Elaborating a Cognitively Enriched Semantic Conceptual Model for Spatial Data Infrastructures to help Blind Pedestrians Navigate in Urban Areas

blaredsnottyAI and Robotics

Nov 15, 2013 (3 years and 8 months ago)

120 views

Elaborating a Cognitively Enriched Semantic Conceptual
Model for Spatial Data Infrastructures to help Blind
Pedestrians Navigate in Urban Areas


Reda Yaagoubi
1,2
, Geoffrey Edwards
1,2,3
, Mir Abolfazl Mostafavi
1,2

1
Centre de Recherche en Géomatique (CRG)

2
Ce
ntre Interdisciplinaire de Recherche en Réadaptation et Intégration Sociale
(CIRRIS)

3
Canada Research Chair in Cognitive Geomatics

Université Laval, Québec, Québec, G1V 0A6

Reda.yaagoubi.1@ulaval.ca, geoffrey.edwards@scg.ulaval.ca, Mir
-
Abolfazl.Mostafavi@s
cg.ulaval.ca


Abstract

Semantic information about the surrounding space plays a fundamental role in
various tasks of navigation and wayfinding especially for visually impaired
pedestrians. Therefore, a suitable Spatial Data Infrastructure (SDI) for helping
the
navigation of pedestrians who are blind should provide useful and relevant spatial
data semantics to help these individuals to better configure their mental
representation of urban areas. The aim of this paper is to propose a design
methodology of a s
patial semantic database, which is cognitively enriched to help
visually impaired pedestrians in their daily navigation activities. The elaborated
semantic conceptual model ensures defining an SDI dedicated to improve situation
awareness of the blind pedes
trian in urban areas. This semantic model has a
hierarchical structure, hence providing information about the environment at different
levels of detail. In addition, this semantic model can be integrated with the ISO
19133:2005 standard for location based
services to extend the capabilities of this
standard in ways that are more supportive of the needs of the blind.


Keywords:
Spatial Data Infrastructure, Semantic Model, Cognitive Principles,
Navigation, Wayfinding, Hierarchy, Level of detail, visually imp
aired pedestrians.



1.

INTRODUCTION


Spatial information plays a major role in human daily life activities. More
specifically, this kind of information is heavily solicited during activities of navigation
and wayfinding (Kitchin and Jacobson, 1997). Howe
ver, access to such information
is often limited for people with various types of disability. These individuals can live in
situations which handicap them if they are not able to obtain spatial information that
is relevant and useful, in a clear, understan
dable and timely manner.


In order to improve the spatial knowledge of visually impaired individuals
concerning their surrounding environment, it is necessary to offer navigational
assistance that provides to users relevant information about the semantics
of the
space. Spatial semantics refer to the meaning of components of navigational space
and to the spatial configuration that prevail among these components (Rüetschi and
Timpf, 2005). Providing a visually impaired person with appropriate information abou
t
the semantics of the navigational environment allows to this person to improve his or
her ‘situational awareness’ (Yaagoubi and Edwards, 2008). To achieve this goal,
navigational assistance must rely on a spatial database that describes appropriately
the
environmental semantics, while taking into account the cognitive needs of this
category of population during navigation activities.


Yaagoubi and Edwards (2008) propose a cognitively informed approach called
‘Cognitive Design’
. This approach
facilitates t
he definition of a clear and appropriate
set of requirements that helps to select and structure relevant spatial data necessary
to provide an effective navigational assistance for visually impaired users. This
method allows the development of solutions tha
t match the specific cognitive needs
of the users. In the case of the blind pedestrian, the main issue is to provide
“situation awareness”, that is useful spatial information about his or her surrounding
environment. Giving instructions to a blind user abo
ut a planned displacement is of
little use without appropriate information about his or her surrounding environment.


Relevant spatial information must be collected from a variety of sources. So, it is
very important to integrate this data in a suitable an
d homogeneous form. Defining an
appropriate spatial data infrastructure plays a major role in order to achieve this goal.


Spatial Data Infrastructure (SDI) in general aims to coordinate the exchange and
sharing of spatial data between active users within
a spatial data community (Erik de
Man, 2006). SDI can be viewed as an initiative intended to create an environment
that allows users to access and retrieve consistent data sets in an easy and secure
manner (Rajabifard et al, 2000).


