OntologyBased User Modeling for Pedestrian Navigation Systems

1
Ontology-Based User Modeling for Pedestrian
Navigation Systems

• Panayotis Kikiras, Vassileios Tsetsos, and
• Stathes Hadjiefthymiades
• Pervasive Computing Research Group
• Communication Networks Laboratory
• Department of Informatics and Telecommunications
• University of Athens Greece
• UbiqUM 06 _at_ Riva del Garda

2
Outline

• Pedestrian Wayfinding and Navigation The Theory
• Ontology-based User Modeling
• Existing User Models
• User Navigation Ontology
• OntoNav A Semantic Navigation System
• Conclusion Future Research

3
What Is Wayfinding

• A cognitive procedure for orientation and
  navigation of the user in huge and complex
  environments
• Involves four main steps
• Orientation
• Route Selection
• Routing Control
• Recognition of destination

4
What Affects Wayfinding

• Individuals characteristics
• Sex, age, perceptual and mental abilities, motor
  abilities, prior knowledge of the environment,
• Environments characteristics
• Luminosity, signage, structure, obstacles,
• Learning processes
• Learning abilities strategies,
• Not everyone has the same navigational skills ?
• Personalized navigation is necessary

5
User Profile Components
Wheelchair Escort-aided mobility
Orientation disability Mental impairment
Hearing abilities Visual quality
Mental/Cognitive Characteristics
Sensory Abilities
Motor Abilities
User Profile
Navigational Preferences
User Interface Preferences
User Demographics
Only audio info Only visual info
Age Gender
Avoid stairs
6
Ontology-based User Modeling

• Ontology is
• a formal, explicit specification of a shared
  conceptualization (Studer, 1998 - original
  Gruber, 1993)
• a core knowledge representation technique in
  Semantic Web (OWL language)
• User types ? ontology classes
• User characteristics ? class properties
• Such modeling enables Semantic Web reasoning
  (e.g., classification) and inference (i.e.,
  rules)

7
Existing User Models

• GUMO (General User Model Ontology)
• Represents user dimensions (e.g., demographics,
  abilities)
• Developed in OWL
• - Provides just a vocabulary (no axioms,
  restrictions)
• UserML
• XML language
• Provides just a syntax layer for higher level
  semantic models

8
User Navigation Ontology (UNO)

• An OWL ontology that specifies classes and
  properties for the components of a navigation
  user profile
• User classes are formally defined
• YoungWheelchairedUser
• ? hasAbility AutonomousWheelchairedMobility
• ? hasAge LessThan18
• Enables dynamic classification of users through
  Description Logics reasoning
• UNO is aligned with GUMO (where applicable)

9
UNO Elements
Classes
Properties
10
OntoNav

• Personalized indoor navigation, through the
  exploitation of user- and building-related
  semantics

User Profile Creator
UNO
Building graph
User-compatible graph
Path Computation and Ranking
Rules
Building Blueprints
IF UNOWheelchairedUser(u) AND INOStairway(s)
THEN INOisExcludedFor(s,u)
INO
11
User Profile Creation

• Users can
• A) Choose from predefined profiles
• B) Create a custom profile through forms
• Both options demonstrate serious limitations
• A is too coarse-grained
• B is not automated and requires (substantial)
  user effort
• Users may be reluctant to disclose (all of) their
  abilities/disabilities (AB)
• Solution (challenge) User Profile Inference and
  Calibration
• from user movement, history, published personal
  information (e.g., homepage),

12
Conclusion

• A first attempt to define an axiomatized
  ontology for describing pedestrian users based on
  theories and existing work
• Integration of UNO with a navigation system
  (i.e., Location Based Services)
• Future Research Issues
• User profile inference
• Evaluation with real users
• National project on Universal Access to Indoor
  LBS (GSRT MNISIKLIS)

13
Thank You!

• Questions???
• UNO v0.1 available at
• http//p-comp.di.uoa.gr/ont/UNO.owl