Semantic IoT Semantic Inter-Operability Practices - Part 1


Published on

G. Cassar Semantic IoT Semantic Inter-Operability Practices-Part1 presented at the IERC AC4 IoT Semantic Interoperability workshop, Guildford, UK, 15 April 2013

Published in: Education
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Semantic IoT Semantic Inter-Operability Practices - Part 1

  1. 1. IoT Semantic Inter-Operability EventPart 1: IoT semantic interoperability practicesPresenter: Gilbert CassarCentre for Communication Systems Research, University of SurreyContributors: Dr. Wei Wang, Dr. Payam Barnaghi, Dr. Martin Serrano,Mr. Phillippe Cousin
  2. 2. Getting Started Install Virtual Box Copy ‘InteropEventVM’ from the USB sticksprovided. Load the VM on Virtual Box. Also on the USB Stick: Sensor Ontologies Quantity Type ontologies.
  3. 3. Getting started with Protégé 4 Protégé is an OWL-specific integrateddevelopment environment (IDE) for developing andmaintaining OWL ontologies. Already installed on your VM. To start Protégé:Home/protégé_4.2/
  4. 4. Getting Started with Protégé 4 Tutorials for Protégé 4.2 can be found at: Creating OWL ontologies: Open existing OWL ontologies Open an ontology at a URL Import existing ontologies Each ontology should have a unique defaultnamespace.
  5. 5. Creating classes Named classes - create a class and assign a nameto it. Two ‘built in’ named classes: owl:Thing andowl:Nothing. Defining subclass: rdfs:subClassof Asserting a class is the same as another:owl:equivalentClass Asserting a class is disjoint with another:owl:disjointWith
  6. 6. Checking ontologies We would like to automatically check our ontologyto ensure that the logical meaning corresponds tothe intended meaning, e.g., an individual of a classshouldn’t be an individual of its disjoint classes. For an ontology that falls into the scope of OWL-DL, we can use a DL Reasoner to infer informationthat isn’t explicitly represented in the ontology.
  7. 7. Reasoning in Protégé DL reasoner can be plugged into Protégé HermiT Fact++ Standard reasoning services: Subsumption checking Equivalence checking Consistency checking Instantiation checking
  8. 8. Creating properties OWL has two main types of properties: Object properties Datatype properties. Object properties relate an individual to anindividual. Datatype properties link an individual to a datavalue. Annotation properties can be used to attach ‘meta-data’ to classes, properties and individuals.
  9. 9. More on properties OWL supports the specification of a propertyhierarchy; in OWL-DL, object properties may onlyhave object properties as super-properties, andsame for datatype properties. Properties have a Domain and a Range.
  10. 10. Exercises 1: use Protégé Study the following ontologies in Protégé: W3C SSN: OWL-S: IoT-A ontologies:
  11. 11. Exercises 1: use Protégé cont’d Open the following ontologies in Protégé: IoT.est ontologies:
  12. 12. What is expected from the semanticinteroperability? Unified access to data: unified descriptions and at the same time an openframework. Self-descriptive data and re-usable knowledge. Deriving additional knowledge. Reasoning support and association to other entitiesand resources. Enabling autonomous interactions with the resources.
  13. 13. Potential solutions Using machine-readable and machine-interpretablemeta-data Well defined standards and description frameworks: XML,RDF, OWL,etc. Variety of technologies and tools for creating/managing/querying andaccessing semantic data, e.g., Jena, Sesame, Protége, etc. Ontologies defines conceptualisation of a domain. Domain concepts modeling Relationships between the concepts Link to existing knowledge, the linked open data cloud
  14. 14. Semantics in IoT – myth and reality #1: If we create an Ontology our data isinteroperable Reality: there are/could be a number of ontologies for a domain Ontology mapping Reference ontologies Standardisation efforts #2: Semantic data will make my data machine-understandable and my system will be intelligent. Reality: it is still meta-data, machines don’t understand it but caninterpret it. It still does need intelligent processing, reasoning mechanismto process and interpret the data.
  15. 15. Semantics in IoT – myth and reality #3: It’s a Hype! Ontologies and semantic data aretoo much overhead; we deal with tiny devices in IoT. Reality: Ontologies are a way to share and agree on a common vocabularyand knowledge; at the same time there are machine-interpretable andrepresented in interoperable and re-usable forms; You don’t necessarily need to add semantic metadata in the source- it could beadded to the data at a later stage (e.g. in a gateway); Legacy applications can ignore it or to be extended to work with it.