According to Rajabifard
and Williamson (2001), an SDI includes five components:
people, access, policies, standards and data. People are an important component in
an SDI, because of their roles in processing data and in decision
-
making. The
relation between people and data is de
termined through network access; policy,
privacy and liability; standards and interoperability (Smith et al, 2004). The following
figure summarizes the relations that hold between SDI components (Rajabifard and
Williamson, 2001):


Figure
1
: Relations between SDI Components (Rajabifard & Williamson, 2001)











The interactions among the component of an SDI are strongly dependent on the
context of use and the user’s needs. Given the specific needs of visually impaired,
each comp
onents of an appropriate SDI have to consider their cognitive
requirements when carrying activities of navigation and wayfinding.



In this paper, we introduce a design methodology for a spatial semantic database,
which has been cognitively enriched, to he
lp the navigation of blind pedestrian in
urban areas. This methodology is focused on the following SDI components: Data,
Standard and People. We highlight the importance of such a cognitively enriched
spatial semantic database to define an appropriate SDI
that helps the navigation of
blind pedestrians. First, we summarize cognitive constraints for a navigational aid for
the blind. Then, we highlight the main semantic information required to meet the
needs of our target users. Next, we present our semantic c
onceptual model
developed by the UML (Unified Modeling Language), that aims to capture the
semantics of various components of the environment and relations among them.
People

Access
Ne
t
wor
k

Sta
n
dards

Policy

Data

Finally, we discuss the relevance of this semantic conceptual model as well as the
possi
bility of its integration with existing platforms such as the standard ISO
19133:2005 for Location Based Services

Tracking and Navigation.


2.

COGNITIVE CONSTRAINTS FOR THE NAVIGATION OF BLIND
PEDESTRIANS


The problem of providing an effective navigation
al aid for blind pedestrians
requires that one understands the mental representations that are developed and
used by these users (Yaagoubi et al, 2009). In general, the mental representation of
space plays a fundamental role in the navigational process (Ju
l and Furnas, 1997). If
this representation is well configured, then the cognitive processing cost associated
with a given navigational task will be reduced.


2.1

Mental Representation of Urban Areas among the Blind


Mental representations are constructed
from our perceptions of space. They
consist of a mental organization of elements that are judged to be characteristic of
the environment and of the qualitative spatial relations that hold among these
elements (Tversky, 2003). Beginning with Lynch (1960), m
ental representations of
urban areas are understood to consist of concepts related to paths, limits, districts,
nodes and landmarks (Figure 2). These concepts can be defined as follow:




Paths are channels along which the pedestrian usually, occasionally or potentially
moves. Other elements of the environment are arranged and linked with regard to
paths
according to a reference frame
(Tversky, 2003)
. This allows
preserving
topological re
lations between mental representation components.



Limits are defined as borders between two areas.



Districts are large areas within the city.



Nodes are strategic points within the urban area that pedestrian can access.
These can be junction points, p
edestrian crossings, points of convergence and so
on.



Landmarks are objects within the urban area that serve as spatial references.
The difference between landmarks and nodes is that landmarks are not
necessarily accessible by pedestrians.


Figure
2
: Components of mental representation of urban areas






Although Lynch’s model has been critiqued and supplanted by more recent
models developed from experiments in spatial cognition, the basic structures he
defined are still useful
and close to those posited by modern psychology (Mark et al,
1999). They provide a useful basis for the configuration of mental representation of
urban environments.


City

Districts

Limits

Landmarks

Paths

Nodes

The principal difference between the mental representations maintained by the
blind compa
red to those of the sighted concerns how the saliency of each element is
determined. For the blind pedestrian, saliency is generally given to non
-
visual cues.
Therefore, the semantic information about the surrounding space is very important
for the blind i
ndividual in order to carry out daily navigational tasks. Thus, it is
necessary to enrich the representation of space for blind users with appropriate and
useful semantic information about the navigation environment.


