Résumé:
The Internet of Things (IoT) operates across various domains, such as healthcare, withthe aim of enhancing performance through the remote and real-time collection of data.
This technology facilitates the monitoring of patients’ health status by measuring theirvital signs. However, to fully exploit the potential of the IoT and take advantage of allthe opportunities it offers, it is necessary to solve the problems of heterogeneity and thelack of interoperability.
The semantic web of things is a solution developed by researchers to solve these problemsby integrating the Semantic Web and the Internet of Things. This framework is based onsemantic technologies such as RDF, RDFS, OWL and ontology. However, using variousontology development methods leads to creating specific ontologies that cause alignmentand sharing problems, especially when the domain represented by this ontology is interconnected with other domains, such as healthcare related to transportation, education,
etc. This framework also suffers from a lack of capacity to deal with the vague andimprecise data that characterize the health domain. This state can lead to inaccurateprocessing and incorrect results, which is unacceptable in this domain where accuracy iscrucial for decision-making. As this framework is designed to collect data from numerousIoT devices, this leads to a large amount of data in RDF format, also known as BIG RDFdata. To process this data efficiently, an efficient storage and retrieval method is necessary. In addition, the framework should enable easy incorporation of new IoT devices aswell as ensure real-time processing.
In this thesis, to overcome these problems, a semantic web of things framework has beendeveloped. The development of this framework is accomplished through three contributions. Firstly, an ontology has been developed, which is the basis of the framework andhas been developed using the neon methodology. The ontology is an extension of thestandard SAREF ontology that supports alignment and sharing. The ontology allowsa representation of their integration of domain health and public transport. Secondly,to address the problem of vague and imprecise data, another more advanced ontologythat can deal with vague and imprecise data in health data was developed. The ontology has represented a transformation of our SAREF ontology extension for COVID-19 tofuzzy ontology, which helps the framework for more accurate and reliable decision-making.
Third, adding the cluster and indexing layers to the framework can process the BIG RDFdata by grouping these outputs into clusters and narrowing the search space. Finally,systems based on an IoT architecture and using early warning systems such as MEWSand NEWS2 have been developed to validate these contributions. These systems allowthe determination of the patient’s health status, which allows the provision of appropriate health services. The results obtained with the developed system were considered verypromising and encouraging
