Research On Microscopic Simulation Modeling For Urban Mixed Traffic Flow

Mixed traffic flow which consists of motor vehicles and non-motor vehicles is still animportant characteristic of urban traffic in our country, especially in small and medium-sizedcity. The microscopic simulation modeling for mixed traffic flow can provide effectivetechnical means for the reproduction and analysis of the actual traffic phenomenon and theevaluation of traffic system. This paper focuses on modeling, implementation and applicationof urban mixed traffic flow microscopic simulation system. The study and discussion includethe urban mixed traffic micro simulation system structure, the simulation road network model,the mixed traffic simulation model with two-wheel vehicles, the design and implementationof microscopic simulation system and the application in the traffic signal control and so on.Researches mainly focus on the following:(1) In this paper, based on the construction of mixed traffic microscopic simulationsystem, the architecture and the components of the simulation system have been analyzed anda composite road network model for the mixed traffic micro-simulation system has beenproposed. The urban mixed traffic is composed of motor vehicles, non-motorized vehicles andpedestrians, so the composite road network model combines the continuous road networkmodel of motor vehicles and the discrete road network model of non-motor vehicles andpedestrians, and the mapping relationship of this two type road network model is established.The traffic flow models of microscopic traffic simulation system are analyzed, and the normaldriving and lane changing driving behaviors on the road are described using Lane DrivingTrajectory and Vehicle Dynamic Driving Trajectory.(2) Aimed at the present situation of urban mixed traffic, especially small andmedium-sized city traffic characteristics of two-wheel vehicle mixed traffic flow, atwo-dimensional cellular automata (CA) simulation model for two-wheel vehicle mixed trafficflow is established based on traditional NS model. The lane changing rules are proposedaccording to the driving features of the two-wheel vehicle which take into account the state ofvehicles on both sides of the adjacent lanes in the current and next time steps so that it makesthe model more realistic. By means of computer simulation, the influences of therandomization probability, lane-changing probability on two-wheel vehicles traffic flow, thetwo-wheel vehicles mixed ratio and pedestrian crossing behavior are studied, which canexplain some traffic characteristics of two-wheel vehicle flow and the actual trafficphenomenon. The practical verification shows the validity of the model.(3) With the analysis of the driving characteristics of non-motor vehicles, the paper puts forward the concept of the velocity transverse compression effect of non-motor vehicles andbuilds a multi-value CA model named as V-EBCA. For security reasons, with the increase ofthe velocity of non-motor vehicle on the road, the transverse road width occupied by thenon-motor vehicle will increase, and the available residual transverse road width for othervehicles will be narrowed, rendering the squeezing effect of the transverse residual width ofthe road. Considering this effect, a multi-value CA model has been used to model andsimulate the non-motor vehicle traffic flow. Based on the analysis of traditional multi-valuedcellular machine model (BCA and EBCA), a high-speed EBCA model with a maximum speedof3(cell/s) and a new V-EBCA model for non-motor vehicles have been established, and thecomparisons and analysis of simulation results and the model verification have been carriedout.(4) The design and development of the urban mixed traffic flow microscopic simulationsystem are executed in this paper including the detailed design and function module design.With the object oriented programming method, a small-scale microscopic traffic simulationsystem is developed and realized.(5) Using the self-developed microscopic traffic simulation system for the mixed traffic,the microscopic simulation analysis for traffic flow are carried out under three kinds ofsituations: the actual mixed traffic section, the intersection with fix-time control and theintersection with bus-priority intelligent control. The simulation results of the actual mixedtraffic section and the intersection with fix-time control using the proposed mixed trafficmodel based on the composite road network model show that the effects are relatively close tothe real traffic. A bus-priority intelligent control method for signalized intersection isproposed. In this method, the degree of traffic signal phase emergency is measured by thepriority values which include the capacity priority value and the threshold priority value. Thecapacity priority is used to measure the phase weighted number of vehicles and the thresholdpriority is for the special case of emergency. The bus-priority phase selection method isproposed based on priority values. The green time is determined by the two-level neuralnetwork fuzzy controller which the fuzzy domain is adjustable and the neural network isoptimized by Particle Swarm Optimization. Compared with the fixed-time control andtraditional fuzzy control method, the simulation results executed by the self-developedmicroscopic traffic simulation system show that this method can effectively improve theutilization rate of green time and reduce the bus stop delay. The influence on social vehicles isnot too much with appropriate parameters.