Wright State University Wright State University
CORE Scholar CORE Scholar
Computer Science and Engineering Faculty
Publications
Computer Science & Engineering
9-2013
A Geo-Ontology Design Pattern for Semantic Trajectories A Geo-Ontology Design Pattern for Semantic Trajectories
Yingjie Hu
Krzysztof Janowicz
David Carral
Simon Scheider
Werner Kuhn
See next page for additional authors
Follow this and additional works at: https://corescholar.libraries.wright.edu/cse
Part of the Computer Sciences Commons, and the Engineering Commons
Repository Citation Repository Citation
Hu, Y., Janowicz, K., Carral, D., Scheider, S., Kuhn, W., Berg-Cross, G., Hitzler, P., Dean, M., & Kolas, D. (2013).
A Geo-Ontology Design Pattern for Semantic Trajectories.
Lecture Notes in Computer Science, 8116
,
438-456.
https://corescholar.libraries.wright.edu/cse/176
This Conference Proceeding is brought to you for free and open access by Wright State University’s CORE Scholar.
It has been accepted for inclusion in Computer Science and Engineering Faculty Publications by an authorized
administrator of CORE Scholar. For more information, please contact library-corescholar@wright.edu.
A Geo-Ontology Design Pattern for
Semantic Trajectories
Yingjie Hu
1
, Krzysztof Janowicz
1
, David Carral
2
, Simon Scheider
3
, Werner
Kuhn
3
, Gary Berg-Cross
4
, Pascal Hitzler
2
, Mike Dean
5
, and Dave Kolas
5
1
Department of Geography, University of California Santa Barbara, USA
yingjiehu@geog.ucsb.edu and jano@geog.ucsb.edu
2
Kno.e.sis Center, Wright State University, USA
carral.2@wright.edu and pascal.hitzler@wright.edu
3
Institute for Geoinformatics University of M¨unster, Germany
simon.scheider@uni-muenster.de and kuhn@uni-muenster.de
4
Spatial Ontology Community of Practice (SOCOP), USA
gbergcross@gmail.com
5
Raytheon BBN Technologies, USA
mdean@bbn.com and dkolas@bbn.com
Abstract. Trajectory data have been used in a variety of studies, includ-
ing human behavior analysis, transportation management, and wildlife
tracking. While each study area introduces a different perspective, they
share the need to integrate positioning data with domain-specific infor-
mation. Semantic annotations are necessary to improve discovery, reuse,
and integration of trajectory data from different sources. Consequently,
it would be beneficial if the common structure encountered in trajec-
tory data could be annotated based on a shared vocabulary, abstracting
from domain-specific aspects. Ontology design patterns are an increas-
ingly popular approach to define such flexible and self-contained building
blocks of annotations. They appear more suitable for the annotation of
interdisciplinary, multi-thematic, and multi-perspective data than the
use of foundational and domain ontologies alone. In this paper, we in-
troduce such an ontology design pattern for semantic trajectories. It
was developed as a community effort across multiple disciplines and in
a data-driven fashion. We discuss the formalization of the pattern us-
ing the Web Ontology Language (OWL) and apply the pattern to two
different scenarios, personal travel and wildlife monitoring.
1 Introduction
The term trajectory is used in many different contexts. It can be defined as a
path through space on which a moving object travels over time. For example, the
path of a projectile can be described by a mathematical model which returns
the idealized position of the projectile at each point in time. In other cases,
such as studying animal movement, trajectories are defined by a sparse set of
temporally-indexed positions or ”fixes”, while the exact path between these fixes
is unknown and has to be estimated, e.g., by using Brownian Bridges [23]. In
some of these cases, the fixes have no specific meaning and are purely an artifact
of the used positioning technology, restrictions imposed by energy requirements,
area coverage, and so forth. In other cases, the fixes denote important activities
and decision points, and researchers may be interested in labeling and classifying
them. We will refer to the latter cases as semantic trajectories [1]. An example
of such semantic trajectories occurs in location-based social networks (LBSN),
where the fixes are user check-ins to places and the labels are the names and
types of these places [39,28]. The user’s location between check-ins is unknown.
The distinction between semantic trajectories and other fixes is not always crisp.
