Building ontology for fire
emergency planning and support
Noel Tay Nuo Wi1, Naoyuki Kubota1, János Botzheim1
1Tokyo Metropolitan University, 6 Chome-5-4 Asahigaoka, Hino, Tokyo 191-0065, Japan
Abstract
Risk management and maintenance operation needs planning on emergency occurrences like a fire, which
require information on building state. During fire emergencies, parts of a building may have higher risks
compared to other parts of the building. This distribution of risks depends on the location of the fire, placement
of hazardous materials in the building, condition of doors and windows around the building, vents and so on.
Before getting to safety, one needs to know where safe places are, and how to approach them safely. This can be
difficult in a large building complex. As the aforementioned attributes can be seen as objects and classes, they
can be assigned their semantic meanings. This includes the description of the rooms, and their relationships with
every other room, such as their adjacency. The building can be modeled by OWL description logic, which
define the building as a graph, where objects are classified into classes, with properties to connect between them.
The generated graph via description logic can be used for further inference for more data and for querying. With
this approach, this work aims to build the building ontology to provide support during fire emergency. Case
studies are done to infer smoke propagation and determination of escape outlet, which can provide the basis for
additional query and inference.
Keywords
Ontology, Description logic, fire emergency, knowledge representation
Full Paper:PDF
EJAM Vol.8 No.2 pp.13-22 "Building Ontology for Fire Emergency Planning and Support"