Database Reference
In-Depth Information
Keywords:
Mobility, Spatiotemporal Data Management, Querying Mobile Objects,
Movement Modeling,
1. Introduction
The use of sensors that capture user location information over time is
rapidly expanding and the various types of sensors are becoming ubiq-
uitous. Location sensors are integrated into a large number of personal
devices including PDAs, smartphones, and watches [104, 105]. Among
the sensors used to acquire spatiotemporal data, the most popular is the
Global Positioning System (GPS). GPS receivers are commonly embed-
ded into vehicles for trip navigation, sports watches to track and monitor
personal progress in hiking and running, and smart phones to provide
general purpose location aware querying. The increased availability of
positioning data has given rise to location based services (LBS), which
utilizes a user's current position in order to personalize results. For in-
stance, suppose a user searches for coffee shops on her smartphone. The
query results will be filtered using both the relevancy from the search
string as well as her position to return popular coffee shops which are
physically close.
As an example of the availability of location data,
figure 1
shows GPS
data uploaded to [106] by users in a region of Los Angeles, CA along
with a comparison to the underlying road network of the same area.
From the figure, we can see two things: first, there is a large amount of
spatiotemporal data available online. Second, the data is generated over
several modes of transportation. Given how well the GPS trajectories
outline the heavily traveled roads, it is clear that most of the data is
collected while users are in vehicles. However, upon closer inspection,
we can see that some trajectories cannot possibly be associated with
an automobile and therefore must have been collected using some other
mode of transportation (i.e. train, subway, or walking).
The adoption of new LBS has increased nearly as quickly as the tech-
nologies have become available. As of 2011, approximately 75% of smart-
phone users are using their devices for navigational purposes (e.g. driv-
ing directions) and 95% use their smart phones for location based search-
ing [104, 105]. Coupled with the increasing number of smartphones year
after year (approximately a 13% increase from 2010 to 2011 [104]), this
signals a steep upward trend for LBS.
Given the availability of large quantities of spatiotemporal data, it is
possible to ask several interesting questions about object movement. For
example, assume we have a database containing the current positions of a
