IoT Systems Security Enhanced Based on Federated Learning and Blockchain Technology
- Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering (LNICST) , 697 (1) : 313-329
Résumé
IoT systems have become very popular in recent years and are present in all areas. They collect and process data that is often sensitive and confidential, making them targets for increasingly sophisticated and frequent attacks. In the field of IoT cybersecurity, Federated Learning and Blockchain Technology are showing promising results. In this paper, we propose an innovative hybrid approach that combines the advantages of multimodal personalised federated learning and blockchain technology to dynamically adapt the model to local resource constraints and secure the learning process, while taking into account the heterogeneity of the domain. This approach stands out for outperforming conventional federated models in its ability to adapt to the heterogeneous resources of IoT devices while guaranteeing effective and reliable detection of cyberattacks on IoT systems. A proof of concept with 10 different IoT clients and 10 different datasets demonstrated feasibility and yielded encouraging results for future work.
Mots-clés
Federated Learning, IoT, Cybersecurity, Blockchain