Poems

4 pages
9 views

SHARE-ODS: An Ontology Data Service for Search and Rescue Operations

of 4
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Share
Description
SHARE-ODS: An Ontology Data Service for Search and Rescue Operations
Transcript
  SHARE-ODS: An Ontology Data Servicefor Search and Rescue Operations Stasinos Konstantopoulos, Georgios Paliouras, and Symeon Chatzinotas Institute of Informatics and Telecommunications,NCSR ‘Demokritos’, Ag. Paraskevi 153 10, Athens, Greece { konstant, paliourg, schatzin } @iit.demokritos.gr Abstract. This paper describes an ontology data service (ODS) forsupporting Search and Rescue (SaR) operations. The ontological modelrepresents various aspects of the command, communication, and organ-isational structure of the SaR forces and the deployment and progressof a SaR operation. Furthermore, the ontology supports the semanticindexing of multimedia documents in the context of SaR processes andactivities. This ODS supports a semantically-enhanced information andcommunication system for SaR forces. Modelling the spatio-temporal as-pects of an operation in alignment with possibly-unreliable informationautomatically extracted from multimedia objects, introduces a numberof challenges for the field of knowledge representation and reasoning. 1 Introduction Search-and-rescue (SaR) operations are conducted by fire-brigade, rescue andmedical units, operating under a complex unified command-and-communicationsstructure. The communication channels of the emergency units are push-to-talkwalkie-talkies and short hand-written message forms read over the radio. Allstatus information necessary to decision making is processed manually.The SHARE project 1 develops a Push-To-Share (PTS) advanced mobile ser-vice that provides communication support for emergency teams during SaR op-erations. We present here the SHARE Ontology Data Service (SHARE-ODS),which supports the PTS service with (quantitatively but also qualitatively) en-hanced information, necessary for the decision-making process at all commandlevels of the operation. Furthermore, the information stored in the ODS servesas a complete log of the operation for the purposes of planning and evaluation.Although several multimedia semantic modelling and spatio-temporal mod-elling ontologies have been proposed, there is no unifying approach of the two.Here we propose a model for the semantic indexing of multimedia objects in thecontext of processes and activities. This model not only unifies these two aspectsof a SaR operation, but it also allows for the semantic cross-checking of possibly-unreliable information automatically extracted from multimedia objects. 1 SHARE: Mobile Support for Rescue Forces, Integrating Multiple Modes of Interac-tion, IST-funded project, URL: http://www.ist-share.org/ G. Antoniou et al. (Eds.): SETN 2006, LNAI 3955, pp. 525–528,2006.c  Springer-Verlag Berlin Heidelberg 2006  526 S. Konstantopoulos, G. Paliouras, and S. Chatzinotas Due to space limitations, the overview of the state-of-the-art and the descrip-tion of SHARE-ODS is very concise. For a fuller overview and SHARE-ODSdescription, please see technical report DEMO-2006-1. 2 2 Related Ontologies Ontological resources that are relevant to our SaR ontology are both generalpurpose ontologies and related domain ontologies.Various general-purpose space, time, and spatio-temporal ontologies havebeen proposed, generally speaking as part of a more complete concept and inter-face specification that aims to enhance interoperability between databases andapplications that make geographic and temporal references. Most prominently: – The Standard Upper Ontology Working Group of IEEE. – ISO geo-reference standards 19107:2003, 19115:2003 and 14825:2004. ISO19108:2002 standardises temporal characteristics of geographic information. – The OpenGIS specification of the Open Geospatial Consortium 3 . – Temporal ontologies, e.g. OWL-Time 4 and OWL-S 5 .Domain ontologies and task ontologies describe the vocabulary for a genericdomain, task, or activity by means of specialised terms. They can be used as abasis which can be expanded and specified into an application-specific ontology.Ontologies and projects that are related to the SHARE ontology include: – The OntoWeb 6 and KnowledgeWeb 7 thematic networks, developing stan-dards, infrastructure and ontologies for semantic annotation. – The aceMedia project 8 on knowledge discovery from multimedia data andontology-driven meta-data extraction. – The Enterprise Ontology 9 , capturing the structure and processes of largecorporations with complex structures and business plans. – The CoSAR-TS project, researching the semantic modelling of military SaRoperations. CoSAR-TS is based on the < I-N-OVA > model. 10 3 The Search-and-Rescue Ontology Data Service Search-and-rescue operations (SaR) are conducted by fire-brigade, rescue andmedical units, operating under a unified command-and-communications struc-ture. Emergency forces use half-duplex channel walkie-talkie technology for sim-ple push-to-talk voice communication. Furthermore they exchange hand-written 2 http://www.iit.demokritos.gr/ ∼ konstant/dload/Pubs/demo-2006-1.pdf 3 http://www.opengeospatial.org/ 4 http://www.isi.edu/ ∼ pan/OWL-Time.html 5 http://www.daml.org/services/owl-s/ 6 http://www.ontoweb.org/ 7 http://knowledgeweb.semanticweb.org/ 8 http://www.acemedia.org/ 9 http://www.aiai.ed.ac.uk/project/enterprise/enterprise/ontology.html 10 http://www.aiai.ed.ac.uk/project/ { cosar-ts/index.html,oplan/inova.html }  SHARE-ODS: An Ontology Data Service for SaR Operations 527 message forms that are typically read over the radio. All status information,reporting and documentation for decision making is processed manually.SHARE proposes replacing walkie-talkies and written message forms with apush-to-share (PTS) system that supports the transmission of audio, video anddigital message forms. The new system integrates the PTS communications sys-tem with the ontologically-indexed data service, supporting the decision-makingprocess by making all relevant information and documents easily retrievable bymeans of semantic indexing and searching. 