Department of Civil, Structural and Environmental Engineering, Engineering Statistics and Econometrics Application Research Laboratory, University at Buffalo, The State University of New York, 204B Ketter Hall, Buffalo, NY, 14260, United States.
Transport Research Institute, School of Engineering and the Built Environment, Edinburgh Napier University, 10 Colinton Road, Edinburgh, EH10 5DT, UK.
Accid Anal Prev. 2020 Apr;138:105361. doi: 10.1016/j.aap.2019.105361. Epub 2020 Feb 24.
This paper investigates the effect of High Visibility Enforcement (HVE) programs on different types of aggressive driving behavior, namely, speeding, tailgating, unsafe lane changes and 'other' aggressive driving behavior types (occurrence of not-yielding right-of-way and red light or stop signs violations). For this purpose, the Second Strategic Highway Research Program (SHRP2) Naturalistic Driving Study (NDS) data are used, which include forward-facing videos and time series information with regard to trips conducted at or near the locations of HVE implementation. To capture the intensity and duration of speeding and tailgating, scaled metrics are developed. These metrics can capture varying levels of aggressive driving behavior enabling, thus, a direct comparison of the various behavioral aspects over time and among different drivers. To identify the effect of HVE and other trip, driver, vehicle or environmental factors on speeding and tailgating, while accounting for possible interrelationship among the behavior-specific scaled metrics, Seeming Unrelated Regression Equation (SURE) models were developed. To analyze the likelihood of occurrence of unsafe lane changes and 'other' aggressive driving behavior types, a grouped random parameters ordered probit model with heterogeneity in means and a correlated grouped random parameters binary logit model were estimated, respectively. The results showed that drivers' awareness of HVE implementation has the potential to decrease aggressive driving behavior patterns, especially unsafe lane changes and 'other' aggressive driving behaviors.
本文研究了高可视性执法 (HVE) 计划对不同类型的攻击性驾驶行为的影响，即超速、跟车行驶、不安全的变道以及“其他”攻击性驾驶行为类型（未让路和闯红灯或违反停车标志的行为）。为此，使用了第二战略公路研究计划 (SHRP2) 自然驾驶研究 (NDS) 数据，其中包括与在 HVE 实施地点或附近进行的旅行有关的正面视频和时间序列信息。为了捕捉超速和跟车行驶的强度和持续时间，开发了比例指标。这些指标可以捕捉到不同程度的攻击性驾驶行为，从而可以直接比较随时间和不同驾驶员的各种行为方面。为了确定 HVE 和其他旅行、驾驶员、车辆或环境因素对超速和跟车行驶的影响，同时考虑到特定于行为的比例指标之间可能存在的相互关系，开发了看似不相关回归方程 (SURE) 模型。为了分析不安全变道和“其他”攻击性驾驶行为类型发生的可能性，分别估计了具有均值异质性的分组随机参数有序概率模型和具有相关性的分组随机参数二项逻辑模型。结果表明，驾驶员对 HVE 实施的意识有可能减少攻击性驾驶行为模式，特别是不安全的变道和“其他”攻击性驾驶行为。
