Dynamic Forecast Method For Traffic Congestion At Signalized Intersection

The traffic congestion and traffic accident are the two major problems for traffic engineering. Doing more research on the traffic congestion forecast is a crucial element in preventing and mitigating urban traffic congestion problems theoretically and practically.First of all, the mechanism of urban traffic congestion at intersection is analyzed, and three types of factors for traffic congestion are summed up: too large traffic demand, not enough traffic supply and inharmony between these two factors. In addition, the research puts forward two kinds of traffic bottleneck: bottleneck point and bottleneck link, and two kinds of traffic congestion: bottleneck point congestion and bottleneck link congestion. Based on this analysis, four main issues are carried out in the following researches: traffic congestion quantification, traffic congestion forecast, analysis for forecast model, and application for forecast model.In the research of congestion quantification, this thesis puts forward two basic kinds of quantification models: demand-supply model for bottleneck point congestion and link loading-rate model for bottleneck link congestion, and two improvement models for ground traffic congestion problem: demand-supply with arrival rate model and link loading-rate with turning rate model. And then link loading-rate with turning rate model is applied and analyzed in three basic types of channeling sections. The advantage and applicability for these four models are also compared in detail.In the research of traffic congestion forecast, dynamic forecast models for traffic congestion at signalized intersection are proposed. This paper presents the method of determining the bottleneck section, concludes two types of congestion occurring situstion, then, bring forward the dynamic forecast models for traffic congestion. Moreover, for the unstable character in traffic system, probability model, which can be more reasonable to describe the congestion state, is put forward for traffic congestion forecast. In addition, the link or network reliability, which is useful for the network reliability research on transportation, is defined with the research of probability model.Based on the traffic congestion forecast, six types of traffic state spaces are put forward to describe the urban road traffic condition, with "space E" is bottleneck point congestion space and "space F" is bottleneck link congestion space. Meanwhile, the evolved law and space character are analyzed in detail, which can provide some basic rules for the determination of traffic management and traffic control strategy.In the research of analysis for forecast model, the effect of parameters in forecast model is analyzed. Firstly, the sensitivity of three main parameters—forecast time, link input/output and effect of relationship (flow relationship and control signal relationship) between input and output-is analyzed. And then based on the analysis, this thesis proposed a basic quantificational utility model which can provide some basic idea for the choice of traffic congestion mitigation strategy.In the research of application for forecast model, based on the analysis and forecast, this thesis provides a basic flow process for preventing and mitigating urban traffic congestion, and optimized method for the traffic management and control strategies. Moreover, five types of traffic congestion mitigation strategies are concluded: optimization in traffic plan level, guidance in traffic policy level, optimization in traffic management level, optimization in traffic control level and optimization in traffic design level. In previous research, the reduction of traffic demand strategy in control and design level do not been considered seriously and the pursuit of local optimization may cause the benefit damage of the whole network, this problem is demoed with the simulation. In addition, this thesis brings forward a "Four-Step Model" for preventing and mitigating traffic congestion which is significant for the improvement of network reliability and travel efficiency.In the finality, this thesis sums up the main conclusions, disadvantages and applications: initiative traffic control, initiative traffic management, and network traffic state forecast and analyzed. However, more researches are still needed, such as some application detail in different case, initiative traffic engineering, traffic information predict and data fusion for different kinds of traffic data.