Research On Indexing Massive Trajectory Data

The traditional spatial indexing technology is difficult to realize the quick query and efficient calculation of trajectory data Analysis.Meanwhile,the massive trajectory data also poses a challenge to the centralized trajectory index.The distributed indexing technology based on grid algorithm and R-tree is problematic in terms of resource occupation and construction speed.In this thesis,the above problems are researched as below:(1)In this thesis,a trajectory data processing method based on bitmap coding is proposed,and a query algorithm based on bitmap technique is designed.Comparing with the query scheme based on spatiotemporal database,the computational efficiency and the feasibility of extending the trajectory query problem are verified by experiments.(2)Based on the proposed trajectory coding method,a trajectory-based range query method is designed.Based on the algorithm of bitmap technology,an algorithm to solve the problem of trajectory data range query is proposed,which is respectively compared with spatiotemporal database and pure memory R-tree index.The impact of different schema on the performance of query is analyzed.(3)Based on the proposed trajectory coding method,this thesis analyzes the query problem of spatial geometric relation of trajectory.Combining with data filtering and conventional geometric relation calculation,this thesis proposes an algorithm to solve the query of trajectory passing through the specified area.The influence of the proposed algorithm on the query performance is analyzed by experiment.(4)Based on the Spark,a distributed computing framework,this thesis analyzes the architecture of the parallel query scheme of spatiotemporal data,and determines the key steps of the distributed query.This thesis completes the realization of the proposed trajectory coding method in the Spark framework,analyzes the construction cost of the query scheme and the requirements for system resources.Through the experiment,the advantage of the trajectory query scheme for the sampling point query scheme is verified.The results show that the proposed trajectory-oriented query scheme can provide the query efficiency for the trajectory data,and the trajectory-oriented query scheme has higher query performance than the general query scheme in the intensive calculation of the massive trajectory data analysis.