2.2

Hierarchical Reasoning in Spatial
Navigation


People are surrounded by a huge amount of spatial information; however, it is
unlikely that they store all the spatial data that reaches them. Instead, they store in
memory only that spatial information which is most useful to them. Furthermore
,
many studies in spatial cognition highlight the fact that humans structure spatial data
in an hierarchical manner (Mennis, 2003). Hierarchical spatial reasoning allows
individuals to deduce and store
knowledge in simpler forms. H
ence
, it allows
to
reduce
the amount of information taken into consideration and to make spatial
reasoning easier and more efficient (Timpf, 1999). This hierarchical reasoning is
more effective when the mental representation has been configured appropriately. In
fact, the elaborat
ion of mental representation of space requires both spatial data and
appropriate knowledge structures used in storing and processing such data. There is
evidence that the structuring of mental representations passes through three stages
(Kitchin et al, 199
7):




Declarative structuring: in this first step, the spatial data are stored as a kind of
mental database containing specific features (landmarks, paths, regions, etc.).



Procedural structuring: several rules are used to synthesize the declarative
know
ledge into information that facilitates spatial actions carried out by the
person in his or her environment.



Configural structuring: finally, information such as angles, directions and
distances are combined within the mental representation of space. Thi
s level
constitutes the most elaborate organization of mental representations of space.


Consequently, we are required to develop a spatial semantic database model that
addresses and structures the spatial information data required for the navigation of
bl
ind pedestrians in similar ways. This spatial semantic database has to provide
information at different levels of detail, in ways that can be adapted to and controlled
by the visually impaired. Hence, the interaction of the blind person with the
navigation
al aid should allow him or her to select the appropriate level of detail to use
in assistance. He or she should be able to increase or decrease this level of detail as
a function of the task at hand (Yaagoubi and Edwards, 2008).


2.3 Semantic Information R
equirements to Meet the Needs of Blind Pedestrians


The semantic information required for successful navigation by the blind is
primarily related to two components of the urban space: road intersections and
landmarks (Yaagoubi et al, 2009).


The Semantic I
nformation Content of Road Intersections


Road intersections play a major role in the configuration of mental representation
of urban areas among the blind (Gaunet, 2006). The same author stipulated that
guidance functions for the blind pedestrian in urban
areas should be composed of
two categories; route instructions and environmental information.


For route instructions, we distinguish the following functions: warnings, orientation
with regard to place, orientation with regard to a street, orientation aft
er crossing an
intersection, crosswalk intersection search, noting progress and noting route ending
(Gaunet, 2006). Environmental information includes localization and orientation,
intersection announcements, intersection descriptions and crosswalk
announc
ements (Gaunet, 2006)].

The research led by Bentzen et al. (2004) on blind pedestrian safety shows that
this depends on a number of variables: starting of the crossing, delay on starting of
the crossing and completing crossing the intersection after the be
ginning of
perpendicular traffic.


The Role of Landmarks in the Navigation Process


According to May et al. (2003), landmarks are the most frequently used category
of navigational information. They provide useful information about directions
especially at
key decision points. Landmarks help also to confirm that the pedestrian
has made appropriate choices during previous navigation maneuvers.


Landmarks are objects or structures that mark a location and they are used as
reference points. These landmarks are
required to have singular characteristics with
regard to other nearby objects. In the case of blind pedestrians, the prominence of a
landmark is defined by:




Semantic attractiveness: such as cultural and historical significance.



Structural attractivene
ss: borders, building materials, etc.


These spatial references are encoded in the mental
representation through
(Raubal and Winter, 2002
):




Location: in relation to other structures in the area (roads, buildings, etc.) and
other landmarks.



Identificat
ion: in general, it is based on a unique and singular characteristic of a
landmark.



Description: it is based on the most attractive characteristics of a landmark.


Yaagoubi & al. (2009) propose to represent the influence of a landmark according
to three
types of spaces that are hierarchically structured. These spaces are derived
from the perceptual regions introduced by Reginster & Edwards (2001).
According to
these authors, ‘
the perceptual regions embrace at least three types of space or sub
-
region, whic
h are denoted here as the vista space, the local displacement space, and
the enlarged displacement space
’. Based on this definition, we can define three
hierarchical spaces as follows:


1.

