Postdoctoral position “Extracting knowledge from concrete”

Linked Data has been recognized as a promising research field for publishing and interlinking heterogeneous data from different Web repositories. For doing so, Linked Data relies on four main principles, derived from traditional Web architectures’ principles. First, resources are uniquely and globally identified by means of URIs (Uniform Resource Identifiers). Second, resources have to be accessed using the HTTP protocol, therefore HTTP URIs have to be used. Third, resource descriptions have to be delivered using the RDF standard model; resource queries have to be specified in SPARQL standard language. Finally, resource descriptions have to include RDF links to other resources (from other datasets), in order to increase the possibilities for discovering new resources. In the context of BIM (Building Information Modelling), such modelling of resources has been identified as an interesting approach for achieving information interoperability. Indeed, today’s ISO standard for BIM information exchange is the IFC (Industry Foundation Classes) model. IFC was designed in order to optimize the size of STEP-based files exchanged among project stakeholders. Still, in the context of ontology modelling, constraints imposed by IFC object-oriented modelling principles can be leveraged. Moreover, by implementing semantic adaptations of the IFC standard one can implement more intuitive building information manipulation. This is one of the reasons why, in construction and facility management domains, several IFC-based ontologies have been developed. Existing IFC-related ontologies were conceived as direct syntax mappings between EXPRESS and OWL languages. One of the latest and the most solid implementations of an IFC ontology is ifcOWL.

Specific context

Selected during the 2017 ANR call for projects, the McBIM project aims at 1) designing a “communicating concrete”, made of concrete equipped with embedded low-energy wireless micro-sensor network, able to manage and exchange data with BIM platforms, and at 2) demonstrating the usefulness of this approach by implementing it for two building lifecycle phases, namely the construction and exploitation phases (for structural health monitoring).

For achieving this vision of a communicating concrete, several scientific issues arise regarding: a) the design of robust wireless communications, not impacted by the concrete environment, b) the definition of innovative RF harvesting techniques to maximize the lifetime of embedded sensor nodes, c) the definition of new data management strategies controlling how data (either generated by sensor nodes, or sent by users) are spread into the WSN for a fast and reliable data storage and retrieval, d) the definition of a native BIM interoperability of the concrete material, based on IFC standard, to ensure a correct communication with BIM platforms. The McBIM consortium is composed of 4 partners (CRAN, LE2I, LAAS, 360SC), all gathering the skills needed for the project success.

Objectives

More specifically, the person to be recruited on this position will be in charge of developping the knowledge model used for information exchange, with respect to concepts presents in the IFC standard. A special attention will be paid to leveraging from all modelling constraints required by the object-oriented structure of IFC schema of the standard version of ifcOWL. More specifically the following tasks will be adressed: (1) define the a minima knowledge model necessary for ensuring interoperability among the different actors, (2) map the required concrete data to concepts in the IFC standard, (3) ensure that the so-defined piece of information can be efficiently and securely exchanged among actors. For doing so, a particular attention will be given to how the knowledge model and security-related issues impact the size of the message to be exchanged.

Expected profile

Applicants for this position must have a solid background in knowledge modelling, notably description logic-based languages such as OWL 2 DL. Skills using Semantic Web technologies for defining and adressing queries (e.g. SPARQL) along with rule languages (e.g. SWRL, SHACL) are also required. Additional knowledge in civil engineering, architecture and/or construction are considered a plus. Applicants must have a good level of French and English. Applicants should be interested by research activities in general and be able to interact with different actors from different domains (industrials from the AEC domain, as researchers from sensor networks).