This paper describes an ontology created for the RESCUER1 (Reliable and Smart Crowdsourcing Solution for Emergency and Crisis Management), a project funded by the European Union and the Brazilian Ministry of Science, Technology and Innovation. RESCUER uses crowdsourcing information for supporting Industrial Parks (InPa) and Security Forces during an emergency situation. The proposal, EDXL-RESCUER ontology, is based on EDXL (Emergency Data Exchange Language), and it aims to be the RESCUER conceptual model related to the coordinating and exchanging of information with legacy systems. The ontology was evaluated with end-users during a workshop and the results show that EDXL-RESCUER is adequate for Emergency and Crisis domain in InPa and Security forces contexts. Specifically, this paper presents an update of EDXL-RESCUER ontology based on a faceted taxonomy approach. Resumo. Este artigo descreve uma ontologia criada para o RESCUER (Reliable and Smart Crowdsourcing Solution for Emergency and Crisis Management), um projeto patrocinado pela União Européia e pelo Ministério de Ciência, Tecnologia e Inovação do Brasil. O RESCUER usa informação do público para apoiar Parques Industriais e Forças de Segurança durante uma emergência. A ontologia proposta, EDXL-RESCUER, é baseada no EDXL (Emergency Data Exchange Language) e pretende ser o modelo conceitual do RESCUER relacionado à coordenação e troca de informação com os sistemas legados. A ontologia foi avaliada com usuários finais durante um workshop, e os resultados mostram que EDXL-RESCUER é adequada para o domínio de Crises e Emergências nos contextos de Parques Industrias e Forças de Segurança. Especificamente, este artigo apresenta uma atualização da EDXL-RESCUER baseada em uma abordagem de taxinomia facetada.
Crowdsourcing information (information that comes from different sources:
people affected by the incident, eyewitnesses, security forces and others) is
becoming widely used as a source of knowledge and solutions for different problems
[
The RESCUER project encompasses four main components as shown in Figure 1: Mobile Crowdsourcing Solution: support eyewitnesses communication with official first responders (police, firefighters, etc.) and command and control centers. The crowd can send information in text, image and video formats. It comprises a set of mobile applications tailored to different platforms and devices; Data Analysis Solutions: composed of the algorithms that will process and filter the data in order to extract the required information; Communication Infrastructure: offers the needed equipment in order to allow the information to flow between the stakeholders; and Emergency Response Toolkit: a set of solutions to manage the analyzed crowdsourcing information and to present them to the command and control center using adequate visualization metaphors.
The InPa (Industrial Park) Brazilian partner is the
From this perspective, the use of a well-referenced standard by the scientific
community, the EDXL [
The Figure 2 shows the ontology-based integration module (green box) in RESCUER architecture, specifically in its interaction with ERTK. This module will have three main parts: 1) the EDXL-RESCUER that works like a global ontology; 2) the database schemas from ERTK and legacy systems involved; 3) mappings between the ontology and the local schemas.
The evaluation of EDXL-RESCUER ontology was performed in two steps:
1. Validation through competency questions - questions that an ontology should
be able to answer. This validation is based on a well know method
in Ontology Engineering (for further information see the TOronto
Virtual Enterprise (TOVE) [
[
This paper is structured as follows: in the Related Works section, research projects related to emergency, ontology and interoperability are presented. Next, an EDXLRESCUER ontology and its update based on faceted taxonomy approach is presented. In the Evaluation section the workshop with end-users is described and the results derived are presented; finally, a conclusion of the work done and future developments are presented.
Ontology has been used on several domains in order to solve interoperability
problems, including emergency and crisis domain [
Based on review of related literature, one project stood out: the DISASTER (Data
Interoperability Solution at Stakeholders Emergency Reaction) project [
In addition to the DISASTER project, several works that use ontologies and EDXL in the context of Emergency and Crisis Management were found. Some of these are presented in this section.
The IC.NET (Incident Command NET) is a system that can be used for Emergency
Services such as incident representation, triage, and more. It is based on EDXL-DE as
a top level loose coupler used for delivery and exposure of operational level Emergency
Services / First Responder data [
The TRIDEC3 project is based on the GITEWS (German Indonesian Tsunami
Early Warning System) and the DEWS (Distant Early Warning System). It provides a
service platform for both sensor integration and warning dissemination. Warning messages
are compiled and transmitted in the OASIS Common Alerting Protocol (EDXL-CAP)
together with addressing information defined via the OASIS Emergency Data Exchange
Language - Distribution Element (EDXL-DE) [
WebPuff is a system sponsored by the U.S. Army CMA(Chemical Material Activity)and developed by IEM, a security consulting firm based in North Carolina’s Research Triangle Park. WebPuff provides users at CSEPP (Chemical Stockpile Emergency Preparedness Program) sites with a suite of planning and response tools that are integrated 3Project Collaborative, Complex and Critical Decision-Support in Evolving Crises with a unique chemical dispersion model that provides an advanced level of science on which decisions about public protection can be based.
