Semantic Location Based Services for Smart Spaces

1
Semantic Location Based Services for Smart Spaces

• Kostas Kolomvatsos, Vassilis Papataxiarhis,
• Vassileios Tsetsos
• Pervasive Computing Research Group
• Communication Networks Laboratory
• Department of Informatics and Telecommunications
• University of Athens Greece
• MTSR 07 _at_ Corfu, Greece

2
Outline

• Introduction
• Spatial Ontology
• GIS metadata and ontology population
• Hybrid Navigation Algorithm
• Conclusions

3
Location Based Services

• LBS The core of smart environments
• Are the most popular context-aware services
• Navigation service One of the biggest challenges
  due to its complexity
• Current Objective
• Universal and optimized access to LBSs
• Advanced user experience
• Current Limitation
• Existing representation techniques lead to
  incapability of smart management and
  exploitation of spatial data

4
Motivation

• Navigation
• Typical representation formalism spatial graphs
• Problems
• Multi-criteria search NP-hard
• Redesign needed to extend an existing algorithm
  with more criteria
• Conclusion Traditional algorithmic approaches
  seem to fail
• Proposed solution Semantic enrichment of LBS

5
Spatial Model

• Indoor Navigation Ontology
• It describes the basic elements of indoor
  environment and the basic relationships between
  them
• It facilitates path searching

http//p-comp.di.uoa.gr/projects/ontonav/INOdoc/in
dex.html
6
GIS Annotation

• Layered architecture
• Each layer corresponds to a basic concept (or set
  of concepts) in INO
• Lower layer Building blueprints
• Second layer Corridors (lines)
• Points are defined upon corridors

• Basic point metadata
• x,y coordinates
• Floor
• Id
• Label, etc.

7
Ontology Population

• Spatial Database creation
• Transform GIS Layers to Spatial Database Tables
• Automatic instantiation of INO through GIS
  metadata

Instances Creation Algorithm
Spatial Database
Ontology Instances
8
Instances Creation Algorithm

• Based on GIS data
• The algorithm involves the following steps for
  all floors in the building
• Find which points belong to each corridor
• Find the ends of each corridor
• Find the neighbors of each point
• Create the instances in INO classes indicated by
  the GIS layers and the information extracted in
  the previous steps

9
Navigation Algorithm

• Hybrid rule-based algorithm. Takes into account
• Route complexity
• Euclidean route distance
• User profile (capabilities and preferences)
• Steps
• Create user compatible building graph based on
  user profile and application of access rules
• E.g. WheelChaired_User(?x) Stairway(?y)
  isObstacleFor(?y,?x)
• Find the k-simplest paths
• Assign the total cost of each path as a function
  of bonuses and penalties of the total path
  distance, preferences and perceptual rules

10
System Functionality (I)
Building blueprints (GIS)
User-compatible graph
Building graph
Data Migration
Spatial DB
INO instances
User-compatible INO instances
Indoor Navigation Ontology (INO)
User profile (capabilities)
11
System Functionality (II)
COM
Complexity Path Computation
SEM
Semantic Path Selection
User-compatible graph
User profile
PRS
Path Presentation
COM
SEM
PRS
K-simplest paths
Best Traversable Path
User
User location and destination
Perceptual Rules
12
Navigation Example
A H 4 possible paths
1) ACFGH shortest path
2) ABDIH simplest path
3) ABDEFGH, node E stairs
4) ACFEDIH , node E stairs
Selected Path ABDIH
A little longer than ACFGH, but much easier to
describe !
13
Implementation Details

• ESRI ArcGIS software
• PostGIS spatial DB
• Protégé Ontology Editor
• Knowledge Representation Languages
• Ontology models in OWL-DL
• SWRL rules
• Bossam for OWL and SWRL reasoning
• Mascopt Library for graph creation and path search

14
Semantics is not everything

• Example Orientation Issues
• Two extra properties storing the real GIS
  coordinates of each door and not only its
  projection to the corridor
• Compute the angle between the line vertical to
  users direction and the line specified by the
  users position to the door.
• If angle ? gt 90, the door is on the left side
• Else, the door is on the right
• Similar process for the turns.

15
Contributions and Open Issues

• Main Contributions
• Semantic representation of GIS metadata with the
  aid of a spatial ontology
• A rule-based hybrid combination of k-simplest
  paths search algorithm with Euclidean distance
  and other application parameters like user
  profile and abilities/preferences.
• Support for flexible navigation schemes
• Content-based navigation
• Presentation-based navigation
• Open Issues
• Immature reasoning engines in terms of
  performance and interoperability with rule
  engines
• Development/Improvement of tools for spatial
  ontology population

16
Thank you!

• QUESTIONS?

http//p-comp.di.uoa.gr