Design of a Blockchain-Based Secure Health Monitoring System Using Decentralized Machine Learning Technique

A health monitoring system (HMS) is an integrated system that includes the sub-systems of health data gathering through sensors, health data analysis, and patient-doctor real-time communication. HMS allows patients to get medical care remotely, and doctors to provide real-time information to patients. The system not only reduces patients' time-cost, but also increases the quality of medical care. To solve the security problem of centralized management of data, blockchain technology has been introduced, as it contains the properties of immutability, transparency, and distribution. To take advantage of the blockchain network structure, this article proposes a system framework integrated with decentralized machine learning, aiming to improve the system performance in terms of throughput and model accuracy. It is a combination of the Hyperledger Fabric network and ModelChain model training method, where Hyperledger Fabric allows users to be grouped and managed in the form of organizations, while ModelChain uses the characteristics of grouped users to explore more valuable information. This article proposes the scheme to allow access control on a Hyperledger Fabric system and the algorithm to implement ModelChain on a Hyperledger Fabric network. Furthermore, the system is built and measured by tools, such as Hyperledger Caliper, Docker, and Weka, and is evaluated in terms of system throughput and accuracy.