion Detection in Medical Networks Based on Hybrid Optimization algorithms

Abstract

The Internet of Medical Things (IoMT) has transformed modern healthcare by enabling continuous patient monitoring, remote diagnostics, and real-time data exchange. While these advancements improve service quality and clinical outcomes, they also expose medical networks to a wide range of cyber threats. IoMT environments are particularly vulnerable due to their reliance on heterogeneous, resource-constrained devices that often lack built-in security mechanisms. Conventional intrusion detection systems (IDS) are often inadequate for these settings, as they struggle to deliver high detection accuracy while remaining lightweight and adaptive. This study proposes a hybrid IDS framework designed specifically for IoMT ecosystems. The approach integrates a Clonal Selection Algorithm (CSA) for dynamic feature selection with a Deep Neural Network (DNN) for accurate classification of network traffic. The CSA component effectively reduces data dimensionality while preserving critical threat-related features, thereby optimizing the model for environments with limited computational capacity. The DNN component captures complex patterns in traffic behavior, enhancing detection capability across diverse attack scenarios. The model is evaluated using benchmark IoMT datasets, demonstrating its suitability for identifying both known and emerging threats. By combining bio-inspired optimization with deep learning, the proposed IDS offers a robust and scalable solution to enhance cybersecurity in sensitive and critical healthcare systems