Traj2bits

Abstract

With the popularity of mobile devices and the rapid development of position acquisition technology, the amount of trajectory data has soared dramatically. It is time-consuming to manage and mine massive trajectory data, because we need to access different trajectory samples or different parts of a trajectory for multiple times. Therefore, it is necessary to devise an efficient data management technology to fast retrieve the desired trajectories. In general, building indexes is a basic step for solving query problems. However, traditional spatial indexing technologies are mostly designed for moving objects, and thus are unable to achieve fast trajectory data query and efficient computing analysis. In this regard, we propose traj2bits, a bitmap-based trajectory data encoding schema, to convert trajectories into binary strings. Based on traj2bits, we also design a trajectory query method. Experiments on two real datasets have shown that traj2bits improves the spatio-temporal efficiency of trajectory query. Compared with other schemes, traj2bitsuery encoding occupies less than 1/10 of the disk space, its encoding efficiency is at least four times faster and its range query time is reduced by at least 65%.