无控制交叉口交通安全影响因素分析

Analysis on Factors Affecting Traffic Safety at Uncontrolled Intersection

为了探究无控制交叉口的交通安全影响因素,降低无控制交叉口的事故风险与安全隐患,采用间接安全评价方法--交通冲突技术分析了成都市7个无控制交叉口。首先,以后侵入时间作为冲突严重程度的定量指标,将冲突严重程度分为潜在、一般、严重冲突这3个有序等级。再采用广义有序logit模型全面鉴别对冲突严重程度具有显著影响的车、路和环境等相关因素。然后,通过边际效应计算反映各种影响因素对交通冲突严重程度等级的影响作用程度。研究结果表明,发生分流、合流及直角冲突的严重程度普遍高于追尾冲突;交叉口车流量、交叉口是否渠化、车辆是否停车等都会显著影响冲突严重程度。其中发生分流、合流和直角冲突时,其严重冲突发生概率分别提高29.4%,38.5%和25.6%,冲突车辆未停车让行导致严重冲突发生的概率提高11.5%,而渠化交叉口可使严重冲突的发生概率降低17.3%。最后根据模型回归和边际效应计算结果,有针对性地提出了无控制交叉口的安全改善建议。

In order to explore the factors that affect the traffic safety at uncontrolled intersections and reduce the crash risk and the potential safety hazards at uncontrolled intersections, the traffic conflict technique, an indirect safety assessment method, is adopted to analyse 7 uncontrolled intersections in Chengdu City. First, taking post-encroachment time as the quantitative indicators of the conflict severity, the conflict severity is classified into 3 ordered levels: potential, slight, and serious conflicts. The factors that significantly affect the conflict severity, in terms of vehicle, road, and environment, are comprehensively identified by generalized ordered logit model. Then, the marginal effects of the influencing factors are calculated for reflecting the influencing degree on each factor that affect the traffic conflict severity level. The result shows that(1) The conflict severities of diversion, merging, and crossing are higher than that of rear-end impact conflict.(2) The traffic volume, the channelization design, and the vehicle stopping at intersection will have significant influences on the severity of traffic conflict. If diversion, merging, and crossing occur, the corresponding probabilities of being serious conflict increased by 29.4%, 38.5%, and 25.6% respectively. The probability of the occurrence of serious conflict increased by 11.5% if the conflict vehicle does not stop, while it decreased by 17.3% if the intersection is well channelized. Finally, several suggestions are put forward for the improving safety of uncontrolled intersections according to the model regression and marginal effect calculation.