A Spatiotemporal Query Processor Based on Simplicial Representation

In spatial databases, efficient filtering of spatial objects using some index is a key issue to improve the retrieval performance R*-tree is a well-known index which can manage points and rectangles. In R*-tree, objects are approximated by their MBRs and input region is also approximated by its MBR. And exact check is postponed to the refinement step. This paper propose a new approach to this filter. Input region as well as each of stored objects is managed by a R*-tree which allows more exact approximation of it and so enables more efficient filtering. This design is used in our spatiotemporal database system Hawks whose data model, called Universe, is based on the simpficial complex theory.