Research On Data Model And Querying Techniques For Moving Object Database

With the development of positioning technology and wireless communica-tion techniques, tracking the position of moving objects becomes increasingly feasible and necessary. The management of moving objects has application background in many fields including traffic control, boats navigating, dynamic computing, weather forecast, digital battlefield, etc. The position of moving objects changes continuously over time. Such characteristics make it inapplicable to support efficient management of moving object information with traditional database techniques, so Moving Objects Database techniques are introduced. The Moving Objects Database(MOD)which manages the moving objects information becomes a new challenge to database.At present, MOST data model is the most popular location model of moving objects. After deep study and farther abstract of MOST model, the MOST* data model is proposed in this paper. In the MOST* model, the speed factor which is the most frequently updated in the function is apart from others, so the frequent update of function is reduced and more objects can share the same function .The update performance of MOST* model is analyzed both on theory and experiment.In Moving Objects Databases, nearest neighbor (NN) query is used to find out one or more queried objects that are nearest to query object. Firstly, this paper propose DF algorithm and BF algorithm based on TPR tree, which could query the future nearest neighbor continuously in dynamic condition. From the experiment we can see that BF algorithm is superior to DF algorithm. Furthermore, we improve these algorithms. This paper propose an algorithm, which can find several nearest neighbors by traveling TPR tree one time. Such algorithm can support KNN query for moving objects and has a better performance than BF algorithm.In the end, we improve TPR tree. Before the improvement, we create the index periodically; after the improvement, we adjust TPR tree partially when a moving object changes its velocity or direction that influences the velocity of Bounding Rectangle at the immediately higher level. We prove that query time of near neighbors decreases after the improvement of TPR tree.