ISA Project

Indoor Spatial Theory - Project Scope and Tasks

Principle Investigator Dr. Mike Worboys.

A fundamental theory of indoor space will be developed. This theory will be hybrid, in that it allows locations in the space to be specified by levels (floor number), cells (rooms, portions of corridors, etc.), and also in some cases (x,y,z)-coordinates in Euclidean space. The theory will determine the types of such locations (specified by dimension, for example, points, linear locations, regions) and their spatial relationships. The theory will also be developed to include the types and behaviors of objects that will populate this space (e.g., indoor moving objects). The theory will generate an ontology, specified in the W3C Web ontology language OWL and also a spatial model, specified in terms of a bigraphical reactive systems model for representing and reasoning about mobile locality and connectivity. Our initial Indoor Spatial Theory will be enhanced by the addition of advanced topological and visualization features, providing support for more accurate representation and reasoning about the properties and behaviors of mobile objects in indoor ubiquitous computing environments.

Project Tasks

The results of tasks 1-3 are presented in the 2008 Project Report and supplemental presentations. A maturation of the models from tasks 1-3 and the results for tasks 4-5 are presented in the 2009 Project Report. Tasks 6-7 will be the focus of efforts later in 2009 and 2010.

1. An investigation into the fundamental nature of indoor space, and a comparison of its properties with conventional, outdoor geo-space. In particular, the investigation will take into account constraints imposed by the sensing technologies and building properties (rooms, corridors, floor levels, etc.)
2. Formal representation of indoor space. The exact format of representation is dependent on the outcomes of task (1), but we can envision a representation as a graph embedded in 3-space. The nodes of the graph represent the cells (e.g., rooms, corridor portions, stairwell-portions) and may be expanded to incorporate more detailed properties of the cells. The labeled edges of the graph represent different types of relationship between cells (adjacency, accessibility, above/below). This graph will be used to show the kinds of reasoning that can be done at the formal level on the space.
3. Formal representation of objects that populate the spaces discussed in (2), including their behaviors and linkages to the space. For example, we may assume that a cell is a part of space for which entry and exit by a moving object is revealed by sensors. We thus assume that it is known exactly which objects are in which cell at any point in time. A cell may be equipped with additional presence sensors. These as well as the entry and exit sensors may be used for inferring more detailed location information of a moving object within a cell. This component of the project assumes a dynamic environment in which a spatio-temporal model is required.
4. Construction of an ontology of indoor space, using the W3C Web ontology language OWL. An ontology editor such as Protege will be used as a tool in this construction. The ontology will include the entities that make up the structure of the indoor space, their properties and relationships. It will also include the entities that populate indoor spaces, their properties, movement characteristics, and relationships to other objects. Particular attention will be given to topological relationships both in the structure and the objects that populate the space.
5. Construction of an Indoor Spatial Theory consisting of elements from tasks 2-4 that provides reasoning and visualization support for modeling indoor spaces. The formal model and theory will be described in a series of reports and papers submitted to peer reviewed conferences and journals.
6. Enhancement of the Indoor Spatial Theory to support a richer set of spatio-temporal relations. For example, if the original theory supports modeling mobile object locality only in terms of containment (e.g., the agent is in the room), the addition of the adjacency relation would support more accurate reasoning about the paths that mobile objects can take when moving through an indoor space (e.g., the agent in the room moves to the next room.)
7. Enhancement of the Indoor Spatial Theory to support a richer visualization of spatio-temporal relations and mobile object behaviours through dynamic image schemas. For example, if the original theory supports modeling mobile object locality only in terms of static positions (e.g., the agent is in the room), the addition of the dynamic spatio-temporal schema PATH would support more accurate visualization of the routes that mobile objects can take when moving through an indoor space (e.g., the agent in the room moves through the building to the exit.)