Development Of Dynamic Traffic Flow Information Platform Based On Multidimensional Data Sources

With the development of urban digitalization and information technology, the urban road traffic data collection techniques have become increasingly sophisticated, providing a rich foundation of data for the evaluation of urban road traffic operational conditions and the implementation of intelligent managements of road traffic. However, the characteristics of different sources of data collections are different, thus how to process the basic traffic flow data efficiently, integrate advantages of data from different sources, and display traffic information visually are the key issues for achieving urban road traffic managements and information services. Therefore, in light of the Remote Traffic Microwave Sensors (RTMS), loop, and video data from the urban road traffic system in Beijing, this thesis is intended to develop the dynamic traffic flow information platform based on the multidimensional data sources from the perspectives of the basic data processing, data fusion, and data display.First, the state-of-the-art on the urban road dynamic traffic flow data sources, the historical research and development of the dynamic traffic information platform, and the data fusion technique and its application in the transportation fields are reviewed and synthesized.Second, the thesis proposes the functional modules and building process of the dynamic traffic information platform. According to the functional definitions, the dynamic traffic information platform is divided into three components:the basic data module, the multidimensional dynamic traffic flow data fusion module, and the traffic information display module. The thesis analyzes their functions and relations.Then, the thesis establishes the basic data module of the dynamic traffic information platform. Based on an analysis of the characteristics and the range of applications of the three traffic flow data collection sources (i.e., RTMS, loops, and videos), it focuses on an introduction of the basic data processing methods, including the data format standardization, data spatial matching and data quality control.Subsequently, the thesis develops a dynamic traffic flow data fusion model and algorithm based on the multidimensional data sources. According to data spatial matching results, the data sources are divided into three dimensions:single source, dual sources, and triple sources. Further, the dynamic traffic data fusion technique based on the GA-BP Model is studied for the dual-source and triple-source dynamic traffic data. Finally, the thesis establishes the dynamic traffic information platform for Beijing based on three sources of data. It achieves the dynamic traffic flow information display by taking the early peak traffic flow data on April 6th and April 12th of 2009 as an example.