A Framework for Bus Trajectory Extraction and Missing Data Recovery for Data Sampled from the Internet

. 2017 Feb 10;17(2):342. doi: 10.3390/s17020342

Algorithm 3 Trajectory fragment cleaning algorithm.

Input:

G_{c}^{(j, m, l)}

C_{M}

u_{\min}

;

Output:

{\hat{G}}_{c}^{(j, m, l)}

{\hat{U}}^{c}

{\hat{R}}_{c}^{(j, m, l)}

;

Process:

(a1) set a temporary noisy set ${\hat{R}}_{c} = ⌀$ ;
(a2) get $U^{c}$ of $G_{c}^{(j, m, l)}$ with $C_{M}$ ;
(a3) get ${\bar{N}}^{c}$ from $U^{c}$ with $u_{\min}$ ;
if ${\bar{N}}^{c} = 0$ then
go to (a5);
else
go to (a4);
end if
(a4) find the index k in ${\bar{N}}^{c}$ with the maximum anomalous indexing;
if k is not unique then
choose $k^{*}$ be the one in k with the minimum row summation;
else
set $k^{*} = k$ ;
end if
update ${\hat{R}}_{c}$ by adding the $k^{*}$ -th data in $G_{c}^{(j, m, l)}$ into ${\hat{R}}_{c}$ ;
delete the $k^{*}$ -th data in $G_{c}^{(j, m, l)}$ , delete the $k^{*}$ -th row and $k^{*}$ -th column in $U^{c}$ ;
go to (a3);
(a5) output the cleaned trajectories:
if $| G_{c}^{(j, m, l)} | > 1$ then
set ${\hat{G}}_{c}^{(j, m, l)} = G_{c}^{(j, m, l)}$ , and ${\hat{U}}^{c} = U^{c}$ ;
else
set ${\hat{G}}_{c}^{(j, m, l)} = ⌀$ , and ${\hat{U}}^{c} = ⌀$ ;
end if
set ${\hat{R}}_{c}^{(j, m, l)} = {\hat{R}}_{c}$ .