Article Preview
TopIntroduction
The mission of a country’s Civil Protection is to guarantee the public safety of citizens against any possible threat brought by destructive natural phenomena such as earthquakes, hurricanes and floods. The work of Civil Protection consists of constantly monitoring the areas at greatest risk, and promptly putting in force the relief whenever natural disasters have occurred or are expected to occur. The relief effort needs to be coordinated from the time the hazard is detected to when the normal situation has been restored. Throughout this time, it is of paramount importance that the coordinating team has access to both the real time data pertaining the specific area interested by the natural phenomenon and the data of other surrounding areas, which may potentially be threatened as well. Also, among the multiple sources of information spread out in the territory at risk, it is useful to identify those capable of producing data which may help to monitor the evolution of the specific phenomenon that has occurred. We believe that a prompt and sensible organization of the data “sensed” on the territory is crucial for the planning of the relief effort.
What emerges from these considerations is the need of integrating some different, yet intertwined, knowledge domains, in order to obtain a complete view of the potential environmental risk: the sensors, intended as sensing elements capable of providing real time data useful to measure the development of natural phenomena; the hazards to which people are subjected; the geographic areas to be put under observation in case a hazard occurs or is about to occur; the potential damage that a hazard could cause on a place and its level of dangerousness. Under a semantic perspective, each domain may be represented through an ontology populated by semantic concepts which relate to each other by means of semantic properties. Independently of the semantic representation that may be adopted to represent the domains, they are heterogeneous to each other, in a way that it is hard to find an overlapping area of semantic concepts which are shared by the different domains. The contribution of this work, which extends the one presented in (Calcaterra, Cavallo, Di Modica, & Tomarchio, 2015) is twofold: a) to define the semantic models for the characterisation of the previously cited knowledge domains; b) to specify semantic properties which build inter-domain relationships.
Final objective of the work is to provide an integrated semantic framework where the above-cited domains provide a common information ground on which to build tools and services that are useful to monitor the territory in case of a natural disaster. As mentioned earlier, primary concern of Civil Protection (CP) is the life of citizens who populate territories which are known to be historically exposed to hazards. Specifically, the CP needs to monitor all those densely or sparsely populated territories for which a concrete risk exists. Actually, looking at a specific geographic area, an event (be it natural or atmospheric) may turn into a natural disaster and thus represent a threat to the population only if that area is “sensible” to that particular event. To make an example, heavy rains that unceasingly fall over a restricted area may trigger a flooding if in that area there are lakes or rivers that may overflow, otherwise no flood will materialise. Again, a volcano eruption may represent a serious hazard only in places which are close to either the crater hole or to the lava’s expected flowing path. Therefore, the tools we offer will enable CP to: a) characterise, and draw on a map, places that are known to be under an environmental risk, also evaluating the level of potential damage in case of different kinds of hazard; b) identify on the map the sensors that populate a place and their capability to measure an event which may turn into a natural disaster; c) in case of hazard, query the system to retrieve the sensors that are useful to monitor and provide useful information about the places affected by the hazard.
The remainder of the paper is organized as follows. In Section 2 related work is reviewed. In Section 3 the semantic model is discussed in technical details by describing the proposed ontologies. In Section 4 we present a web tool that grounds on the proposed model to offer a monitoring service. Two simple use cases are also discussed along. Finally, Section 5 concludes the work.