This study sheds light on pedestrian and driver's behavior towards crossing in Kuwait. It examined the effect of different factors on pedestrian and vehicle violations at 10 locations in Kuwait including signalized, nonsignalized intersections, locations with/without pedestrian's bridge, and locations with/without pedestrian crossing signals. Four types of pedestrian's violations were observed; crossing outside the marked crossway, not obeying traffic signal, distracted crossing (eating, drinking, or using a smartphone while crossing), and not using pedestrian's bridge. For vehicle violations, two types were observed; not yielding to pedestrians and stopping over the marked crossway. The pedestrian and vehicle violations observed are considered as “violations” according to the existing Highway Code in Kuwait. A total of 24,393 pedestrian crossing violations were recorded over 40 days. Pedestrian violations were analyzed using a multinomial logit model while vehicle violations were analyzed by frequencies analysis. Video data were used for validation applying a trajectory-based approach showing a high accuracy level of 98.85%. The results showed that all the studied factors; time of observation including weekdays and weekends, and peak and off-peak hours, location of observation, demographic characteristics including pedestrian gender, and age, and site properties (with/without pedestrian's bridge and with/without pedestrian crossing signal) were significant and had a strong relationship with the type of violations. The results also indices the importance of providing pedestrian's bridge in locations with a high number of violations.

中文翻译：

---

检查科威特的行人过马路行为

这项研究揭示了行人和司机在科威特过马路的行为。它在科威特的 10 个地点检查了不同因素对行人和车辆违规行为的影响，包括信号灯、非信号灯十字路口、有/没有人行天桥的地点，以及有/没有人行横道信号的地点。观察到四种类型的行人违规行为；在标记的十字路口外过马路、不遵守交通信号、分心过马路（在过马路时吃、喝或使用智能手机），以及不使用人行天桥。对于车辆违规，观察到两种类型；不礼让行人并在有标记的十字路口停车。根据科威特现行公路法规，观察到的行人和车辆违规行为被视为“违规”。共24个，40 天内记录了 393 起行人过马路违规行为。行人违规使用多项式 Logit 模型进行分析，而车辆违规则通过频率分析进行分析。视频数据用于验证应用基于轨迹的方法显示 98.85% 的高精度水平。结果表明，所有研究因素；包括工作日和周末、高峰和非高峰时段在内的观察时间、观察地点、包括行人性别和年龄在内的人口统计特征以及站点属性（有/没有人行天桥和有/没有人行横道信号）是显着的并且与违规类型有密切关系。结果还表明了在违规行为较多的地点设置人行天桥的重要性。