3.1 SaR Operations Emergency units participating in SaR operations, operate under a unified com-mand and communications structure and are deployed in sections (B-Level de-ployments) and subsections (C-Level deployments). Deployments have threeaspects: (a) operational, e.g. fire-fighting, first-aid, water supply, etc., (b) ge-ographical, that is, the area they are responsible for, and (c) operation struc-tural, defining command and communications channels. In addition to the B andC-Level units, the operation establishes (on site) an A-Level command and con-trol centre, which is in charge of the whole operation. 3.2 The Ontology Data Service The SHARE Ontology Data Service (ODS) is an intelligent storage, indexingand retrieval mechanism for (a) meta-data of documents created and trans-mitted during an operation, (b) spatio-temporal information pertaining to theoperation, and (c) information regarding the structure of the operation. Thisinformation is accessed through the ODS interface, which offers functions forpopulating and querying an application-specific ontology through a Web Service(complying to the W3C SOAP 11 messaging recommendation). The web serviceuses Prot´eg´e 12 to manipulate the OWL representation of the ontology, and areasoning engine (currently Jena 13 ) to provide the back-end for the queryingfunctionality, allowing for the retrieval of implicit (inferred) knowledge. 3.3 Ontological Model The ontological model of the operation is organised into three sub-ontologies: SaR, multimedia  and event  . In addition, there are two auxiliary sub-ontologies( time and space ) that represent spatio-temporal references and actual geograph-ical features (buildings, streets, etc) present at the theatre of the operation.The SaR ontology holds the concepts that are related to the Search andRescue operation, the personnel involved, and the communications system. SaRconcepts include deployment types (e.g. A-, B- or C-Level), operational rˆoles andactual personnel, units, vehicles, equipment, etc. Deployments comprise units 11 http://www.w3.org/TR/2003/REC-soap12-part0-20030624/ 12 http://protege.stanford.edu/ 13 http://jena.sourceforge.net/  528 S. Konstantopoulos, G. Paliouras, and S. Chatzinotas and are linked together in a partology, and also connected to operational rˆolestheir require (e.g. deployment leader, dispatcher, etc) which, in their turn, areconnected to the actual personnel members that fulfil each rˆole.The Multimedia conceptual model holds meta-data from all documents (textmessages, audio, video, and infrared video) generated and transmitted during theoperation. This meta-data includes ‘logistical’ information (creation time, sessionthe document was transmitted in, sender and recipient, etc) and content meta-data extracted automatically by image, speech, and text processing systems.Finally, events relate a temporal instance with some characteristic or instanceof the operation. Action events, in particular, relate temporal instances with anagent (e.g. a PTSUser instance), an object (e.g. a PTS Session ), and otheraction-specific properties (e.g., for session participation events, floor time). 4 Future Plans We propose an ontological model that unifies SaR operation modelling withsemantic annotation of documents, to offer an integrated model for an operationand all documents pertaining to it. Furthermore, we are putting together a setof tools for using the ontology at an actual SaR operation. These tools includethe Ontology Data Server for updating and accessing the semantic data and thereasoning facilities that will augment the srcinal data with inferred facts.As a SaR operation unfolds, the ontology gets populated by various sources,some reliable (e.g. GPS) and some not (e.g. information extraction modules).Faulty data can be caught (and, possibly, corrected) when create inconsisten-cies, which can be resolved in favour of the more reliable source. In this man-ner, feedback can also be provided to the module responsible for the error, sothat it can improve its performance over time. In cases where multiple sourcescorroborate towards accepting or rejecting multiple pieces of information, theproblem of deciding which to accept as most reliable (and, inversely, distri-bution of responsibility in order to provide feedback) becomes a non-trivialproblem. This problem has been approached in various domains, but not inthe domain of responsibility distribution among multiple information extractionsources.The other interesting direction we plan to pursue is spatio-temporal repre-sentation and reasoning for the purposes of operation planning and evaluation.At this point, the ontology models only the current situation, with a limitedtemporal Event ontology. In order to represent the operation through time,the Event ontology will be expanded. The temporal information will be usedby a reasoning engine that supports spatio-temporal reasoning. Spatio-temporalreasoning applies to cases where we don’t have precise, quantitative informa-tion about space and time, but only qualitative relationships between instances.Constraint-satisfaction spatio-temporal reasoning can be applied in these casesin order to effectively query the Knowledge Base.
Related Documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x