城市快速路匝道合流区车辆交互行为模式

在城市快速路匝道合流区，驾驶任务难度主要来自于车辆与其周边车辆之间的动态交互，目前对这种交互行为的特征和机理的认识还不十分清楚。基于从无人机视频中提取的高精度车辆轨迹数据，提取出表征车辆交互行为的指标TTC和GAP，并结合速度、加速度、车道位置等其他指标，对车辆的交互过程加以刻画，从中获得了大量交互行为实例，并在此基础上归纳总结出9种典型的车辆交互行为模式。通过分析各模式特征发现：即使在相同的外部环境下，车辆交互行为模式也可能存在差异，这表明交互行为不仅与车辆之间的相对位置、时空距离、速度状态等环境因素有关，还与驾驶人的应对能力、动机及风险意识等认知心理有关；另外，不同的交互模式面临的风险不同，并且该风险既可能是周边车辆行为发生改变而被迫卷入，也可能是驾驶人自身主动寻求的结果；9种不同类型的交互行为模式，构成了驾驶人自行感知的4种风险状态互相转换的具体实现形式；在驾驶过程中，驾驶人努力寻找契机并选择某种交互行为模式在各个风险状态之间来回切换，并非仅由心理压力较大的危险态向压力较小的自由态转换，也会发生反向转换，前者主要由降低事故风险和减少认知努力的动机驱动，后者旨在追求行车效率，但同时驾驶人会付出更多的认知努力以对抗风险的增加，这充分反映了驾驶人试图在行车效率、事故风险与认知努力三方面取得平衡。研究成果对深化理解驾驶行为及其背后的决策机制具有积极意义。

Vehicle Interaction Patterns at On-ramp Merging Area of Urban Expressway

In the on-ramp merging area of urban expressways, the difficulty of driving tasks mainly arises from the dynamic interaction between vehicles and surrounding vehicles. However, the characteristics and mechanisms of these interactions are not well understood. In this study, based on the high-precision vehicle trajectory data extracted from UAV video, two indicators, time to collision (TTC) and GAP, were extracted and used to measure the vehicle interactions. Additionally, combined with several other measures including velocity, acceleration, and lane ID, multiple indicators were comprehensively realized to describe the characteristics of the vehicle interaction process. A large number of interaction examples were obtained. Accordingly, nine typical types of vehicle interaction patterns were proposed. By analyzing the characteristics of each pattern, it was found that the target pattern selected by drivers may differ, even if the drivers were under the same external context. This indicates that interaction patterns are not only related to external environmental factors, such as a vehicle's relative position, spatial and temporal distance, and speed value, but also to internal cognitive factors, such as a driver's capability, motives, and risk awareness. Moreover, each interaction pattern is accompanied by different driving risks, which may be caused by either forced involvement due to changes in the behavior of surrounding vehicles or the active seeking of a driver's own behavior. Further, the nine interaction patterns constitute the realization forms of the mutual transformation of the four different risk states perceived by the driver. While driving, drivers continuously try to find an opportunity and choose a suitable interaction pattern to switch back and forth between various risk states. In fact, drivers not only change from a high-risk state with high psychological pressure to a low-risk state with low psychological pressure, but also vice versa. The former is mainly driven by the motivation to reduce crash risk and cognitive effort; whereas, the latter aims to improve driving efficiency; however, the drivers will make more cognitive efforts to combat increased risks, demonstrating that drivers always try to make a tradeoff between driving efficiency, crash risk, and cognitive efforts. This study has positive significance for a better understanding of driving behavior and the basic mechanisms of driver decision-making.