Semantics

IoT is facing many challenges: interoperability, data consistency, dynamicity... These issues ask for data models with specific properties: decentralization, flexibility, interoperability and scalability. Semantic web and linked open data principles [1] cover these requirements.

The semantic web is a web of data, linked to each other and described with expressive vocabularies called ontlogies. Ontologies allow an enrichment of data in order to make it meaningful to systems. They also provide a level of formalism that allow automatic reasoning to infer new data.

Our team proposed IoT-O [2], a core domain ontology for the IoT. It aims at enriching data coming from an IoT system with meta-data describing said system. IoT-O will be soon showcased in an open data to enrich real data and make it more easily reusable. IoT-O is also used to enable the autonomic control of a smart building model.

Many projects integrate semantic web principles and technologies to overcome IoT issues: for example,  oneM2M proposes an ontology to describe its data model [3], and the W3C [4] is also working semantics for the IoT. We take part in the oneM2M MAS working group, in charge of the semantics in the norm.

[1] Christian Bizer Tim Berners-Lee, T. H. (2009). Linked Data - The Story So Far. International Journal on Semantic Web and Information Systems (IJSWIS), 5(3), 1–22. http://doi.org/10.4018/jswis.2009081901

[2] Nicolas Seydoux, Mahdi Ben Alaya, Nathalie Hernandez, Thierry Monteil, Ollivier Haemmerlé. Sémantique et Internet des objets : d'un état de l'art à une ontologie modulaire. 26es Journées francophones d'Ingénierie des Connaissances, Jun 2015, Rennes, France. 26es Journées francophones d'Ingénierie des Connaissances, 2015, <http://pfia2015.inria.fr/conference-ic>. <hal-01166052>

[3] http://www.onem2m.org/ontology/Base_Ontology/

[4] https://www.w3.org/WoT/IG/