A blockchain based federated deep learning model for secured data transmission in healthcare Iot networks
Gopinath Ganapathy, Sujatha Jamuna Anand, M. Jayaprakash, S. Lakshmi, V. Banu Priya, Samuthira Pandi
Abstract
The wide use of sensors in healthcare applications has made it necessary to have secure communication in healthcare Internet of Things (IoT) networks. The sensor data is sensitive, and can contain extremely confidential information such as medical diagnosis, clinical records, vital signs and health data of patients. The emergence of blockchain as a technology ensures consensus and trust among systems, and is now considered to be a new trend used to achieve high scalability, data integrity and privacy. Federated learning is a new technology based on distributed learning that exploits the concept of trust. In federated learning, each user builds an individual distributed model to help a central server that is accessible only to a trusted user group. This paper harnesses the potential of these approaches and proposes an attack detection model to discern normal user behaviours from that of adversaries in a IoT network. This model is called the Blockchain enabled Federated Learning model for secured communication in healthcare IoT (BFL-hIoT), to secure data in healthcare IoT networks. This model is trained and tested on a standard dataset and demonstrates the highest classification accuracy of 97.16% for normal, 0.9546 for backdoors, 0.9618 for XSS etc., outperforming other blockchain and deep learning models.