近年来，随着车内信息设备的广泛使用，由其产生的驾驶人分神行为严重危害道路交通安全，已成为交通事故的主要致因之一。为探究驾驶分神对交通事故形成的影响及其机理，本文分别对正常驾驶状态下和冲突状态下的驾驶分神行为进行了系统性实验研究，阐明了正常行驶环境下驾驶分神对车辆横、纵向操控表现的影响及其内在机理，从而揭示分神导致交通冲突形成的机理；进一步，阐明了交通冲突状态下，分神对驾驶人响应及事故概率的影响机理。基于驾驶模拟器构建的城市道路跟车场景，对分神下的车辆横、纵向操控表现开展了系统性实验研究。采用统计检验手段，考虑了不同类型、程度的驾驶分神的影响，阐明了涉及视觉的分神对车辆横、纵向操控表现的不利影响，从而揭示其导致交通冲突形成的机理。同时，对比了4种不同车道偏离时距计算方法，提取出车辆横向安全性指标，发现了认知分神同时降低车辆横向位置标准差和车道偏离时距。此外，研究发现在认知视觉复合分神中，视觉分神占据主导地位。从驾驶人视觉、生理层面，对分神下的车道保持内在机理进行了建模研究。通过对分神下眼动、皮电数据的分析，阐明了视觉分神影响车辆操控表现的视觉机理，揭示了认知分神下视点集中度、意识觉醒、转向操控及车道保持的时变规律。同时，对比已有假说，建立了预测转向操控和车道保持变化的多水平线性回归模型，通过模型比较方法，得到了用于解释车道保持变化的最佳模型，揭示了认知分神影响车辆操控表现的内在生理机理。利用实验及建模方法，对典型交通冲突下分神驾驶人的响应及事故风险进行了系统性研究。分别在横向行人和纵向追尾冲突形态下，考虑了分神类型、冲突紧迫度（包括不同行驶车速、车头时距、前车减速度）、前车制动灯状态等因素，建立了驾驶人制动反应时间和事故概率模型，揭示了驾驶分神对驾驶人反应时间及事故风险的影响机理。同时，基于上述模型，提出了分神下的理论车头时距补偿量。在工程上为涉及分神的驾驶风险监测及车辆主动安全系统中驾驶人反应时间的设定提供了依据。
In recent years, with the extensive use of In-Vehicle Information System (IVIS), the induced driver distraction impairs road traffic safety severely, which has become one of the main contributors to traffic accident. To explore the influence of driver distraction on traffic accident and the underlying mechanism, the present study has conducted systematic experiments for investigating distracted driving on both normal driving condition and traffic conflict condition. The effects of driver distraction on both lateral and longitudinal driving performances, and their underling mechanisms on normal driving condition, are illustrated, thus revealing the mechanism of distraction causing the formulation of traffic conflict. Furthermore, the influences of distraction on drivers’ response and crash probability under traffic conflicts conditions are illustrated.Based on a simulated city-road car following scenario, both driver’s lateral and longitudinal performance are investigated experimentally. Statistical test method is used, so that different types and degrees of distraction can be considered. The adverse effects of visual-related distractions on driver’s lateral and longitudinal performance are illustrated, thus revealing the mechanism of visual-related distraction leading to the formulation of traffic conflict. In addition, different methods for time-to-line crossing calculation are compared, thus vehicle’s lateral safety indicator is extracted. Based on it, we observe reduction in both time-to-line crossing and standard deviation of lateral position during cognitive distraction. Besides, the conclusion that visual distraction dominates cognitive-visual combined distraction is verified in our results.The mechanism underlying distracted lane keeping is investigated by mathematical model, from the aspect of driver’s visual and physiological behavior. The visual mechanism of visual distraction affecting driving performance is illustrated, while the temporal dynamics of gaze concentration, arousal, micro-steering activity, and lane keeping performance is revealed, through analyzing eye movement and skin conductance data. Meanwhile, micro-steering activity models and lane keeping performance models are constructed, based on the existing hypotheses. The best model for explaining changes in lane keeping performance is determined, by model comparison method, thus revealing the underlying physiological mechanism of cognitive distraction affecting driving performance.Distracted driver’s response and crash risk, under typical traffic confits, are investigated systematically, by using experimental and modeling method. Driver’s brake reaction time and crash risk models are constructed, with the consideration of types of distraction, situational urgency (including different driving speed, following time headway, and lead vehicle deceleration), and lead vehicle brake light states, in lateral pedestrian crossing and longitudinal rear-end conflicts. As a result, we reveal the mechanism of distraction affecting driver’s reaction time and crash risk. In addition, based on these models, we propose the theoretical compensation of time headway in distracted driving. These will provide bases for the monitor of distraction-related driving risk and the design of driver’s reaction time in advanced driver assistant systems.
