Abstract :
As GPS datasets are becoming more available with spatial and temporal information, the needs of high quality, predictive information about spatio-temporal are turning out to be a feasible goal. However, many GPS devices are prone to failure, which results in missing GPS data. Moreover, the missing data may affect the accuracy of the result. To make the matter complicate, the missing data is always not missing at random, and is characterized by long time during which no data is observed. In this paper, there exists a missing trajectory in a truck´s traces which lasts a few of seconds. In order to make the trajectory complete, an STZ algorithm is proposed to solve this problem for the first time. The STZ algorithm is aimed to compute the area between any two trajectories. In this paper, we have known many possible routes which are from the original point to the destination. Next, this approach calculates the area between any two trajectories which represents the similarity of the two routes. Last, the K-modes algorithm is adopted to group the similarity of the routes and complete the missing trajectories.
