Geohash-Trees: An Adaptive Index Which can Organize Large-Scale Trajectories

Trajectory data which contains mining valve is widely distributed and large-scale. How to organize trajectory data and retrieve trajectory data efficiently becomes very difficult to solve. We present a framework of adaptive index based on Geohash to organize the worldwide and large-scale trajectory data set. Different trajectory data sets will be covered by Geohash grid which is deepest, and then we take the grid as root node to generate adaptive Geohash-Trees. In order to quickly locate the corresponding index, we design trie on the basis of the feature of Geohash. Adaptive Geohash-Trees is a spatial index based on grid. It can divide the space according to the track density by adopting a variety of strategies which improves the efficiency of range query. Meanwhile, we design the algorithm of incremental insertion and update for the supporting of real-time update of trajectory data. Furthermore, this framwork has been migrated in Oracle. The experiment results verify that our approach in several aspects such as range query and occupied disk size performs much better than R-Trees.