Font Size: a A A

Urban Mixed-Traffic Flow Modeling Techniques Based On Cellular Automaton

Posted on:2005-09-08Degree:DoctorType:Dissertation
Country:ChinaCandidate:J ZhangFull Text:PDF
GTID:1102360152970884Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
Transportation is significant on tactical development of national economy and is always the key factor in country building. With rapid increasing number of many kinds of vehicles, traffic jam is more serious and even traffic accident happens sometimes. These induce a set of severe problems on society and environment. Additionally, these restrict economic developments strongly. Therefore, transportation obtains more and more attentions of government officials, scientific research institutions, and academic communities, even citizens. Transportation system is a complicated one with serious non-linear, strong stochastic, large time varying characteristics and uncertainties, so it is much difficult to be controlled. Besides making the most use of road resources in existence, the utility of scientific traffic flow theory on reasonable traffic layout, traffic control and management is more important to solve traffic problems. Traffic flow theory is a developing scientific. Though there are theory productions that already probe into traffic phenomena, the integrated theory system has not been formed till now. The traffic flow in our country is mixed-traffic flow, which is composed of motor vehicles, no-motor vehicles and pedestrians. The particular composing doubles the number of traffic conflicts, and increases the probability of traffic jams and accidents. These bring on more difficulties on traffic flow organization, control and management. So special severe requirements are put on our researchers. Thus, researching on traffic flow theory, especially mixed-traffic theory, not only has application background but also research meaning on theoretical.Combined with research on traffic flow theory of cellular automaton (CA) and the moving characteristics of motor vehicles, no-motor vehicles and pedestrians, this dissertation realizes the motor vehicle CA model, bicycle CA model and pedestrian crossing CA model respectively. And then proposes a mixed-traffic flow model on these work. At last, a distributed mixed-traffic flow model based on techniques of multi-agent systems (MAS) is designed and a structure of urban mixed-traffic simulation system is brought forward. The main results achieved in this dissertation can be summed up as follows.a) A one-dimension CA traffic flow model of motor vehicle, named two-car following CA model, is proposed. The model can reveal traffic flow's metastable characteristics with better robustness and higher critical density. And the update rule is completely parallel, so it can adapt to large-scale traffic simulations. Theoretical analysis and experimental results reveal that the metastable character is relevant withthe relation degree between cars and the random slow probability.b) A one-dimension CA traffic flow model of motor vehicle, named variable time-headway CA model is presented. The model assumes that every vehicle tends to keep expected time-headway. The vehicle synchronous updates its velocity by estimating the velocity of head vehicle and the distance-headway. The distributing of the simulation system's time-headway can be tuned by two simple parameters according to practical unsaturated traffic characteristics. Thus the simulation system can describe rather complicated traffic states on multi-lanes highway and indicate traffic capacity of different highway grade.c) A one-way bicycle model based on two-dimension CA is brought forward for the first time. Theoretical analysis and experimental results indicate that the model describes the moving characteristics of bicycle flow effectively, and the simulation data on velocity-density-flux accord with practical or experiential data. The influence of lane number on bicycle capacity is investigated using this model. The simulation result shows that one-lane capacity increases alone with the added lane width, but the increasing magnitude is descending and will cease till a certain number of lanes. This trend completely obeys qualitative analysis. The model can be conveniently extended. More importantly, it can offer theoretical base of furt...
Keywords/Search Tags:traffic simulation systems, cellular automaton, traffic flow model, bicycle model, pedestrian's crossing model, mixed-traffic flow, agent, multi-agent systems
PDF Full Text Request
Related items