(The) $B^{st}$-tree index and a dynamic histogram method for spatiotemporal query processing

Abstract

Rapid advances in wireless communication and positioning technologies enable new application areas such as the location-based services. Representative location-based applications in various areas include mobile e-commerce, traffic control and military C4I systems and eventually lead the pervasive computing environment. These services can be achieved by the spatiotemporal database technology that concerns about the time-varying geometric information. In mobile environments, the positions of moving objects are represented as coordinates at a certain time or trajectories. Hence those two classes of location information are needed to be treated appropriately by the requirements of applications. However, managing exact geometric location information is difficult due to the continual changes of moving objects with traditional databases. Indexing is one of the major research issues in the spatiotemporal databases and traditional indexes focus on optimization of node accesses during index construction that do not refer on-line continuous updates. For efficient handling of frequent updates as well as the query performance, we investigate previous indexing approaches and evaluate their performances firstly. As a result, we propose the $B^{st}$-tree that is based on quad partitioning index structure and space filling curve for point moving objects and use constraint database approach to improve the efficiency of future spatiotemporal query performance. Our method provides an indexing framework for future location queries based on space partitioning and the dual transformation. And the multi-version structure increase its size as time progresses. Hence the performance enhancement is achieved by our cost-based time-window management algorithm that determines the appropriate index reorganization probabilistically induced by query cost functions. This approach can be applied for the trajectory indexing specifically in a next generation mobile communication systems. ...