Abstract: In order to reduce the traffic accidents of pedestrians and vehicles crossing the street at non-signalized intersections in low visibility, a study on the identification of vehicles and pedestrians crossing the street considering the impact of visibility was carried out. Combined with the relative position, velocity, acceleration, vehicle size and other information of vehicles and pedestrians, the conflict recognition model between pedestrians and vehicles was constructed. The visibility measurement method of traffic environment based on the distance between people and vehicle was determined, and the relationship model between vehicle speed and visibility was given. On this basis, the proposed model was modified, and the effectiveness of the proposed model was verified. The results show that the proposed conflict recognition model can effectively identify the conflicts between pedestrians and vehicles crossing the street, and the accuracy of conflict recognition is 82.4%. This study can provide decision-making for pedestrians and vehicles at non-signalized intersections with intelligent vehicle infrastructure cooperative systems, and thus improve the safety of pedestrian and vehicle in low visibility.

Key words: traffic engineering, visibility, pedestrians crossing the street, conflict identification, traffic safety

摘要： 为减少低能见度下无信号交叉口过街行人与车辆的交通事故，开展了考虑能见度影响的车辆与过街行人冲突识别研究。结合车辆行人相对位置、速度、加速度、车辆尺寸等信息，构建了过街行人与车辆冲突识别模型，确定了基于人-车间距的交通环境能见度测量方法，给出车辆速度与能见度之间的关系模型，在此基础上对模型进行修正，并验证了模型的有效性。结果表明：该冲突识别模型可对过街行人与车辆冲突进行有效识别，冲突识别的准确率为82.4%，该研究可为车-路协同下的无信号交叉口行人和车辆冲突识别提供决策，进而提高低能见度下行人与车辆的安全性。

关键词: 交通工程, 能见度, 行人过街, 冲突识别, 交通安全