不同交通状态下直左交通冲突模型研究

Study on Direct-Left Traffic Conflict Model Under Different Traffic Conditions

在分析直左冲突与交通流影响关系的基础上,依据信号交叉口周期时长、饱和车头时距、相位绿灯时间和交通冲突影响系数,计算实际交通流量与道路通行能力的比值,依据该比值对信号控制交叉口的交通状态进行划分。并选择碰撞时间(time to collision,TTC)指标作为交通冲突的判定指标,以全贝叶斯方法为基础,构建独立状态下单维泊松对数正态贝叶斯组合回归(Poisson logarithmic normal Bayesian combination regression,PLCR)直左交通冲突模型和不同交通状态下多维泊松对数正态贝叶斯组合回归(multidimensional Poisson logarithmic normal Bayesian combination regression, MPLCR)直左交通冲突模型,分析表明:MPLCR模型的精度是PLCR模型的2.2倍左右,MPLCR模型可定量分析不同交通流与交通冲突的影响;交通状态不同,直行交通流和左转交通流对直左交通冲突的影响结果不同。

Based on the analysis of the relationship between direct-left conflict and traffic flow, and the length of the signal intersection period, the time distance of the saturated head, the phase green time and the influence coefficient of the traffic conflict, the ratio of the actual traffic flow to the road capacity is calculated, and the traffic state of the signalized intersection is divided according to the ratio. The time to collision (TTC) is chosen as the judgment index of traffic conflict. Based on the whole Bayesian methodology, the straight-left traffic conflict model of the one-dimensional Poisson logarithmic normal Bayesian combination regression (PLCR) in independent state and the straight-left traffic conflict model of multidimensional Poisson logarithmic normal Bayesian combination regression (MPLCR) in different traffic states are constructed. The analytical results show that the accuracy of MPLCR model is about 2.2 times that of PLCR model;the MPLCR model can quantitatively analyze the influence of different traffic flow and traffic conflict;the impact of direct traffic flow and left-turn traffic flow on direct-left traffic conflict is different under different traffic conditions.