  16. 16. Exercises 2: create an ontology for IoT Considering reuse of the existing ontologies (using‘import’ in Protégé) Consider the following concepts in the IoT domain: Resource (sensor, actuator, RFID) Other resources (gateway, directory, server) Service (related to IoT resources; as well as servicelifecycle related information) Systems, subsystems Observation and measurement Relationships among the concepts Link to existing knowledge (location)
  17. 17. Ontology matching for improvinginteroperability Also known as ontology alignment or ontologymapping. Formally, is the process of determiningcorrespondences between semantically relatedentities from (two) ontologies. A set ofcorrespondences is also called an alignment. Can be used to support various tasks Ontology merging Assisting ontology engineering for humans
  18. 18. A simplified ontology matching task Two ontologies: Os (source) and Od (destination) To establish correspondence between two conceptsCs from Os and Cd from Od: Check equivalence for classes and relations Check similarity if equivalence cannot be confirmed A similarity or confidence value is calculated using some mechanisms No matching Produce report: equivalence, similarity, and those concepts which cannotbe matched This will help us in the ontology engineering process.
  19. 19. Matching algorithm based on lexical andstructural information Two classes are equivalent if: Their URIs are same They are both equivalent to a third class If no equivalent relation found between two classes,then we try to find out if two classes haverelatedness: subclass/superclass/subproperty/superproperty sibling have Common Ancester lexically similar: check two classes’ labels (e.g., edited distancealgorithm)
  20. 20. Pseudo-codeFor 0<i<m (vector c1)For 0<j<n (vector c2)if cheEuqivalence(c1[i], c2[j]) assert equivalent;elseif checkRelatedness(c1[i], c2[j]) assertcheckRelationType (c1[i], c2[j]);End ifEnd ifEnd ForEnd for
  21. 21. Exercise 3: Check the interoperability ofyour model against existing ones. Ontology matching tool: http://localhost:8080/InteropOntologyCheckingTool/ Input ontologies: The IoT ontology developed in exercise 2 The existing ontologies for sensors (SSN), services (OWL-S)and IoT (IoT-A, IoT.est) Discussion: How similar to existing models is your model?
  22. 22. Thank you
  23. 23. Linked Open Data~ 50 Billion Statements
  24. 24. Linked Data as an independent layerin the Internet architectureImages from Stefan Decker,; linked data diagram:
  25. 25. Linked data and interoperability Linked Data is becoming an accepted best practiceto exchange information in an interoperable andreusable fashion. Many different communities on the Internet useLinked Data standards to provide and exchangeinteroperable information. We have seen methods mainly for improvinginteroperability at ontology (schema) level, now welook at interoperability at data level.
  26. 26. Building interoperability Metadata standards: Dublin core, FOAF, SSN and IoT.est (domain specific) Existing vocabularies: NCI, SSN-QU Other knowledge base and ontologies DBPedia, Geonames Relationships: SKOS closeMatch, exactMatch, broadMatch,narrowMatch, relatedMatch owl:sameAs, rdf:seeAlso
  27. 27. Linked Data and interoperabilitybased on links “The Web of data proposes a style ofinteroperability which doesnt rely on synchronousquery of separate databases, nor on reducingdatabases into a common format, but on thecreation of a global information space, using links tobrowse seamlessly between resources.”Emmanuelle Bermes, "Convergence and Interoperability: a Linked Data perspective"
  28. 28. Linked data principles using URI’s as names for things: Everything isaddressed using unique URI’s. using HTTP URI’s to enable people to look up thosenames: All the URI’s are accessible via HTTPinterfaces. provide useful RDF information related to URI’sthat are looked up by machine or people; including RDF statements that link to other URI’s toenable discovery of other related concepts of theWeb of Data: The URI’s are linked to other URI’s.
  29. 29. Linked data in IoT Using URI’s as names for things;- URI’s for naming IoT resources and data (and also streaming channels anddata); Using HTTP URI’s to enable people to look up those names;- Web-level access to low level sensor data and real world resource descriptions(gateway and middleware solutions); Providing useful RDF information related to URI’s that arelooked up by machine or people;- publishing semantically enriched resource and data description: temporal,spatial, thematic; Including RDF statements that link to other URI’s to enablediscovery of other related things of the web of data;- linking and associating the real world data to the existing data on the Web;
  30. 30. Creating and using linked sensor data
  31. 31. Sensor discovery using linked sensor data