Research And Implementation Of Query Technology Of Moving Objects In Road Networks

With the rapid development of Chinese economy and the increase of the motor vehicles, traffic congestion has become a social problem that can not be ignored. Urban transportation networks and moving objects’ information have spatio-temporal characters of large scale and complex structure. As a result, it will be an effective way to resolve the traffic congestion that we accurately deal with a large amount of time and space information and provide the query of traffic congestion. This thesis studies the index and query technology of moving objects in road network. The main contributions are described as follows:(1) The problem of urban traffic congestion is analyzed and summarized. According to the characters of large scale and complex structure in road network and moving objects, this thesis describes the research on moving objects index and continuous density query technology and emphasizes their value to alleviate urban traffic conditions.(2) Most of the traditional continuous density query methods regularly perform clustering algorithm based on density calculation, which will cause the query response time delay because of frequent clustering. Moreover, existing road network models only considers the road and the intersection of factors, ignoring the driving direction and width factor of the road. Thus, road network model is redefined and the concept of the road complexity and the road condition are proposed in this thesis. Finally, RNCDQ algorithm is designed. Through two-stage query processing method, RNCDQ can avoid the continuous global query of road network and moving objects, reduce the number of moving objects and the global dense area of computing resources. The experimental results show that the RNCDQ algorithm has better query efficiency and accuracy, compared with TBDQ, SDQ, MODQ and CBDQ algorithms.(3) Existing index structures of moving objects only optimize two layer structure of memory / disk, ignoring the cache sensitivity of index nodes in the memory. Thus, a novel full-temporal index structure named DFTBx-tree is proposed, which is based on the distributed memory database. The size of memory and data transfer chain index nodes is set according to many conditions of the cache line and the size of disk data pages. Moreover, the historical data migration chain is constructed to realize the persistence of moving object’s historical data and support the full-temporal index of the moving objects’ positions. The experimental results show that DFTBx tree has better query and update efficiency, compared with Bx tree, Bdual tree, TPR* tree and STRIPES.(4) Based on Baidu map API interface and the result of research, a prototype system of moving object query is designed and implemented. The system mainly realizes the functions of display of real-time road conditions and vehicle location, vehicle trajectory query, public transportation line query and transfer program query. The system verifies the validity and practicability of the proposed full-temporal index of moving objects and the technology of continuous density query. They can support the query of real-time traffic, vehicle position and trajectory query in prototype system.