Research On Traffic Flow Characteristic With Lane-changing Based On A Cellular Automaton

Lane-changing is a common behavior in multilane traffic flow. There is a famous paradox “faster is slower”, that is, to change lanes for speeding up, on the contrary to slow down. However, this paradox does not always come into existence, but take place in special range.Inspired by thought that combining macro and micro, a cellular automaton model that considering macroscopic property of the continuous traffic flow characteristics was built, which is based on the analysis of lane-changing behavior, mixed traffic flow characteristics, and the continuous traffic flow characteristics.In view of continuous traffic flow on urban unidirectional road segment, symmetrical lane-changing rules were used to analyze the changes of the overall state of traffic flow, which were caused by free lane-changing. It was found that the effect on traffic flow switched from good to bad gradually with the increase of vehicle arriving probability. In order to determine the critical range of good effect or bad effect, average delay and transitable flow were set as guideposts. Then surfaces were obtained, which adopted the arriving probability of vehicles as well as the green ratio as dual independent variables, and guideposts as dependent variable. Thus by the comparison of two surfaces, free lane-changing and straight proceeding, the horizontal projection of the intersecting lines was solved, and the critical range of occurrence for the paradox was obtained. In the following sections, three parts of the state of the traffic flow(normal region, transition region, and paradox region) were discussed, and the reasons for paradox caused by free lane-changing were analyzed, as well as corresponding conclusions and suggestions were acpuired.In view of overtaking traffic flow on the bi-directional road segment, asymmetrical lane-changing rules were used to comprehensively simulate the overtaking phenomenon in one-way traffic flow, asymmetry and symmetry traffic flow, respectively. Average density, average velocity, average delay and transitable flow were set as guideposts. Then surfaces were obtained, which adopted the arriving probability of vehicles in the home-lane and the passing-lane as dual independent variables, and guideposts as dependent variable. Afterwards, several critical ranges that divided positive and negative effect of overtaking behavior were acpuired. In the following sections, different types of road mid-line were simulated, thus the best type on the bi-directional road segment was determined by qualitative and quantitative analysis.In this article, after considering the continuous traffic flow characteristics, a cellular automaton model was set up. In theory, traffic flow theories about undirectional and bi-directional were proposed. In practice, the paradox was verified, and the critical ranges of both free lane-changing and overtaking were determined. These ranges were applied to various road and traffic facilities, and could be used to improve traffic management or optimize traffic organization to some extent.