Vista space: this sub
-
region which can be perceived from a single p
oint of
perspective without locomotion.
In the case of visually impaired people, the vista
space is perceived by senses other than vision.

2.

Local displacement
-
reinforcement space: the space surrounding the vista space;
it consists of places and locations
with frequent visits, usually by foot.

3.

Enlarged displacement
-
reinforcement space: this space is a network of local
displacement
-
reinforcement spaces and other activity islands, usually connected
by vehicle displacements.








Figure
3
: The three hierarchical spaces embraced by perceptual regions














Accordingly, the localization and orientation instructions that describe landmarks
near the blind pedestrian will be provided with regards to the three hierarchical
space
s cited above. We believe that this manner will help users to encode landmarks
in hierarchical fashion, which improve their configuration of mental representation of
urban area.


Hierarchical Structure of Semantic Spatial Data for the Navigational Assistan
ce
of Blind Pedestrians


Yaagoubi et al. (2009) proposed a hierarchical data structure adapted to the
problem of assisting navigation for the visually impaired.


This hierarchical structure provides three levels of details that reinforce all three
sub
-
reg
ions within the concept of perceptual regions (see Figure 3). Each level of
details will contain landmarks that correspond to the appropriate sub
-
region. These
three levels of abstraction are:




The ‘Level of Imminent Reinforcement’, the most detailed and the least
abstracted. At this level, the assistance will offer details of the neighborhood of
the blind pedestrian, according to his or her position (using a deictic reference
system
centered o
n the person
).



The ‘Level of Local Reinforcement’, an intermediate level of abstraction. The
information offered at this level is les
s detailed than the first level.

T
he description
of the space will be organized according to the most salient landmarks
(intrinsic
reference system
centered on object of the surrounding space
).



The ‘Level of Enlarged Reinforcement’, with more abstractions and fewer details.
Such a level will serve as a link between various ‘pieces’ of the mental
representation of the urba
n space. This will be use an extrinsic reference system

(eg. North, South, East, West)
.


The following figure summarizes the three levels of the hierarchical structure of
the spatial data used for navigation assistance as well as the reference systems
corr
esponding to each level.





Vista Space

Local
displacement
-
reinfor
cement
space

Enlarged
displacement
-
reinforcement
space



Figure
4
: The three levels of hierarchical structure of spatial data used and the corresponding
reference systems





















3.

THE SEMANTIC CONCEPTUAL MODEL PROPOSED FOR PROVIDING
NAVIGATIO
NAL AID TO THE BLIND PEDESTRIAN


In order to provide effective assistance to blind pedestrians, it is necessary that
data is structured appropriately and semantically enriched. To achieve this, we
introduce a cognitively enriched semantic conceptual model
of spatial data oriented
specifically to support visually impaired pedestrians. Classes and relations of this
semantic conceptual model are developed in the following sections via the Unified
Modeling Language (UML).


3.1

The City and its Main Components


As discussed by Lynch (1960), a city is built across a large geographical area,
which we cannot fully perceive except after long periods of exploration. A city usually
consists of several urban housing areas and contains a set of social, economic and
cultu
ral activities.


From this definition, districts may be modeled as a zonal area characterized by a
name, a size and a well
-
defined form. Boundaries
(Limits)
of districts can be of
various types such as rivers, roads, highways, streets, etc.

However, it can
be noted that there is an important component of mental
representation of the city that does not appear in Lynch’s work. This is the concept of
transition between districts.
Indeed,
although the boundaries between districts are
often fiat, the individual
is aware of his transition between two neighboring districts
.

Edwards and Ligozat (2004) define a transition zone as a zone where views change
quickly but not discontinuously. This concept describes the various structures that
facilitate the transition bet
ween two districts
,
for example, to cross the boun
dary
between these two districts.



Extrinsic
reference system

Intrinsic
reference system


Deictic reference
system

Level of Imminent
Reinforcement


More abstract


More details

Level of Local
Reinforcement


Level of Enlarged
Rei
nforcement


Useful Spatial Data

Figure
5
: Aggregation relations among classes City, District, Transition and Limit




The aggregation relations in
Figure 5 are composites; they imply that Districts,
Transitions and Limits must belong to the class City.


