Mesoscopic Traffic Flow Model Basing On Wiedemann Car Following Behavior

Traffic congestion has become a major bottleneck restricting urban development.Establishing a scientific,effective and efficient traffic flow model is an important mean of recognizing and understanding traffic flow characteristics,which can provide a theoretical support for the measures of alleviating traffic congestion.In this paper,we start by analyzing the shortage of Nelson’s traffic flow model,and introduce Wiedemann’s car-following behavior model to establish the new mesoscopic traffic flow model.We obtain traffic flow characteristic by processing and analyzing the traffic flow data of east second ring road in Changsha,The numerical calculation verifies the improved model can better reflect actual operating characteristics of urban traffic flow.The main results are as follows:Firstly,Nelson’s model improved the Prigogine-Herman model by using correlation model and mechanical model.However the acceleration and deceleration in the mechanical model are so large that it is inconsistent with the actual situation.In this paper,we introduce the Wiedemann’s car-following behavior model to improve the Nelson’s mechanical model.The speed change is determined by judging which behavior section the vehicle situation locates.The mesoscopic traffic flow model integrating Wiedemann’s model is proposed.Secondly,traffic flow data are collected by roadside laser detector on east second ring road.According to Wiedemann’s behavior threshold model,the statistical results of behavioral region in speed difference and headway plot are fitted by 1stopt software and the corresponding joint probability function of speed difference and headway is obtained.To a certain extent,the problem of mathematical assumption without real data suppot is partly solved.Thirdly,collected traffic flow data of east second ring road are processed and traffic flow characteristics such as flow rate,speed,density,speed difference and headway are analyzed.These data are used to analyze and evaluate the proposed model in this paper.