Department of Civil and Environmental Engineering, Florida International University 10555 W. Flagler Street, EC3725, Miami, FL, 33174, United States.
Accid Anal Prev. 2020 Feb;135:105355. doi: 10.1016/j.aap.2019.105355. Epub 2019 Dec 5.
Large truck rollover crashes present significant financial, industrial, and social impacts. This paper presents an effort to investigate the contributing factors to large truck rollover crashes. Specific focus was placed on exploring the role of heterogeneity and the potential sources of heterogeneity regarding their impacts on injury-severity outcomes. The data used in this study contained large truck rollover crashes that occurred between 2007 and 2016 in the state of Florida. A random parameter ordered logit (RPOL) model was applied. Various driver, vehicle, roadway, and crash attributes were explored as potential predictors in the model. Their impacts were examined for the presence of heterogeneity. Interaction effects were then added to the random variables in order to detect potential sources of heterogeneity. Model results showed that the impacts of lighting conditions and driving speed had significant variation across observations, and this variation could be attributed to driver actions and driver conditions at the time of the crash, as well as driver vision obstruction. Findings from this study shed light on the direction, magnitude, and randomness of the factors that contribute to large truck rollover crashes. Findings associated with heterogeneity could help develop more effective and targeted countermeasures to improve freight safety. Driver education programs could be planned more efficiently, and advisory and warning signs could be designed in a more insightful manner by taking into account specific roadway attributes, such as sandy surfaces, downhill, curved alignment, unpaved shoulders, and lighting conditions.
大型卡车翻车事故造成了重大的财务、工业和社会影响。本文旨在研究导致大型卡车翻车事故的因素。研究特别关注探讨异质性的作用以及异质性对伤害严重程度结果的潜在来源。本研究使用的数据包含了 2007 年至 2016 年佛罗里达州发生的大型卡车翻车事故。应用了随机参数有序逻辑回归（RPOL）模型。模型中探讨了各种驾驶员、车辆、道路和碰撞属性，作为潜在的预测因素。研究了它们对异质性存在的影响。然后，将交互效应添加到随机变量中，以检测潜在的异质性来源。模型结果表明，照明条件和行驶速度的影响在观测值之间存在显著差异，这种差异可以归因于碰撞时驾驶员的行为和条件以及驾驶员的视线受阻。本研究的结果揭示了导致大型卡车翻车事故的因素的方向、大小和随机性。与异质性相关的研究结果可以帮助制定更有效和有针对性的措施来提高货运安全。通过考虑特定的道路属性，如沙质表面、下坡、弯道、未铺砌的路肩和照明条件，可以更有效地规划驾驶员教育计划，并以更有见地的方式设计咨询和警告标志。