3.2

Pedestrian Paths and their Components


Pedestrian paths are channels through which pedestrians move in urban areas. In
our model, we distinguish
between two basic concepts; ‘Pedestrian path section’ and
‘Pedestrian path segment’. The segment is defined as a continuous part of the path
over which the blind pedestrian does not perceive any significant change in
orientation or direction of displacemen
t. Where there is a change that strongly
enough to notice, it is necessary to define a new segment.




Figure
6
: Aggregation relations among classes District, Pedestrian_Path_section,
Pedestrian_Path_segment and Crosswalk






Typically, one district contains multiple paths, and these paths may pass through
many districts. However, in our model, we assume that a pedestrian path section has
to belong to only one district. Thus, limits of this pedestrian p
ath section will be
defined according to limits of the district. In addition, a pedestrian path section is a
result of an aggregation of a set of adjacent pedestrian path segments, which belong
to the same path, and may include a pedestrian crosswalk (Figu
re 6).



3.3

Intersections and their Associated Components


Research undertaken by Bentzen & al. (2004) show that blind pedestrians run
high risks when crossing intersections. Therefore, it is necessary to provide a rich
semantic description of intersectio
ns and their associated components.


The concept of Intersection corresponds perfectly to the concept of node referred
by Lynch (1960). In our model, we define an intersection as a junction between two
or more streets. Intersections therefore connect three
or more street segments. A
‘Street segment’ is a section of street that serves vehicles, while its boundaries are
always defined by a start intersection feature and end intersection feature. The class
‘Street section’ is an aggregation of a set of segment
s, which belong to the same
street. The limits of such section are defined with regard to the district to which it
belongs (Figure 7).


Figure
7
: Relations among classes Intersection, Intersection_signs, Crosswalk,
Pedestrian_Path_
segment, Street_segment, Street_section and District



Each intersection is usually accompanied by a set of road signs that have to be
observed by pedestrians (Figure 7).
These signs have to be non
-
visual in order to be
perceiv
ed by visually impaired pedestrians.
The crosswalk allows pedestrian to cross
the intersection safely. It also provides a connection between two or more pedestrian
path segments if they are separated by a road or a space reserved for vehicles.


3.4

The Inf
luence of Landmarks and Obstacles on Navigational Activities


The landmark is an important element for navigation because it can be seen as a
spatial reference. It is characterized by its prominence in the surrounding
environment. However, landmarks may ha
ve two opposing effects or forces: they can
have an attracting effect, or a repulsive effect. The attracting effect corresponds to
landmarks that help the pedestrian to know that he or she is in the right location and
on the right path. If the landmark sho
ws that the blind individual is in the wrong
location or on the wrong path, the landmark acts as repulsion.

Obstacles are structures in the environment of navigation that can limit movement
of pedestrians. Albeit Lynch (1960) has not mention this concept a
s an element of
mental representation of urban areas, obstacles have a strong influence on the
success of navigational activities, especially for blind pedestrians. The effects of
these barriers can be either to block the blind pedestrian, so he or she has
to change
the current path, or it is possible for him or her to encounter such an obstacle while
continuing in the same path.


In our model, we present all the effects of landmarks and obstacles in one class
named ‘Force’. This class is used to define the
cognitive effects of landmarks and
obstacles (Figure 8). These effects have a strong influence on decisions to be taken
when performing navigational tasks.


Figure
8
: Modeling the influence of Landmarks and Obstacles on intersecti
ons and pedestrian
path segments





3.5

The Semantic Conceptual Model and the Three Hierarchical Levels


During the design process that led to our semantic model to assist the navigation
of blind pedes
trian, we kept in mind th
e
objective of developing a
hierarchical model.
Hence, our
semantic model
is based on
three levels of abstraction mentioned in
section 2 (Level of Imminent Reinforcement, Level of Local Reinforcement and Level
of Enlarged Reinforcement). The availability o
f information at these three hierarchical
levels helps to improve the configuration of the mental representation among blind
pedestrians.



The next figure shows the semantic conceptual model and the corresponding
hierarchical levels.