In order to ensure interoperability with civilian jurisdictions, the system uses the
Emergency Data eXchange Language (EDXL) Common Alerting Protocol (CAP)
developed by the Organization for the Advancement of Structured Information Standards
(OASIS) [
The German Research Centre for Geosciences developed a model for integrating
the national tsunami warning system on a large scale. They proposed a system based on
existing protocols such as EDXL Common Alert Protocol (EDXL-CAP) and the
Distribution Element (EDXL-DE) [
EDXL is a set of packages of XML-based messaging standards that favor emergency
information sharing between organizations and systems. EDXL standardizes messaging
formats for communications between these parties. It was developed by OASIS
(Organization for the Advancement of Structured Information Standards) [
EDXL is a broad enterprise to generate an integrated framework for a wide
range of emergency data exchange standards. The EDXL has several packages:
EDXLDE (Distribution Element); EDXL-RM (Resource Messaging); EDXL-SitRep (Situation
Reporting); EDXL-HAVE (Hospital Availability Exchange); EDXL-TEP (Tracking of
Emergency Patients); EDXL-CAP (Common Alerting Protocol) and EDXL-RIM
(Reference Information Model) [
An ontology for the semantic integration of data exchange between the RESCUER platform and legacy systems has been defined based on EDXL standards. The current version of EDXL has seven (7) packages and covers a full range of message contexts in an emergency. The extended scope of EDXL has raised several questions, including: (i) Should an ontology be constructed for all packages? (ii) What message contexts are important for RESCUER? (iii) What kind of information will be exchanged with legacy systems?
In order to clear up these doubts, other RESCUER documents related to Requisites and Architecture tasks were analyzed. They were chosen because they provide useful information that can be used in semantic integration of RESCUER with legacy systems. Based upon this study, a list of competency questions can be designed, which serve as a basis for the selection of EDXL packages for RESCUER domain.
Therefore, in order to address these questions, four packages were chosen:
EDXLDE, EDXL-RM, EDXL-SitRep and EDXL-CAP. Four new ontologies were created, one
for each chosen package. These were based on ERM and Data Dictionary of their
associated standard. These four ontologies comprise the EDXL-RESCUER ontology
and the formalization of the first version of them can be found in our previous paper
[
With this first version of the ontology, a validation through competency questions, where each competency question is related with the correspondent ontology elements can be performed, as seen in (Table 1). In this way, the selection of EDXL packages can be validated. This validation also contributes to a first step of evaluation of the ontology.
Where was the incident?
In order to update the EDXL-RESCUER [
In line with this approach, in [
b) Entity listing: we listed all entities found; c) Facet creation: we arranged all entities that resembled under a main entity, the facet (main entity was chosen to represent a domain segment EDXL); d) Facet arrangement: we made sure that the entities resembled to the associated facets, reorganizing them when appropriate, (the checks were made through the EDXL documentation, which contains the description and data model for the entities). e) the citation order and f) classification – phases that refer to how the taxonomy would be implemented. In our case, the goal was the creation of an ontology, then we defined what every element under a facet and the facet itself would be in an OWL ontology, i.e. what is a class, sub-class, object property, and data property. g) The last phase included revision, testing, and maintenance: the result of this phase is EDXL-Rescuer v2.
Figure 3 summarizes the entire process of the EDXL-RESCUER update. The first version of the ontologies that composed the EDXL-RESCUER relied on EDXL documentation and the ERM models available there. Hence, a faceted taxonomy based on the same documentation, which allowed one to better detail the domain of each chosen pattern was created. Moreover, we were able: (i) to determine the main concepts with higher precision; and (ii) to use the results for reviewing and revalidate the ontologies created at the first iteration.
For instance, the concepts Severity, Urgency and Certainty found in EDXL-CAP. After the procedure previously mentioned (the concepts reviewing), those concepts became classes instead of DataProperties. Those classes received sub-classes with the ability to have different values according to the EDXL documentation as can be seen in Figure 4.
Another improvement from the previous version is that we were able to reuse common concepts among more than one type of EDXL pattern. Hence, Severity, Urgency and Certainty, which EDXL-SitRep also employs, they are imported concepts from EDXL-CAP; therefore the URI is the same as found on the original ontology.
