Simulation Of Urban Road Traffic Flows By Cellular Automaton Approaches Based On Deceleration Effects

As a common phenomenon in urban road traffic, vehicle decelerating often influences the driving state of following vehicles, and even changes the status of the traffic flow. In the study of urban road traffic flow, vehicle decelerating must be treated. Research of the microscopic properties of urban road traffic flow under vehicle decelerating environment can offer accurate reference data and decision basis for traffic management and road traffic system optimization. Cellular automaton with flexible rules and the simple algorithm can be used to simulate the road traffic under various conditions. So cellular automata traffic flow models based on deceleration effects are established and are used to simulate space-time characteristics of traffic system in this thesis.(1) Driving characteristics of vehicles on urban road under the many different influence factors are analyzed and a cellular automata traffic flow model with deceleration effects is established, which is suitable for the urban environment and vehicle motion rules. Different with Nagel-Schreckenberg model(NS model), in which vehicle moves with its current speed at a timestep, the concept of "memory" speed of the vehicle is introduced and vehicle displacement is expressed by used the average velocity of "memory" speed and current speed at current time-step. Because deceleration effects are introduced and speed-updating rules are redefined in the cellular automata model, continuities of vehicles driving states are hold and the safe distances between vehicles are conformed to the physical rules much more. The simulation results show that motive path curves of vehicles and fluctuation of vehicles headways are all smoother in cellular automata models for urban traffic flow.(2) A dual-lane cellular automata model with deceleration effects is established based on single-lane cellular automata model. In this model, the lane-changing rules are conformed to the practical traffic states much better, which including vehicles lane-changing motivations are decided by expected velocities of the vehicles on current lane and the target lane and safety states are decided by related velocities of the vehicle on current lane and the following vehicles on the target lane. The simulation results show that the lane-changing rules are more flexible than the rules of symmetric two-lane cellular automata model(STCA model) and are conformed to the practical lane-changing conditions much better and the lane-changing efficiencies are improved under the same vehicles density in this model.(3) According to the characteristics of urban intersections, dual-dimensional cellular automaton mixed traffic flow models are used to simulate mixed traffic flow of urban intersections, which simplified cellular automata traffic flow models with deceleration effects. Because many vehicles would select to pass the intersection quickly when the lights are about to turn red in fact, many vehicles conflicts rules are redefined firstly in those models. The time-space evolutions characteristics of traffic flow under different roads intersections including isolated intersections and long-link grouped intersections and short-link grouped intersections are simulated. The effects on intersections traffic flow of different intersection spacing signal periods and phase offsets are analyzed. Also the relationships of intersection spacing and phase offsets under green-wave traffic signal control are obtained.(4) Dynamics characteristics of vehicles and behavioral characteristics of pedestrians at unsignalized and signalized crosswalks are studied. Two concepts about critical safe distances and critical decision distances are defined. Therefore, cellular automata models for crosswalks are established. In those models, the probabilities of pedestrian crossing crosswalks are estimated by using critical safe distances and critical decision distances and the probabilities of pedestrian to hazard are estimated according to pedestrian’s waiting time. Also, changing process of probabilities of pedestrian crossing crosswalks along with the changing of vehicles speeds and distances at unsignalized crosswalks are simulated, and the relationships between the probabilities of pedestrian to hazard and pedestrian’s waiting time threshold are researched, and the effects of different control signals on vehicles flow and pedestrian flow crossing crosswalks are analyzed.(5) Cellular automata models for different buses stopping styles are established. In those models, buses deceleration effects during forced lane-changing and many new lane-changing rules are adopted, which makes the vehicles changing lane modes closer to the reality. Using those models, vehicles driving characteristics with different buses stops such as non-harbor, harbor, and multi-platform buses stops are simulated, and the effects of the length of platforms and the distances between platforms on traffic flow are studied, and applicable situations of different buses stops are determined.(6) A traffic flow model combined with harbor-style buses stops and intersections are established. Based on the model, driving characteristics of different vehicles driving on the connection roads of intersections with buses stops are studied, and dual effects of buses stops and intersections on traffic flow under different signal cycles, offsets, buses stops locations and turning probabilities are analyzed and compared.