Figure
9
: The semantic conceptual model and the three levels of hierarchy




The level of imminent reinforcement corresponds to the heart of our model. It
contains the classes that provide a highly detailed description of the space around
the user. T
hese classes include intersection, crosswalk, Pedestrian_Path_segment,
Intersection_sign, Landmark, and obstacle, as well as corresponding relations. These
classes and relations facilitate reinforcement of the mental configuration among
neighboring element
s with respect to the blind pedestrian. Hence the use of a deictic
reference system centered on the person.


The level of local reinforcement connects the various components from the level of
imminent reinforcement, and will therefore result in a better co
nfiguration of the
mental representation. This level includes the classes Street_section,
Street_segment, Pedestrian_Path_Section, and Landmarks. The communication of
information at this level uses an intrinsic reference system. Blind pedestrians can use
t
his level of detail if they consider that their knowledge of the environment is
sufficient. Otherwise, they may access more detail by choosing the level of imminent
reinforcement.


The level of enlarged reinforcement is the most abstract level; it contains
the
classes District, Limit, and City. This level allows the assessment of a more global
perception of the elements of the urban space; therefore, the communication of
information will be according to an extrinsic reference system.


4.

TOWARDS THE INTEGRA
TION OF THE PROPOSED SEMANTIC
CONCEPTUAL MODEL WITH ISO 19133:2005 STANDARD


Standards are key components of any Spatial Data Infrastructure. They help to
define, to describe and to manage geographic information and geospatial services.
The adoption of sta
ndards in geographic information increases the understanding
and usage of geographic information, and facilitates access, integration and sharing
of geographic information (Albrecht, 1999). The International Organization for
Standardization ‘ISO’ is a worl
dwide federation of national standards bodies. The
main objective of this federation is to prepare international standards in several fields.


The technical committee ISO/TC 211 is in charge of preparing international
standards related to geographic inform
ation. This committee works on developing the
methods, rules, and services needed to acquire, process, manage, analyze and
access geospatial data (Brodeur & al., 2000). The ISO/TC 211 technical committee
has defined a special standard ISO 19133:2005 for lo
cation based services called
‘Geographic Information

Location
-
based services

Tracking and navigation’. This
international standard describes data types and associated operations for
implementation of tracking and navigation services (ISO 19133:2005). T
his standard
uses the Unified Modeling Language UML to represent different types, classes and
relations among them.


The following conceptual model summarizes the most important classes for
tracking and navigation services, as well as the relations among t
hese classes. Note
that the most important classes are NT_Network, NT_Link, NT_Turn and
NT_Junction.








Figure
10
: Simplified schema of main classes for tracking and navigation services according
to the ISO 19133:2005 standard






























In this model, the class NT_Network results from the aggregation of three classes:
NT_Link, NT_Turn and NT_Junction. The class NT_Link is a basic oriented curve
element of a network (i.e. an instance of NT_Network); each link
must have two
instances of NT_Turn; one as a start turn and another as an end turn. The class
NT_Turn represents the mechanism relating one instance of NT_Link to another;
each instance of NT_Turn corresponds to a specific node of the network and will be
c
onsidered as an entering or an exiting link. The NT_Junction is an aggregation of
the class NT_Turn instances, meaning that an instance of NT_Junction corresponds
to a number of instances of NT_Turn that occur in the same location. It holds thus all
the po
ssibilities to turn at this junction of the network.


The ISO 19133:2005 standard has been designed to be adopted by a broad
community of location
-
based services. However, our analysis indicates that the
specifications required to meet the needs of the bli
nd pedestrian are not included in
this standard. In attempting to create a unified and generic standard, the committee
has dropped key features necessary to support users with visual deficits.



To overcome this weakness, it necessary to extend this ISO st
andard. In this way,
we may take advantage of its strengths but extend its capabilities in ways that are
more supportive of the needs of the blind.


To achieve this goal, our semantic conceptual model can be integrated with the
platform ISO 19133:2005. In
fact, this semantic model can be considered as an
upper layer, which will be related to the ISO 19133:2005. For instance, the class
ISO 19133

Interesection_signs to the class NT_Advisory, and the class Crosswalk to the class
NT_Link, the class Landmark can be linke
d with the class NT_Waypoint. In the case
of the link between Landmark and NT_Waypoint, it is necessary to extend the
definition of NT_Waypoint given by the ISO 19133:2005 in order to include spatial
objects that are useful for the navigational process and
which are not necessary
accessible by the pedestrian. The effort to achieve this integration as still ongoing,
and we it will be subject of further publication.


