Transcript An ontology-based approach for the semantic modelling and

An ontology-based approach for
the semantic modeling and
reasoning on trajectories
Miriam Baglioni
Dept. Computer Science,
Pisa, Italy
Jose de Macedo
EPFL, Lausanne, Switzerland
Chiara Renso
KDDLab, ISTI, CNR, Italy
Monica Wachowicz
Technical University of Madrid, Spain
SeCoGIS - Bercellona, October 20, 2008
Introduction

Mobile communications and ubiquitous computing
pervade our society

Wireless networks sense the movement of people and
vehicles, generating large volumes of mobility data.

A scenario of great opportunities and risks:


analyzing mobility data can produce useful knowledge;

individual privacy is at risk, as the mobility data contain
sensitive personal information.
A new multidisciplinary research area is emerging at this
crossroads, with potential for broad social and economic
impact.
SeCoGIS - Bercellona, October 20, 2008
2
The scenario
3
Trajectories
Managing location information gives the possibility to
access space-time trajectories of personal
devices.

Trajectories are the traces left behind by
moving objects and individuals

Trajectories offer opportunity to extract behavioral
patterns
SeCoGIS - Bercellona, October 20, 2008
4
Raw Trajectories
A trajectory is the user defined record of the evolution of the
position of an object that is moving in space during a given
time interval in order to achieve a given goal.
trajectory: [time interval] → space.
Mobile devices produce raw location data (obj-id, x, y, t)
SeCoGIS - Bercellona, October 20, 2008
5
The scenario: the need for semantics
Which are the
malicious/tourist/
… activities?
6
From Raw Trajectories to Semantic trajectories
Semantic Trajectory defined as sequence of Stops and Moves.
Stop is where the position of the object stays fixed
Move is the part of the trajectory where the position changes.
Monument [10:30 – 12:30]
Hotel [09:00 – 10:00]
Restaurant [13:00 – 15:00]
7
The need for semantics and reasoning

Movement data is complex

Movement happens always in a geographical space, that
affects and gives meaning to the movement itself.

Adding semantics may support the user in the
interpretation and understanding of trajectories as people
behavior or activity: suspicious behavior or tourist activity
can be inferred from a motion pattern.
8
The trajectory enrichment process
9
The role of the ontology

Ontology (Description Logics) as a knowledge
representation and reasoning formalisms : Trajectories +
Geography + Domain expert knowledge.

Geographic and domain knowledge may fit well a concept
hierarchy and this give an additional power to the system
since it can encode and infer new knowledge.

Reasoning tasks:


Consistency of ontology:
Classification by instance checking of individuals (instances)
into concept categories.
10
Semantic representation of movement data
Semantic
Trajectory
CityPlace
stop_is_at
trajCompOfStop
Monument
Church
…
Bridge
Stop
Museum
is_a relationship
object property
Malicious
Activity
Time
Move
stophasDuration
11
Motivating Example
General lack of knowledge of recreation planners
about the actual movement of players within a
recreation site.
The assumption is that game activities are not always
performed in planned recreational zones.
Thus an analysis of movements would allow a better
understanding of how people behave during the
game.
12
Motivating Example
Paper chase is an old children’s game in which an area is explored by
means of a set of questions and hints on a sheet of paper. Each team tries
to find the locations that are indicated on the questionnaire.
Once a place is found, they try to best and quickly answer the questions on
the sheet, note down the answer, and proceed to the next point of interest
(POI).
Game Frequency 1550 by Waag
Society. Students are transported
to the medieval Amsterdam of
1550 with the mobile phone.
13
Understanding movements
A malicious activity of the tracked kid is represented by a trajectory that has some
short stops at a bridge (kids launching stones in the water).
The reasoning engine checks if an individual (a given trajectory) is an instance of a
concept (malicious activity).
The class Malicious Activity is an implicit class defined in the ontology by an axiom: a
combination of logical operators
NOT_GAME_STOP  (StopHasTime some (hasDuration some LONG)) or
(StopHasTime some (hasDuration some SHORT))
GAME_STOP  StopHasTime some (hasDuration some MEDIUM)
MaliciousActivity  trajIsCompOfStop some (NOT_GAME_STOP and
(StopHasTime some (hasDuration some SHORT)) and (is_at some
BridgePlace))
14
Ongoing work
We are currently extending this work in two main directions:

Integrating data mining in the scenario: understand people
behaviour analyzing large amount of data. From millions of
trajectories, data mining extracts movement patterns.

Developing a prototype based on Oracle11g Semantic
Technologies, OWLPRIME language: ATHENA. We are
currently experimenting the system in a larger dataset
coming from GPS on cars and moving in Milan.
15
Athena: Trajectories and city places
16
Athena: Tourist trajectories
17
Athena:
The Ontology
18
Conclusions
Movement data is complex and strictly related with the
geographical space where people move.
Raw trajectory data is not enough to understand people
behaviour: the need for semantics and reasoning
An approach to semantic enrichment process, based on
ontology formalism, to infer behavior classes.
www.geopkdd.eu
19