A Privacy-Preserving Deep Learning Framework for Real-Time Intrusion Detection in Internet of Medical Things (IOMT) Networks

With the explosive growth of Internet of Medical Things (IoMT) devices in contemporary healthcare, the attack surface that is vulnerable to increasingly sophisticated cyber threats has grown to encompass a variety of devices and systems. The current intrusion detection systems (IDS) built with legacy machine learning algorithms have a high false-positive rate, feature engineering takes a lot of manpower, and the algorithms show limited generalisation when faced with zero-day attacks. To tackle these limitations, this research introduces a novel deep learning framework named CNN-HBO-GAN-FL which combines a Convolutional Neural Network (CNN) optimised using the Honey Badger Optimisation Algorithm (HBOA), a Generative Adversarial Network (GAN) for class-balanced data augmentation, and Federated Learning (FL) for privacy-preserving, distributed model training. This framework was then translated into Python using TensorFlow/Keras and tested on the CICIoMT2024 benchmark dataset. The experimental results demonstrate that the proposed framework attained 99.18% accuracy, 98.96% precision, 98.72% recall, 98.84% F1-score and 1.7% false-positive rate whereas traditional CNN, CNN-LSTM and CNN-PSO baselines achieved 96.99%, 97.55% and 97.74%, respectively. The model convergence time improved by around 35% compared to the baseline models when optimized using hyperparameter tuning based on HBOA, and the minority-class attack sensitivity was improved by 16% when the model was optimized using the augmentation based on the GAN. The federated learning architecture maintained the accuracy of centralised training by less than 0.4% variance, which has the potential to help ensure adherence to data-protection laws and regulations including HIPAA, the GDPR, and Nigeria's Data Protection Act 2023, but does not necessarily ensure compliance. The results demonstrate the potentiality of the proposed framework as a resource-efficient solution for preserving privacy in intrusion detection of future IoMT infrastructure.