Research And Application Of Trajectory Data Spatio-Temporal Index In NoSQL Database

Nowadays,with the rapid development of wireless sensor technology,GPS technology,Internet technology and cloud computing technology.Acquisition of location data is becoming more and more easy.Location-based service(LBS)is booming.The moving track data are acquired and accumulated day by day.Trajectory data carry a large number of hidden information,it have characteristics of massive and high dynamic.The efficient management and indexing of trajectory data is of great significance to mining the potential value of massive trajectory and serving traffic and travel.In this paper,the spatio-temporal index of trajectory data is explored from database selection,index strategy selection,storage structure and index algorithm design three aspects,based on the practical application scenario.In view of the characteristics and applicability of the existing mainstream NoSQL databases,this paper chooses the Cassandra database to store trajectory data which with good lateral expansibility and fastest write performance.The spatio-temporal index is designed for the bottleneck of reading performance of Cassandra database.We construct composite spatio-temporal index based on Cassandra use time division strategy.The spatial index is constructed by quadtree and Hilbert curve,and the time index is constructed by equal interval partition.The index update cost of key-value database is high,this paper choose the fixed level S2 index to construct spatial index for real-time trajectory data.For the fixed level of space partition,the response efficiency of different precision query scene is different.In order to realize the rapid response of common application scenarios,this paper proposes a partition strategy of spatial level and time interval,which realizes the rapid response of data storage and query operations on the premise of maximizing the use of server and database storage resources.A new algorithm for constructing query region is proposed to solve the problem of query missing.In this paper,we use 10000 taxis in Beijing about 15 million trajectory data to carry out the experiment.The response efficiency of spatio-temporal range query,adjacent query and concurrent access was tested.Experiments show that the proposed method can efficiently store and query massive trajectory data.On the basis of the above research,we designed a trajectory data management function of big data management platform and proposed an algorithm for estimating the passage time of complex routes based on spatio-temporal index.This realized the practical engineering application of big data spatio-temporal index strategy.