Core Incident Model Engineering Reports | OGC

Copyright © 2015 Open Geospatial Consortium. 15 of core ontologies spatial, temporal, identifier, event, SKOS to provide a coherent model for geospatial reasoning on incidents. Figure 4 gives an overview of the Incident Ontology Model. Incidents are based on the Event ontology defined in the Core Geospatial Ontologies. Figure 4: Core incident ontology model The core Incident Model can be extended using the built-in mechanism in OWL to add specific information used within an application domain or jurisdiction, thus becoming specialized profile of the core Incident model. Existing standards based on syntactic and schematic standards NIEM, EDXL, NDEX,… can be integrated in Semantic Incident Management Systems by the use of adapters converting the standard to an RDF representation using the different incident profiles see Figure 5. The benefit of this approach is to provide a unified semantic representation of incidents that can integrate different systems without breaking incompatibilities. The knowledge-based model allows machines to interpret information without ambiguities, perform triaging, reasoning and fusion of information. The system will assist the user by reducing the cognitive burden to make sense of the information. The common knowledge-based representation of the incidents can be used to convert back the information to existing data-centric standard. The semantic layer introduces a layer of ‘smartness’ in the existing IMS that can adapt easily to future change in the model. 16 Copyright © 2015 Open Geospatial Consortium. Figure 5: Semantic layer with adapters to the core incident model and derived profiles

8.3 Incident model and data for demonstration

During the Testbed, we were unable to get datasets for incidents at national level that use HSWG or Canadian EMS taxonomies and could exercise our core incident ontology. We recommend strongly that in future Testbeds, datasets are provided by sponsors or OGC at the start of the testbeds to be able to run scenarios. As a fallback, close to the end of the testbed, we used open-source data from the San Francisco Police Department SFPD https:data.sfgov.orgPublic-SafetySFPD-Incidents-from-1-January-2003tmnf-yvry , as the demonstration was occurring in the San Francisco Bay. Copyright © 2015 Open Geospatial Consortium. 17 Figure 6: SFPD Incident samples from SF OpenData The data were exported in CSV format and the values of category of the incident were replaced with concepts from the HSWG EMS taxonomy in order to leverage the SKOS encoding of HSWG. We defined intentionally two different incident models to exercise the semantic mediation of taxonomies and ontologies. Here a sample of a HSWG Incident using the HSWG EMS Taxonomy. prefix ks: http:www.usersmarts.comont200506ks . prefix spatial: http:www.opengis.netontspatial . prefix hswg: http:www.opengis.nettestbed11ontincidenthswg . prefix rdfs: http:www.w3.org200001rdf-schema . prefix geosparql: http:www.opengis.netontgeosparql . prefix time: http:www.knowledgesmarts.comontologytime . prefix evt: http:www.knowledgesmarts.comontologiesevent . prefix xsd: http:www.w3.org2001XMLSchema . prefix owl: http:www.w3.org200207owl . prefix wgs84: http:www.w3.org200301geowgs84_pos . prefix place: http:www.knowledgesmarts.comontologiesplace . prefix skos: http:www.w3.org200402skoscore . prefix evt-type: http:www.smartrealm.comcttypesevents . prefix ptype: http:www.knowledgesmarts.comontologiesplacetypes . prefix incident: http:www.opengis.netontdomainemergencypoliceincident . http:ows.usersmarts.comldappows11demoemssfpdincidents11616059406244 a hswg:HSWGIncident ; hswg:hasAddress [ a hswg:Address ;