基于实测数据的信控环交驾驶行为建模与仿真

通过图像识别得到单进口轮流放行信控环交的车辆轨迹数据，发现车辆在信控环 交内的换道行为主要集中在入环阶段和出环阶段发生.在换道空间分布特征及其原因剖析的 基础上，引入驾驶心理参数，分别建立了入环、环中阶段和出环阶段的信控环交元胞自动机模 型.模型的数值模拟结果表明，由于进口道车道功能未渠化，车辆在环内换道引发的冲突与阻 滞是导致信控环交运行不畅的重要原因，换道动机与时机取决于转向比例与渠化方案，与进 口道绿灯时间的增减相关性甚微，过长的放行时间反而会导致环内车流密度加大，换道所造 成的阻滞更难以消散，不利于信控环交的运行.

Driving Behavior Model and Simulation of the Signalized Roundabout Based on Vehicle Trajectory

By analyzing the vehicle trajectory data of the single-approach-entering operation in signalized roundabout, which is obtained by the image recognition, it is found that vehicles’lane-changing behavior mainly concentrated in the entry and exit area. Based on the analysis of the spatial distribution characteristics and reasons, the driving psychological parameter is introduced and the cellular automaton models for the different areas of the roundabout are established. Numerical simulation results indicate that, due to entrance lane’s construction not channelized, retardation that induced by lane-changing and traffic conflicts is the main factor of traffic congestion in signalized roundabout, and lane-changing motivation and occasion depends on turning ratio and channelization, which has little relevance to the length of entrance’s green time, and given long time may cause traffic flow density within roundabout increasing, then the retardation caused by lane-changing, which may block the operation of signalized roundabout, is more difficult to dissipate.