A Trajectory Collaboration Based Map Matching Approach for Low-Sampling-Rate GPS Trajectories

. 2020 Apr 6;20(7):2057. doi: 10.3390/s20072057

Algorithm 3: Sliding-Window Algorithm for Trajectory Resampling
Input: a set of GPS trajectories $t r j s = {t r j_{1}, t r j_{2}, \dots, t r j_{n}}$ , search radius $r_{s}$ Output: a collaborative GPS trajectory $r e s u l t T r j = {p_{1}, p_{2}, \dots, p_{m}}$
1.	$r e s u l t T r j \leftarrow \emptyset$ ;
2.	$f i r s t P s \leftarrow t r j s . F i r s t P s$ , endPs $\leftarrow t r j s . L a s t$ Ps;
3.	$p S e e d \leftarrow r a n d o m (f i r s t P s)$ ;
4.	while ( $p C_{1} C_{2}$ .contains(endPs))
5.	create $C_{1}, C_{2}$ with radius $r_{s}$ and $2 * r_{s}$ ;
6.	$p C_{1} \leftarrow$ rtree.query( $p S e e d$ , $C_{1}$ );
7.	$p N e w$ $\leftarrow m e a n (p C_{1})$ ;
8.	$p C_{1} C_{2}$ $\leftarrow$ rtree.query( $p S e e d$ , $C_{1}$ , $C_{2}$ );
9.	$p S e e d$ $\leftarrow$ $r a n d o m (p C_{1} C_{2})$ ;
10.	$r e s u l t T r j \leftarrow p N e w$ ;
11.	end while