Traffic Information Estimation Based On Multi-source Data

In recent years, with the accelerating process of urbanization, the urban population increased dramatically, making the traffic travel demand of urban residents increased significantly, traffic motorization process developed rapidly and private car ownership increased sharply, resulting traffic congestion is getting worse.The problems such as economic losses and environmental pollution caused by traffic congestion are increasingly serious plagued the daily lives of urban residents. In this context, Intelligent Transportation System (ITS) came into being, which, as a new means of urban traffic congestion relief, has been widely recognized. Accurate traffic information, as the most basic part of ITS, is the key to ensure the reliable and good operation of ITS.Installation of GPS equipment taxis and urban monitoring systerm are common means used to detect the traffic information. However, due to the taxi driving characteristics and taxis’ differernt coverage on road links, use taxis GPS data can not be very accurately reflect the overall road traffic information. Urban monitoring system for cost reasons cannot cover the entire city road network, even worse, it is often out of order due to system causes and artificial cause, and thus it cannot continue to detect the overall traffic flow.Based on the above issues, this thesis combines taxis GPS data and urban monitoring systerm data and analyzes these data, mainly to do the following research:Firstly, this thesis preprocesses with taxis GPS data and uses coordinate transformation and map matching methods. Use of urban monitoring systerm data to obtain link travel speed of traffic flow, with the analysis of data collection characteristics, the reason of fault data and missing data occurring is clarified. The corresponding processing and restoration methods are developed. Secondly, contrast exact value obtained by urban monitoring systerm data and the estimated value obtained by taxis GPS data, found that under different traffic conditions, the error is quite different. Therefore, this thesis established a model to estimate link travel speed of traffic flow based on traffic status by use of taxis GPS data. By way of example verification, the results show that the accuracy and reliability of the proposed model are better than conventional direct estimation models. Finally, it is the application of proposed model. In the area of urban monitoring systerm cannot be fully covered, using taxis to represent the overall traffic flow, analyzed the link path delay characteristic information of office place, commercial land, cultural and recreational purposes and residential land.