Trajectory Collection and Reconstruction - Mobility Data - page 41

illustrated in Figure 2.2 . Quddus et al. ( 2007 ) proposed a technique for replacing
each position of the original trajectory by the point on the network that is the
most likely position of the moving object. Greenfeld ( 2002 ) proposed a method
based on topological analysis using the observed position of the individual with-
out assuming any knowledge of the expected traveling route and the speed or
heading information supplied by the GPS. Furthermore, Newson and Krumm
( 2009 ) used hidden Markov model approaches to find the most likely road route
corresponding to a sequence of positions.
Meratnia and de By ( 2004 ) proposed the Top-Down Time Ratio (TD-TR)
and Open Window Time Ratio (OPW-TR) algorithms for the compression of
spatio-temporal data. Potamias et al. ( 2006 ) proposed the two algorithms, called
Thresholds and STTrace, respectively, for online trajectory data compression.
Kellaris et al. ( 2009 ) present a different approach by replacing certain episodes
of a trajectory by selected shortest paths between the beginning and ending
position of these episodes. As for the trajectory reconstruction topic, Marketos
et al. ( 2008 ) presented a method for determining different trajectories as part
of a trajectory reconstruction manager. On the other hand, Ya n e t a l . ( 2011 )
presented a technique for reconstructing semantic trajectories from the raw GPS
mobility records.
With regard to privacy issues, Gruteser and Grunwald ( 2003 ) introduced the
concept of location k -anonymity in the context of LBS; Jensen et al. ( 2009 )intro-
duced the dichotomy of identity privacy versus location privacy; Casper ( Chow
et al. , 2009 ) is a major privacy preserving framework supporting location k -
anonymity; the velocity-based attack is described in more detail in Ghinita et al.
( 2009 ); Damiani et al. ( 2010 , 2011 ) introduce the semantic location cloaking
paradigm.