Federated Learning Based Intrusion Detection Framework using Weighted Voting Ensemble in IoMT

The Internet of Medical Things (IoMT) has transformed healthcare delivery with the integration of connected medical devices for real-time data collection and analysis, better patient monitoring, and personalised treatment. However, the heterogeneous and the distributed nature of IoMT devices, along with their constrained resources, have created major security and privacy problems, making standard security methods inadequate. This paper proposes a new Federated Learning (FL)-based framework for intrusion detection system (IDS) in IoMT. For the purpose of enhancing generalisation and imbalanced data handling, the proposed framework supports decentralised weighted voting ensemble training among clients, leveraging multi- dataset usage and a cross-validation technique. This framework aims to achieve a strong balance between robustness and privacy for intrusion detection in IoMT.