Microscopic Traffic Simulation Modeling For Urban Freeway

Urban freeways, undertaking the long-distant and rapid transportation of city, play an important role in relieving the stress of heavy traffic brought about by the rapidly increase in the needs of traveling. However, with quick economic growth, this kind of needs as well as initial underestimation of it make the current freeways have not enough capacities to accommodate more and more vehicles, leading to many serious troubles such as traffic congestion, frequent accidents and environmental pollution, etc. On the one hand, to solve these problems, building more freeways to meet the needs is of course a feasible approach. On the other hand, it is another effective way to make use of the current road resources efficiently with reasonable management.Nowadays, as a flexible, safe and low-cost means of studying on traffic phenomenon, traffic simulation systems are more often used successfully in some fields of traffic engineering, such as road design, traffic management, analysis of traffic safety etc. Based on the traits of urban freeways, developing traffic simulation software accordingly will contribute to the understanding of the phenomenon of freeway traffic and seeking the approaches to various problems occurring on the freeways.In this thesis, a traffic simulation system is developed for a road segment of the North-south Viaduct in Shanghai, the length of which is about 4.5 kilometers. At the current stage, our aim is to establish the models for the critical components of freeway environments and develop a basic simulation system to make the results of simulation fit the actual traffic states as well as possible. Firstly, driver-vehicle unit is modeled to describe the characteristics of drivers and vehicles. Secondly, by means of analysis and comparison between several classical car following models, the Generalized Force Model is chosen according to the traits of freeways and used as the car following model in our simulation system to describe the drivers' following behaviors. Three stages of drivers' discretionary lane changing behaviors are modeled to characterize the demand generation, gap acceptance and execution of lane changing, respectively. In addition, in the thesis, we put forward an idea that the drivers' desired speeds are variable and focus more on how to model the variation of derivers' desired speed at high flow. Finally, a simulation experiment has been conducted and the results show that the simulation data fit the field data preferably and then these models are validated.