Algorithm 2: Trajectory Similarity Measurement

Input: Trajectory dataset $T{r}_i$ preprocessed by Algorithm 1, similarity threshold $\epsilon$. Output: Trajectories $T_{s{r}_i}.$

1: for $T_{s_i}$ in ${\left[T_{s_1} T_{s_2} \cdots T_n\right]}^T$

2: calculate the distance $D_{SSP{D}_i}\left(T_{s_i},T_s\right)$ between $T_{s_i}$ and $T_s$

3: end for

4: for $D_{SSP{D}_i}\left(T_{s_i},T_s\right)$ in $\left[D_{SSP{D}_1} D_{SSP{D}_2} \cdots D_{SSP{D}_n}\right]$  5: save the trajectory which $D_{SSP{D}_i}$ < $\epsilon$  6: end for  Return: $\left[T_{s{r}_1} T_{s{r}_2} \cdots T_{s{r}_m}\right]$