Q+Rtree: Efficient Indexing for Moving Object Databases

Moving object environments contain large numbers ofqueries and continuously moving objects. Traditional spatial index structures do not work well in this environmentbecause of the need to frequently update the index which results in very poor performance. In this paper, we present anovel indexing structure, namely the Q+Rtree, based on theobservation that i) most moving objects are in quasi-staticstate most of time, and ii) the moving patterns of objectsare strongly related to the topography of the space. TheQ+Rtree is a hybrid tree structure which consists of bothan R*tree and a QuadTree. The R*tree component indexesquasi-static objects-those that are currently moving slowlyand are often crowded together in buildings or houses. TheQuadtree component indexes fast moving objects which aredispersed over wider regions. We also present the experiental evaluation of our approach.