The approach based on faceted taxonomy seemed to be adequate, considering that
this technique for classifying concepts is characterized by randomly choosing the terms
that represent concepts within a domain (facets). Furthermore, it chooses the relationship
between other domain terms and the terms previously chosen, creating categories (each of
which is related with a facet). Finally, the faceted approach selects the terms and the
relationship between them within the same category or between categories [
Due to space limitation, only part of the EDXL-RESCUER ontology is shown. The concepts that make up the EDXL-CAP Ontology and their definitions are:
Info: Refers to all component parts of the info sub-element of the alert message. Resource: Necessary element to deal with an emergency. A Resource contains information about its Identity, Description and Status.
Incident: Term referring to occurrences of any scale that may require some form of Emergency Response and Management, and that requires tracking and information exchange.
ResponseType: Refers to the type of action recommended for the target audience. Area: Refers to all component parts of the area sub element of the info sub element of the alert message.
Category: Refers to the category of the subject event of the alert message MsgType: Refers to the nature of the alert message.
Status: Refers to the appropriate handling of the alert message.
Scope: Refers to the intended distribution of the alert message.
Sender: The originator of an alert.
Certainty: The certainty of the subject event of the alert message Severity: The severity of the subject incident or event.
Urgency: The urgency of the subject event of the alert message
Zero or more objects of <Concept1> <Relationship> with <Restriction> objects of <Concept2>.
Where <Restriction> can be some, all, Max 1, Min 0, Exactly 1. Min 0 is the default value.
Some axioms have also been defined, for instance: (i) Private, Public and Restricted - subclasses of Scope – are disjoint concepts; (ii) Actual, Draft, Exercise, System and Test – subclasses of Status - are disjoint concepts too.
4.1. Goals The evaluation occurred during the RESCUER Brazilian Consortium Meeting on July 21-23, 2014 and had the goal of validating the terms with potential RESCUER users in Brazilian side. Next, the Goals, Method and Results of this evaluation will be presented.
To present some ontology terms to the stakeholders - terms which were chosen because they represent the main classes of the selected EDXL packages and were the most controversial for both industrial parks (InPa) and large-scale events (LSE); To match those terms with the vocabulary the stakeholders use on a daily basis in order to extract synonyms and verify differences, if differences exist, between InPa and LSE.
The “brainstorm technique” was used in order to capture stakeholder feedback concerning the ontology terms.
Large-scale events (CICC)
During this session, the EDXL concepts were shown to the experts and they tried to find synonyms or correlated terms used in their contexts. At the end of the session, there was an open discussion about the findings related to main concepts of EDXL-RESCUER ontology.
Based on the activity conducted with the stakeholders, it can be deduced: The concepts related to EDXL-SitRep package, in the COFIC context, were suitable; Some concepts, for instance the term “incident”, had minor variations between the two groups; Almost all EDXL terms had related instances or synonyms according to this activity.
The exception was the term “Jurisdiction”, which did not have an instance or a synonym for COFIC. However, at CICC, was found a related instance. Some collected terms can be used as instances for populating the EDXLRESCUER ontology in the future.
A deep investigation of the differences between industrial parks and large-scale events in Brazil; and The need to replicate this activity in the European scenario
It is important to note that the differences between the scenarios (COFIC and CICC) emphasize the need for an Interlingua and the relevance of this proposal - EDXLRESCUER as a common basis for communication.
This paper discuss the conceptual model for semantic integration – EDXL-RESCUER ontology. It aims to integrate, semantically, the RESCUER system with legacy systems. In particular, this paper presents an updated version of the ontologies that composed EDXLRESCUER based on a faceted taxonomy approach. This approach relied on a bottom-up analysis of the EDXL documentation in order to synthesizing the subject statements of these documents. It is important to note that the construction of facets provides different perspectives and views of the domain. In this way, we were able to review our first version of EDXL-RESCUER ontology and adjust its concepts and relationships.
Moreover, in regards to the evaluation, the legacy systems information and data are still missing, as well as the data from RESCUER base. After populating the EDXLRESCUER ontology, we are going to validate it using reasoning algorithms and queries. Another step is to implement the ontology-based integration module between RESCUER and legacy systems.
Some further investigations will be carried out as well: (i) the use of LOD (Linked Open Data) in this context; (ii) the use of the EMERGEL-knowledge base as an additional controlled vocabulary or just as a synonym-base.
This research is part of the RESCUER (Reliable and Smart Crowdsourcing Solution for Emergency and Crisis Management) project funded by European Union under grant reference 614154 and by CNPq under grant reference 490084/2013-3. This work also was partially supported by the National Institute of Science and Technology for Software Engineering (INES), funded by CNPq, grant 573964/2008-4.