Information Technology Reference
In-Depth Information
Behavior-Based Indoor Navigation
Julio Abascal, Elena Lazkano, and Basilio Sierra
Euskal Herriko Unibertsitatea - The University of the Basque Country, Spain
julio.abascal@si.ehu.es
,
{
ccplaore,ccpsiarb
}
@sc.ehu.es
1
Introduction
Ambience provides large amounts of heterogeneous data that can be used for
diverse purposes, including indoor navigation in semi-structured environments.
Indoor navigation is a very active research field due to its large number of possible
applications: mobile guides for museums or other public buildings [36], oce post
delivering, assistance to people with disabilities and elderly people [34], etc.
The idea of using indoor navigation techniques to develop mobile guides is not
new. Among the pioneers, Polly, a mobile robot acting as a guide for the MIT AI
Lab [35], and Minerva, an autonomous guide developed for the National Museum
of American History in Washington [69], are well known. A particular case are
mobile guides for blind people which experienced a notable interest in the last
years [40]. Another interesting application field is devoted to smart wheelchairs,
which are provided with navigation aids for people with severe motor restrictions
[64, 75]. All these applications share the need for a navigation system, even if its
implementation may be different for each of them. For instance, the navigation
system may act over the power stage of a smart wheelchair or may communicate
with the user interface of a mobile navigation assistant in a museum. Evidently
the implication of the user is different in each system, leading to diverse levels
of human-system integration. Therefore, there are two key issues in the design
of mobile guides: navigation strategy and user interface. Even if most of the
mentioned systems use maps for navigation [36], there exist alternative, behavior-
based systems, that use a procedural way to represent knowledge. Therefore, the
selection of the approach not only conditions the navigational architecture but
also the design of the human interface.
This chapter analyzes alternatives for navigation models and focuses on how
properties of the environment can be intelligently exploited for indoor naviga-
tion tasks. In addition, it describes, in detail, an illustrative example based on
behavior decomposition. Its navigational characteristics and influence upon the
human interface design are also discussed.
1.1
Data for Navigation
All agent-environment interaction systems rely on the data obtained from sen-
sors, and have to cope with their quantity and diversity. Sensors are needed to
perceive the environment which is tightly coupled to the agent. In order to obtain
Search WWH ::
Custom Search