Research On Continuous Nearest Neighbor Queries Of Moving Objects In Spatial Databases

As the growing importance of spatial database, its scope of application has far exceeded the traditional fields of GIS.The nearest neighbor query is one of the most important queries in spatial databases. In the traditional nearest neighbor search method,query point is stationary. As wireless communications and the location technology to track moving objects achieved great development in recent years,queries issued by a dynamic object become a new hotspot.Continuous nearest neighbor query is one of the most important queries in spatial databases,it is used to query the nearest neighbor of each point in a given track, as an extension of the nearest neighbor query, because of wireless communications and the location technology to track moving objects achieved great development in recent years, it is possible to track and record the position of moving objects. Makes the CNN query an important and difficult of study. Effective way of obtaining data is always accomplished through the use of the index in spatial databases. The basic idea of spatial index, actually all of the basic idea of spatial query processes the use of approximation. This method can make index structure to manage the object in accordance with one or more space codes, these space codes are more simple than the objects themselves geometric objects. A basic example is MBR(the smallest rectangle with axes parallel which surrounds objects).This subject analyzes some kinds of current continuous nearest neighbor query.It is the most effective that the query algorithm based on R-tree of Y.Tao etc. Therefore, the method of this algorithm is described in detail. Previous research results is for the CNN of two-dimensional space, but in real life, the world around people is in the form of three-dimensional space, to compensate for this deficiency. The continuous nearest neighbor query of three dimensional space is proposed to meet the higher demand for the query. Some concept is put forward,such as perpendicular bisector of plane,break point and adjacency ball. The CNN query algorithm of three dimensional space is implemented based on R tree. For the mass of spatial data, use prune rule and spatial index structure to reduce the number of end points of access,greatly improves the query efficiency.This thesis gave the related pseudo code of the core algorithm,analyzed the execution procedure combining a specific example,at last the performance of this algorithm was analyzed through simulation experiment.