Sharma Vinay Kumar, Sil Gourab
Department of Civil Engineering, Indian Institute of Technology, Indore, India.
Traffic Inj Prev. 2025;26(6):739-747. doi: 10.1080/15389588.2025.2450710. Epub 2025 Jan 29.
The study aims to develop continuous trajectory profiles along curves with minimal error. It also focuses on formulating a percentage trajectory transection rate model as a function of geometric parameters (e.g., radius and curve length) for through and passing maneuvers on four-lane divided highways. Additionally, the study sought to identify the critical maneuver between through and passing on horizontal curves of four-lane divided highways.
A total of 15 horizontal curve sites on a four-lane divided rural highway in plain terrain were selected. An instrumented vehicle, equipped with a GPS data logger and cameras, was used for data collection. Naturalistic driving data from 34 drivers' trips across the curves were collected for analysis, ensuring free-flow conditions. This dataset was used to develop continuous trajectory profiles, with GPS positioning errors minimized by calibrating the profiles at the Point of Curvature (PC) and Point of Tangency (PT) using the vanishing point method. Moreover, multiple linear regression technique was used to develop the mean percentage trajectory transection rate models i.e., for through () and passing () maneuvers.
The trajectory profiles were effectively developed using the vanishing point method, which restricts the mean absolute error to 16 cm. The developed models revealed that curve length significantly influences PTTR in through maneuvers, while deflection angle, approach tangent length and curve length are the most critical factors in passing maneuvers. Notably, curve length emerged as a common predictor for both maneuvers. Additionally, a 15% higher trajectory transection rate was observed in passing maneuvers.
Combining geometric features such as curve length, deflection angle, and approach tangent length provides a more comprehensive understanding of drivers' trajectory behavior, especially during passing maneuvers. The developed models can serve as a supplemental tool in geometric design to minimize lateral deviations and trajectory transitions across horizontal curves on four-lane divided highways. Passing maneuvers are riskier than through maneuvers, often involving lane changes and unpredictable paths. This highlights the need for roadway departure safety measures such as raised pavement markings, centerline rumble strips, and passing restrictions at specific curves.
本研究旨在开发误差最小的沿曲线的连续轨迹剖面。它还着重于制定一个百分比轨迹横断率模型，该模型是四车道分隔高速公路上直行和超车操作的几何参数（如半径和曲线长度）的函数。此外，该研究试图确定四车道分隔高速公路水平曲线上直行和超车之间的关键操作。
在平原地形的一条四车道分隔农村公路上共选择了15个水平曲线地点。使用一辆配备了GPS数据记录器和摄像头的仪器车辆进行数据收集。收集了34名驾驶员通过这些曲线的行程中的自然驾驶数据进行分析，确保自由流条件。该数据集用于开发连续轨迹剖面，通过使用消失点法在曲率点（PC）和切点（PT）对剖面进行校准，将GPS定位误差降至最低。此外，使用多元线性回归技术开发平均百分比轨迹横断率模型，即用于直行（）和超车（）操作。
使用消失点法有效地开发了轨迹剖面，将平均绝对误差限制在16厘米。开发的模型表明，曲线长度在直行操作中对PTTR有显著影响，而偏转角、接近切线长度和曲线长度是超车操作中最关键的因素。值得注意的是，曲线长度是两种操作的共同预测因子。此外，超车操作中的轨迹横断率高出15%。
结合曲线长度、偏转角和接近切线长度等几何特征，可以更全面地了解驾驶员的轨迹行为，尤其是在超车操作期间。开发的模型可作为几何设计中的补充工具，以最大限度地减少四车道分隔高速公路水平曲线上的横向偏差和轨迹过渡。超车操作比直行操作风险更大，通常涉及车道变换和不可预测的路径。这突出了道路偏离安全措施的必要性，如凸起的路面标记、中心线隆隆带以及特定曲线处的超车限制。
