Health Parameters Monitoring Through an Integrated Multilayer Digital Twin Architecture

This paper evaluates the possibility to use a digital twin architecture to support the management and improvement of the vital health parameters. The proposed approach considers an architecture that integrates machine learning and deep learning algorithms. The digital twins are automated programs that learn from training data and new data, and then assist the end user in various situations. The digital twin model that is proposed in this paper, which is called MyMLTwin, assists the user in monitoring and keeping the health parameters within the correct threshold values. Threshold values can be set as a target by users or suggested and modified by the digital twin’s artificial intelligence core. Data are securely collected from the end users and properly anonymized. Furthermore, the data can be used to digitally assist the user in maintaining their health parameters by suggesting the proper nutrition plan, and physical activity patterns relative to the proposed digital twin model. The model includes elements of web semantics and appropriate ontologies, which are used to create a proper equivalent of the required nutrition plan. The proposed model integrates state of the art conceptual models and technologies that relate to machine learning, deep learning, and distributed systems. The paper describes a functional prototype that explores different types of features related to different algorithms and methods used to train and support the decisions of the digital twin.