Research On K-nearest Neighbor And Group K-nearest Neighbor Query For Moving Objects In Obstructed Spaces

The spatial database,which was researched in the 1970 s,has the features of a huge amount of data,high accessibility and complicated spatial data model.It is used to store,manage and retrieve spatial data.And spatial data query and measurement is a major research direction of spatial database.Although there have been many literatures on spatial data query and measurement has done extensive and detailed research,but it still need to constantly improve the existing query methods to meet the real-life people's query needs.Nearest neighbor query is one of the most common query methods in query types and processing methods of spatial database.With the rapid growth of intelligent mobile devices(such as PAD and smart phones),and the rapid growth of wireless communication and positioning technologies(such as GPS),more and more users issue a query request when they are moving,mobile object dynamic query also developed rapidly.Based on this,this dissertation analyses the two existing nearest neighbor query types-continuous nearest neighbor query,group nearest neighbor query and nearest neighbor query under obstacle,and proposed a new query mode: COPk NN(continuous obstructed possible k-nearest neighbor)query and its extension of Pk OGNN(probabilistic k obstructed group nearest neighbor query)query.Where the COPk NN query refers to the k-nearest neighbor query in the presence of an obstacle,and Pk OGNN refers to the group k-nearest neighbor query in the presence of an obstacle.Because of the inherent uncertainty of the moving objects itself,it can't use a certain point to represent the moving objects,but rather an indefinite region.In addition,in the obstacle space,the distance between two query points can't be simply expressed by the Euclidean distance,which needs to be represented by theobstacle-distance.For the two types of queries,this paper firstly gives the corresponding formal definition and reveals its uniqueness.Secondly,it gives the method to determine the uncertain region model of the moving objects,and it is convenient to determine the approximate position of the moving objects.Thirdly,it gives the corresponding calculation method of the obstacle distance.In the experimental part,the performance of the two algorithms is evaluated by comparing the experimental methods respectively.The performance of the two algorithms is improved effectively.