4.

DISCUSSION AND CONCLUSION


Semantic information about the surrounding space has a fundamen
tal role in
various tasks of navigation and wayfinding, especially as carried out by blind
pedestrians. This paper proposes a cognitively informed methodology to design a
conceptual semantic model to assist visually impaired pedestrians in their daily
navi
gational activities. We mention the importance of adopting a hierarchical
structure of semantic spatial data according to cognitive principles of navigation.
Consequently, this semantic model has to have a hierarchical structure in order to
help users to b
etter configure their mental representation of urban areas.


The elaboration of this semantic conceptual model takes into account the main
three components of an SDI: People, Data and Standard. For the People
Component, the proposed model aims to meet the
specific needs of visually impaired
pedestrians. These needs are generally omitted in SDI dedicated to general public.
The importance of Data Component is enhanced in our semantic conceptual model
through the use of classes that allows communicating useful
information to help the
navigation of the blind, and the adoption of a hierarchical structure with three levels
of detail. In addition, we proposed to enrich the Standard ISO 19133:2005 (which
corresponds to the third component of an SDI) by introducing t
he semantic model as
an upper layer which will be integrated with this standard.



The proposed model can be seen as a promising solution that allow the definition
of appropriate Spatial Data Infrastructure, which is necessary to develop navigational
assistive technologies for people with different types of disabilities. In fact, the
integration of our semantic model with standards that are widely adopted and used,
such as ISO 19133:2005, allows filtering the relevant spatial data for assistance and
en
riching the semantics of spatial information that are provided to blind pedestrians
during their daily activities of navigation and wayfinding. The process of the
integration between the proposed semantic model and the standard ISO 19133:2005
is still ongo
ing. This integration will consist of the definition of relations between
similar concepts extracted from the two models.


Note that it is essentially to validate the consistency and effectiveness of this
semantic model. We need to evaluate the relevancy o
f classes at each level of detail
offered to the blind pedestrian with regard to navigation performance. Hence, it is
necessary to elaborate validation procedures through the definition of
experimentation protocols involving blind pedestrians in a variety
of realistic urban
spaces.


Finally, we may also observe that the process adopted here to support people with
visual impairments may also be adapted to address other individuals with sensory or
motor deficits who have a need for navigational aids that take
into account the
particulars of their impairments. Indeed, this work needs to be generalized to
accommodate broader groups.


In conclusion, we hope that this paper highlights the importance of the semantic
content of the navigational space to provide effe
ctive aids for individuals with
impairments, in this particular case, people with visual deficits. We also expect that
this research will help not only blind people, but also other people with disabilities to
better integrate their activities within societ
y in active and independent manner.


REFERENCES


Albrecht, J. (1999). Geospatial information standards: A comparative study of
approaches in the standardization of geospatial information.
Computers &
Geo
s
ciences,
25:9
-
24.

Bentzen, B.L. Barlow, J.M. and Bon
d T. (2004). Challenges of Unfamiliar Signalized
Intersections for Pedestrians Who Are Blind: Research on Safety.
Transportation Research Record
, 1878: 51
-
57.

Brodeur, J. Bédard, Y. and Proulx, M.
-
J. (2000), Modellilng Geospatial Application
Databases usin
g UML
-
based Repositories Aligned with International
Standards in Geomatics.
Proceedings of 8th ACM sy
m
posium on GIS
, 10
-
11
November 2000, Washington, D.C., USA. 36
-
46.

Edwards, G. and Ligozat, G. (2004). A formal model for structuring local perceptions
of
environmental space.
Cognitive Processing
, 5(1): 3
-
9.

Erik de man, W.H. (2006). Understanding SDI; complexity and institutionalization,
International Journal of Geographical Information Science,
20(3) : 329
-
343.

