D2.3 Evaluation of the semantic model in real world scenarios

The document is the deliverable “D2.3 - Evaluation of the semantic model in real world scenarios” of the VICINITY project, funded by the European Commission’s Directorate-General for Research and Innovation (DG RTD), under its Horizon 2020 Research and Innovation Programme (H2020).  The VICINITY ontology network developed in the context of this project consists so far of five ontology modules, i.e., the VICINITY Core (Core), the Web of Things (WoT), the WoT mappings (Mappings), the VICINITY Adapters (Adapters), and the Datatypes (Datatypes) ontologies, belong to the VICINITY ontology network and aim to provide interoperability in the IoT domain. The Core ontology represents the information needed to exchange IoT descriptor data between peers through the VICINITY platform; this ontology is being created by following a cross-domain approach and implements requirements from different domain experts. The WoT ontology aims to model the Web of Things domain according to the W3C WoT Interest Group19 descriptions. The Mappings ontology represents the mechanism for accessing the values provided by web things in the VICINITY platform. The Adapters ontology aims to model all the different types of devices and properties that can be defined in the VICINITY platform. Finally, the Datatypes ontology aims to model the required and provided datatypes that are used in the interaction patterns of the platform. The methodology for developing this ontology network is iterative and based on the NeOn methodology.

 This document covers the following main topics: 

  • Validation regarding the model, to assure that there are no inconsistencies in the ontologies by using semantic reasoners) and that there are no modelling errors (by using the tool OOPS!).
  • Verification regarding their ontological requirements, to guarantee that all the requirements asked by the domain experts are satisfied by the ontology.
  • Validation regarding pilot data, to analyse how the data is used in VICINITY relying on the VICINITY ontology.
  • Verification regarding IoT standards, which analyses the coverage of the VICINITY ontology regarding the ontological commitments of a set of well-known IoT standards.

From the evaluation presented in this document it could be concluded that the VICINITY network does not have inconsistencies or modelling errors, and that it covers all the requirements given by the partners. Additionally, it was also concluded that, even though the VICINITY ontology network does not cover all the requirements in those standards, there are no inconsistencies between the standards and the VICINITY ontologies and that VICINITY ontology network has partial conformance with the IoT standards. However, this coverage analysis also shows that the VICINITY ontology network is out of scope of the analysed IoT standards, which is expected because the IoT standards are more generic than the VICINITY ontology network.

We have analysed the Things and Thing Descriptions registered by the pilots to verify which parts of the ontology are used the most. It is well known that VICINTIY relies on several core components that work with this data, and therefore, the correct specification of it is paramount. The results of our analysis advocates that pilots are correctly using the ontology, although some parts should be use more; like the parts that refer to the contextual data, or the mappings.