Article Preview
Top1. Introduction
Emergency management (EM) is the continuous process by which individuals, groups, and communities manage hazards to avoid or mitigate the impact of disasters resulting from them. Emergency response is triggered by the occurrence of geospatially localized adverse events or incidents of a very diverse nature and that typically chain or combine with others as the situation evolves (Fan and Zlatanova, 2010). As we have seen in recent events, a global approach is required in order to achieve as complete an awareness as possible of the situation to assess the possible courses of action for the unfolding emergency. Such a global approach calls for models of different domains in which it is possible to connect incidents with resources and means, that are able to provide support for situation awareness during the course of the incident. The links between these aspects are determined by the potential effects that some kinds of incidents are known to have in particular contexts. For example, certain kinds of fires have different behaviors depending on the infrastructure or area affected, and consequently the means to deal with them are also different.
The complexity of emergency management requires models covering different aspects that are part of the situation triggered by the incident. A survey of situation awareness ontologies can be found in Baumgartner & Retschitzegger (2006). There are different approaches to deploying an Emergency Management (EM) concept. Di Maio (2007) - based on Rex Brooks (OASIS - http://www.oasis-open.org/committees/emergency/) – distinguishes between Emergency Response Management Systems (ERMS) and Emergency Response System Management (ERSM). In addition to this, there are some other proposals where ontologies are used in different ways in Emergency Management Systems, for instance Malizia et al. (2010) define an ontology to manage alerts called SEMA4A. Zhong et al. (2017) focus on the use of ontologies for geospatial information. Authors like Othman and Beydoun (2010) found references in major models for the concept of adverse event or incident, identifying them as the triggers of coordinated actions.
There exist several adverse event taxonomies, but they differ significantly and the ontological criteria they build on in the breakdown into event categories is in most cases not made explicit. This is hampering their extension and reuse. Also, these taxonomies are not associated to the kinds of actions and resources that are known to be effective in handling them, which represents a critical kind of expert knowledge in unfolding situations. Some other authors like Truptil et al. (2008) and Ford et al. (2015) first proposed a mediation tool dedicated to support interoperability, but this approach does not replace the need for a shared meta-model.
Several proposed ontologies for EM have included incidents in their structures. However, the relation of incidents to particular kinds of other elements has not been subject to formalization in most of them (Fan and Zlatanova, 2010; Galton and Worboys, 2011). However, Sicilia and Santos (2009) described BFiaO, which provides a model covering the common high-level aspects of infrastructures (electrical, gas, etc.), causes of incidents (be them natural or caused by humans) and the incidents themselves as situations that in some cases evolve or are qualified as emergencies. Rules could then be used to combine the structure of events representing the evolving situation to infer potential situations of risk associated to infrastructures. The model proposed here, as an extension to BFiaO, allows for the codification of EM knowledge in a reusable way that could later be used for the planning of actions accounting for available resources and other contextual circumstances.