Gaunet, F. (2006). Verbal guidance rules for
a localized wayfinding aid intended for
blind
-
pedestrians in urban areas.
Universal Access in the Information Society
,
4: 338
-
353.

ISO 19133:2005. Geographic information

Location
-
based services

Tracking and
navigation.
International Standard ISO 19133
:2005
, 150 p.

Jul, S., and Furnas,
G. W. (1997). Navigation in Electronic Worlds.
Special Interest
Group on Computer
-
Human Interaction Bulletin
, 29(4): 44 p.

Kitchin, R.M. Blades, M. Golledge, R.G. 1997. Understanding spatial concepts at the
geographic sca
le without the use of vision
. Progress in human Geography
,

21(2):225
-
242.

Kitchin R.M. and Jacobson R.D. (1997). Techniques to collect and analyse the
cognitive map knowledge of persons with visual impairment or blindness:
Issues of validity.
Journal of Vi
sual Impairment & Blindness
, 91(4): 360
-
376.

Lynch, K. (1960).
The Image of the City
. Cambridge : MIT Press, 194 p.

Mark, D.M. Freska, C. Hirtle, S.C. Lloyd, R and Tversky, B. (1999). Cognitive models
of geographical space,
International Journal of Geograp
hical Information
Science
, 13(8): 747

774.

May, A.J. Ross, T. Bayer, S.T. and Tarkiainen, M.J. (2003). Pedestrian navigation
aids : information requirements and design implications.
Pers Ubiquit
Comput.
, 7: 331
-
338.

Mennis, J.L. (2003). Derivation and impl
ementation of a semantic GIS data model
informed by principles of cognition.
Computers, Environment and Urban
Systems
, 27: 455

479.

Raubal, M. and Winter, S. (2002). Enriching Wayfinding Instructions with Local
Landmarks.
Lecture Notes in Computer Science

Geographic Information
Science
,

2478:243
-
259
.
Sp
ringer Berlin / Heidelberg.


Rajabifard, A. Escobar, and F. Williamson, I.P. (2000). Hierarchical Spatial
Reasoning Applied to Spatial Data Infrastructures,
Australian Cartography
Journal,
29(2) : 41
-
50.

Ra
jabifard, A. and Williamson, I.P. (2001). Spatial Data Infrastructures : Concept,
SDI Hierarchy and Future Direction,
Proceeding of GEOMATICS'80
Conference
, Tehran, Iran, April
-
2 May 2001.

Reginster, I. Edwards, G. 2001. The Concept and Implementation of P
erceptual
Regions as Hierarchical Spatial Units for Evaluating Environmental
Sensitivity. URISA Journal 13(1):5
-
16.

Rüetschi, U. and Timpf, S. (2005). Using Image Schemata to Represent Meaningful
Spatial Configurations.
Lecture Notes in Computer Science
, 3
762: 1047
-
1055. Springer, Heidelberg.

Smith, J. Kealy, A. Mackaness, W. and Williamson, I. (2004). Spatial Data
Infrastructure : Requirements for Mobile Location Based Journey Planning,
Transaction in GIS,
8(1) : 23
-
44.

Timpf, S. (1999). Abstraction, level
s of detail, and hierarchies in map series. Freska
C, Marks DM editors.
COSIT 99; LNCS
, 1661: 125
-
140.

Tversky, B. (2003). Structures of Mental spaces : How people think about space.
Environment and behavior
, 35(1): 66
-
80.

Yaagoubi, R. and Edwards, G. (200
8). Cognitive Design in action: developing
assistive technology for situational awareness for persons who are blind.
Disability and Rehabilitation: Assistive Technology
, 3(5): 241
-
252.

Yaagoubi, R., Edwards, G., and Badard, T. (2009). Standards and Spatial
Data
Infrastructures to help the navigation of blind pedestrian in urban areas. In A.
Krek, M. Rumor, S. Zlatanova, & E. M. Fendel (Ed.),
Urban and Regional
Data Management: UDMS 2009 Annual
, pp. 139
-
150. Leiden: CRC
Press/Balkema. Ljubljana, Slovenia, Ju
ne 24
-
26.


Proceedings of Global Geospatial Conference 2012

Québec City, Canada, 14
-
17 May 2012