Automated Urban Travel Interpretation: A Bottom-up Approach for Trajectory Segmentation

. 2016 Nov 23;16(11):1962. doi: 10.3390/s16111962

Algorithm 1 Pre-processing of a GPS trajectory in two stages

1:
INPUT 1) $Π_{R} :$ rawTlist(), 2)lowSpeed: LST
2:
OUPUT $Π_{P} [spatialFiltered : sfTlist (); velocityFiltered : vfTlist ()]$
3:
PROCEDURE Spatial_Filter()
4:
$r a w T l i s t . s i z e () = k_{1}$
5:
for i=0 to $k_{1}$ −1 do
6:
if $r a w T l i s t . g e t (i) . g e t A c c u r a c y () > 40$ then
7:
$s f T l i s t . a d d (P_{i})$ # rawTlist.get(i)= $P_{i}$ , where $P_{i}$ is a GPS point in raw trajectory $Π_{R}$
8:
end if
9:
end for
10:
END PROCEDURE
11:
PROCEDURE Velocity_Filter()
12:
$s f T l i s t . s i z e () = k_{2}$
13:
for i=1 to $k_{2}$ −1 do
14:
if $s f T l i s t . g e t (i - 1) . g e t S p e e d () > L S T | | s f T l i s t . g e t (i + 1) . g e t S p e e d () > L S T$ then
15:
vfTlist.add( $P_{i}$ )
16:
else
17:
if $s f T l i s t . g e t (i) . g e t S p e e d () < L S T$ then
18:
if $s f T l i s t . g e t (i - 1) . g e t S p e e d () \leq L S T | | s f T l i s t . g e t (i + 1) . g e t S p e e d () \leq L S T$ then
19:
vfTlist.add( $P_{i}$ )
20:
end if
21:
end if
22:
end if
23:
end for
24:
END PROCEDURE