For instance, the OCEARCH’s Global Shark Tracker
1
can only record pings of
tagged sharks if they surface for a certain amount of time. One could argue that
these fixes do not carry any semantics and just reflect technological limitations
of the used positioning technology. However, they reveal some important infor-
mation, namely the event of surfacing and, thus, can be meaningfully labeled.
Summing up, with the fast development of location-enabled mobile devices, it
has become technically and economically feasible to record a large number of
(semantic) trajectories generated by vehicles, animals, humans, and other mov-
ing objects (e.g., from the Internet of Things). While GPS has been widely used
to detect the outdoor locations of moving objects, WiFi[11,10], RFID[31], and
other sensor-tracking techniques have been employed to extend the geo-locating
capability to indoor environments [20,32].
There are multiple ways to publish trajectory data in order to make it ac-
cessible to others. During the last few years, Linked Data [4] has become one
of the methods of choice. It opens up data silos by providing globally unique
identifiers for physical objects and information entities, links between them, and
semantic annotations to foster discovery, retrieval, and integration. The seman-
tic annotations are realized using shared vocabularies. In a highly heterogeneous
and dynamic environment, such as the Web, arriving at commonly agreed and
stable domain ontologies is a difficult task and progress has been slow over the
last years. Foundational ontologies, such as DOLCE [16], have been usefully
applied as a common ground for geo-ontologies [7]. In a Linked Data context,
however, foundational ontologies tend to be too abstract and introduce a hardly
comprehensible set of ontological commitments difficult to handle for laypersons.
Ontology design patterns [14] have emerged as more flexible, reusable, manage-
able, and self-contained building blocks that help to model reoccurring tasks
and provide common ground for more complex ontologies. To reach a higher
degree of formalization and further improve interoperability, these patterns can
be combined and ultimately aligned with foundational ontologies that act as
glue between patterns. An increasing number of geo-ontology design patterns
has been developed as joint community effort by domain experts and ontology
engineers during so-called Geo-Vocabulary Camps (GeoVocamps) [9,8].
In this paper, we propose an ontology design pattern for semantic trajec-
tories and demonstrate its applicability. While trajectory ontologies have been
developed before [37,34], they were confined to specific application areas and
1
http://sharks-ocearch.verite.com/
were not optimized for querying Linked Data, e.g., via the GeoSPARQL query
language [2]. The proposed pattern is developed with two major goals. First, it
should be directly applicable to a variety of trajectory datasets and, thus, reduce
the initial hurdle for scientists to publish Linked Data [3]. Secondly, it should
be easily extensible, e.g., by aligning to or matching with existing trajectory
ontologies, foundational ontologies, or other domain specific vocabularies.
The remainder of this paper is structured as follows. First, we introduce some
background materials and related work supporting the understanding of the pro-
posed ontology design pattern. Section 3 introduces the conceptual foundation
for the pattern. Next, in section 4, we discuss the formalization of the pattern
using the Web Ontology Language (OWL). In section 5, we demonstrate how to
annotate two trajectory datasets using the proposed pattern, in order to evalu-
ate its applicability. We conclude by summarizing our results and pointing out
directions for future work.
2 Background and Related Work
In this section we introduce related research and background materials relevant
for the presented geo-ontology design pattern.
2.1 Semantic Trajectories
A trajectory consists of a series of spatiotemporal points generated by the mov-
ing object. These points are often represented as {x
i
, y
i
, t
i
} (with x
i
, y
i
denot-
ing a position in the 2D geographic plane, and t
i
representing a time point)
or {x
i
, y
i
, z
i
, t
i
} (with z
i
denoting the elevation information) if the trajectory
should be analyzed in a 3D space. While such spatiotemporal points support an
exploration of the mobility pattern of a moving object [13], many applications
require an understanding of additional information to interpret the trajectories.
For example, a traffic analysis based on car trajectories may not be able to de-
rive meaningful results without incorporating information about road networks.
Similarly, studies on bird migration patterns may require an understanding of
the features of the particular bird species (e.g., their body sizes, food sources,
and competitors) as well as information about the weather conditions during
their flight.
Semantic trajectories fill this gap by associating the spatiotemporal points
and segments with geographic and domain knowledge, as well as other related
information [5,1,33]. These semantically enriched trajectories facilitate the dis-
covery of new knowledge, which otherwise may not be easily found. For example,
human trajectories are best understood when the positional fixes can be labeled
with activities performed at these places and the places are associated with
semantic categories such as restaurant or grocery store.
