Research On Traffic Flow Characteristics Of Freeway Based On Cellular Automata

In China, traffic congestion still occurs frequently even though the actual traffic volume of many freeways has not exceeded their designed capacity. Two major reasons can account for this phenomenon:one is the moving bottleneck induced by an increasing number of slow trucks that often occupy the fast lanes, resulting in accidental congestion; the other is the fixed bottleneck resulting from the poor traffic facilities and improper control measures, leading to recurrent congestion. The present measures have failed to eradicate traffic congestion, which’s ultimate solution must be based on a thorough study of traffic flow evolution, contributing to the understanding of the essence of various traffic phenomena and the extraction of the evolution laws. This dissertation is aimed to analyze the traffic flow characteristics of four-lane freeways on which cars and trucks are allowed.In this dissertation, on the basis of the national traffic laws and the measured traffic data, a comprehensive analysis is conducted over speed limit rules, the longitudinal driving rules, lane-changing rules as well as the classification and running characteristics of vehicles on four-lane freeways. Then, this dissertation describes the longitudinal driving behaviors and the lane-changing behaviors, which can be subdivided into either voluntary or mandatory lane-changing. The voluntary lane-changing behavior can be further divided into tendentious and non-tendentious behavior. Based on this, lane-changing rules are classified into four categories:symmetric rule, asymmetric rule, old-law rule and new-law rule. The concepts of giving way and cutting-in behaviors are also defined. It also analyzes the road characteristics, longitudinal running characteristics and lane-changing characteristics of vehicles on special sections, including speed zone, long steep slope, and construction zone.Cellular automata, as a concise and effective model, has unique advantages in traffic flow modeling. Based on single-lane NaSch model, a two-lane cellular model is established to describe all kinds of lane-changing rules, driving behaviors and special sections. Cellular simulation software is developed through VC++ programming and the periodic boundary condition is adopted. In order to compare the difference when less trucks and more trucks are in system, the truck ratio is set at 0.05 and 0.5 normally. Traffic density ranges from 0 to 60veh/km. When analyzing the spatiotemporal or spatial characteristics of traffic flow, the density is generally set at 25veh/km. Then, the influence of control parameters, such as initial distribution pattern of vehicles, statistical time, sample size and cellular length, on simulated results is analyzed to determine all parameter values. The effectiveness of cellular model and the rationality of parameters are verified through measured traffic data.In this dissertation, factors influencing the traffic flow are investigated, including lane-changing rules, the ratio, length and performance of trucks, the probability of lane-changing and moderation, the behavior of giving-way and cutting-in and the running differences of same-type vehicles. This dissertation also explores the impact of speed limit value and deceleration behavior on speed zone, the impact of truck ratio and climbing lane on long steep slope as well as the impact of closedown of different lanes and traffic density on construction zone. The findings of this dissertation are demonstrated and analyzed in the following three aspects:(1) the correlation curve between density and such parameters as flow rate, speed, lane-changing rate, utilization rate of lane, conflict rate is drawn to analyze the running characteristics of traffic flow in a macroscopic perspective; (2) the spatiotemporal pattern of cellular evolution is drawn to analyze spatiotemporal distribution characteristics of vehicles; (3) the spatial distribution curve of flow rate, speed and density along the road is drawn to analyze the spatial distribution characteristics of traffic flow on special section.The following innovative findings are made in this dissertation:(1) by defining cellular rules of driving behaviors, like lane-changing behavior, a traffic flow cellular model of four-lane freeways is established to comprehensively demonstrate the traffic flow characteristics (2) cellular model is used to reveal the relative laws concerning the impact of such factors as lane-changing rules, truck ratio and car cutting-in on the traffic flow of four-lane freeways. (3) cellular models are established to describe the traffic flow characteristics on speed zones, long steep slopes and construction zones in China, and the running characteristics and vehicle flow laws in related sections are acquired through simulation analysis.The study can reinforce the knowledge and understanding towards the traffic flow characteristics of freeways from both microscopic and macroscopic perspectives, for example, the influencing laws and mechanisms of lane-changing rules and truck ratio on traffic flow and the traffic running laws in construction zones, providing theoretical foundation for the optimization of traffic design and management. The cellular model, when calibrated, can be employed in real traffic flow stimulation. If the correctness, accuracy and arithmetic speed of the model are satisfactory, it can also be applied in online simulation, short-time prediction and early warning of traffic status, so as to alleviate and prevent traffic congestion.