Study On Integration Techniques And Methods About The Mobile Transport Management

Traffic congestion is a worldwide problem and how to control it has caused wide public concern. With the development of technology on terminals, software and internet, we have entered into the era of ―mobile internet‖ and ―big data‖. Under this circumstance, the fundamental way to solve traffic congestion problem is to achieve balance between supply and demand in transport network. And the way to achieve balance is to improve the utilization efficiency of existing transport infrastructure and the management of transport. The improvement on overall efficiency of urban road network is subject to the rational allocation of transport resources and scientific decisions made by transport administrators. It has been proved by practice that to improve the performance of overall transport network, relying on traditional Dynamic Transport Assignment model, has faced many difficulties, such as spatiotemporal variation of transport demands, complexity of road network, dynamic transport bottleneck location, inadequate adaptability to emergencies etc. However, the potential knowledge and laws, such as spatiotemporal transport pattern mining, spatiotemporal similarity clustering, congestion propagation etc, behind the massive transport data accumulated in traffic information platform could be converted to rules and methods for advance warning and diverting after analysis and processing. They could provide scientific and effective support for transport administrators on solving traffic congestion and improving overall performance of road network.In this dissertation, it is proposed to establish the framework of a mobile transport management system to improve the overall efficiency of transport network and achieve the balance between supply and demand in transport. The main researches include:(1)Establishing framework of collaborative transport information acquisition and sharing;(2) Finding sptiotemporal variation transport pattern and dynamic transport bottleneck location;(3) Influencing and managing demands in transport;(4)Analyzing the transport patterns and laws of urban transport network with the big data in application, summarizing the potential transport knowledge, raising advance warning and processing transport congestion, to maintain effective operation of entire transport network.This framework contains four sections, which are acquisition and generation of transport information, integration of multi-source heterogeneous transport information, acquisition of transport knowledge and application. It is also proposed two ways to collect real-time transport information, Location-based Service(LBS) and framework of collaborative transport information acquisition and sharing. All data collected are stored in Transport Information Database(TIDB) which is the database as well for further data analysis to find various Spatiotemporal Pattern(STP) and dynamic transport bottleneck location. In this dissertation, the transport data and knowledge extracted from the data are applied to two systems, travel time prediction system and transport management decision support system. The former system provides inquiry function of shortest travel-time route to travelers, meanwhile the latter one provides suggestions to transport administrators on decisions to improve the efficiency of transport network. Both systems adopt technique of database system while the model and algorithm design derive from existing data models and analytical methods with the collaboration from experts and scholars in relevant domains. This research is based the meta-rules contributed by experts and scholars and new rules discovered in this dissertation. New rules are the hidden patterns and laws mined, via scientific analysis, from the transport phenomenon which is in the form of data recorded. New rules can be guidance for practice and can be summarized from new phenomenon in practice again. Not only is it the methodology in era of mobile internet and big data, but also is ultimate idea of this dissertation.The main contributions of this dissertation are as below,(1)Establishing a framework of transport management and knowledge acquisition, with mobile internet and big data as background, proposing concept of mobile transport management and laying the groundwork for further development of smart transport system;(2)Proposing several methods to find the spatiotemporal traffic patterns and bottlenecks in the way of analyzing and mining spatiotemporal data;(3)Proposing dynamic hydrid travel time prediction method, which is much more precise than the prediction method solely dependent on real-time travel time and travel time recorded historical traffic data;(4)Indicating that the system of transport resources allocation and decision support, which adopts the knowledge of spatiotemporal transport pattern, dynamic transport bottlenecks, traffic status prediction etc, is more applicable on enhancing the performance and management of urban transport network than traditional transport allocation way.