Using Federated Artificial Intelligence System of Intrusion Detection for IoT Healthcare System Based on Blockchain
Priyanka Tyagi, S.K. Manju Bargavi
Abstract
Recently Internet of things (IoT)-based healthcare system has expanded significantly, however, they are restricted by the absence of an intrusion detection mechanism (IDS). Modern technologies like blockchain (BC), edge computing (EC), and machine learning (ML) provide a robust security solution that is well-suited to protecting patients' medical information. In this study, we offer an intelligent intrusion detection mechanism FIDANN that protects the confidentiality of medical data by completing the intrusion detection task by utilising Dwarf mongoose-optimized artificial neural networks (DMO-ANN) through a federated learning (FL) technique. In the context of recent developments in blockchain technology, such as the elimination of contaminating attacks and the provision of complete visibility and data integrity over the decentralized system with minimal additional effort. Using the model at the edges secures the cloud from attacks by limiting information from its gateway with less computing time and processing power as FL works with fewer datasets. The findings demonstrate that our suggested models perform better when dealing with the diversity of data produced by IoT devices.