Semantic web of things-based architecture for healthcare services

Abstract

Background Unlocking interoperability challenges towards medical devices is the core of this research work. This work reports an adaptable Semantic Web of Things (SWoTs) architecture that requires due consideration of Internet of Things (IoTs) devices/applications to set out the fundamental techniques and technologies from device level to software defined approach through the implementation of ontological methods—for high level data/information representation. Methods This work adopts design science research approach proposing SWoTs based architecture for medical domain. The methodology covers six distinct activities from problem definition and motivation up to communication. Evaluation of the proposed approach relies on patient data obtained from Tirunesh Beijing General Hospital which is approved by the Addis Ababa Regional Health Bureau Research Review Board. In addition, patient consent was taken for all data under consideration. Results In this study, a total of 120 patients participated, out of which only 73 are applicable for our purpose. Extracted sensors data from body temperature and heart rate measurements of IoT devices are converted to common presentation format (JSON—JavaScript Object Notation) with data rate of 10 s where the issue of interoperability is dealt. The acquired vital sign values are associated with other objects/classes (with full attribute definition) using ontological method and reasoned via Semantic Web Reasoner Language (SWRL) to support medical decision. The ontology is evaluated against known evaluation metrics. The outcome of this work comes up with a scoped and focused set of software engineering techniques; including algorithms, flowcharts, Python code level representations, illustrative methods, architecture, deployment approach for healthcare settings, including Unified Modeling Language (UML), architectural illustration at component level and evaluation of the work using K6 test tools. Discussion The proposed architecture comprises: (i) an acquisition layer for context parsing; (ii) data aggregation layer for pre-processing and classification; (iii) repository layer for medical datasets and knowledge representation; (iv) cognition layer for semantic reasoning; and (v) application layer with mobile and web application program interfaces (APIs). A security layer protects patient data throughout. The architecture was tested using microcontroller development modules with Bluetooth Low Energy (BLE) to capture heart rate and temperature readings as basic vital sign information with varying formats. The deployment of the architecture assumes cloud, fog, and edge computing techniques, with detailed component mapping.