Research Of Spatio-temporal Urban Vehicle Trajectory Data Indexing Design

With the development of wireless sensor and communication technology,trajectory data gradually becomes a hot research field.Trajectory data having strong characteristics of time and space,carrying the behavior characteristics of pedestrian or public transport,is of great significance for data mining.The characteristics of the large data itself need huge amounts of data mining resources,so fast track data storage and efficient retrieval has become an important factor restricting the development of trajectory data.Indexing technology is the key technology of database storage.Regardless of the application of traditional relational database nor NoSQL database,indexing technology is closely related to data distribution and data query in databases,and has profound impact on the performance of the storage and query.Therefore proper data index structure adapting to the trajectory characteristics is in great demand of improving the convenience of trajectory data storage and query.The paper creates index from the perspective of space and time.The first index put particular emphasis on spatial query,in consideration of roads.It is composed of two layers:the upper layer adopts regional blocks,the bottem layer adopts quartree,with time stored in its leaves.The second index put particular emphasis on temporal query.It stores multiple time quantums with interval tree based on red-black tree.Inside every time quantum,there is a corresponding trajectory,with one R-tree.Experimental results show that the first index in this paper performs better in spatio-temporal POI query that emphasizing on space,while the second index performs better in spatio-temporal trajectory query that emphasizing on time,compared with the composite index of spatio-temporal index in MongoDB.In the future,more factors will be taken into consideration despite road to create spatio-temporal index,and it will cost less to update spatio-temporal index.