【作者】 漆巍巍；

【导师】 裴玉龙；

【作者基本信息】 哈尔滨工业大学 ， 交通运输规划与管理， 2014， 博士

【摘要】 道路基础设施供给与交通需求矛盾日益加剧，导致我国大中城市交通拥堵频发，交通拥堵对驾驶人身心健康和驾驶行为的影响可逐步转化为潜在的安全隐患。目前国内外学者主要从形成机理、疏解策略和环境污染等方面研究道路交通拥堵的危害；而关于交通拥堵对驾驶人身心健康和驾驶行为影响的研究成果则鲜见报道。因此，论文以拥堵状态下驾驶人的生理和行为特性为研究对象，揭示驾驶人的心电和换道指标随交通拥堵压力变化的规律，量化交通拥堵对驾驶人生理和行为的不良影响，提出拥堵状态下驾驶人换道的PSR(压力-状态-响应)模型，探讨基于PSR模型的风险换道警示技术，可以取得较好的创新性研究成果。从交通流客观运行的角度提出的交通拥堵状态划分方法已经非常成熟，但这种分类方法不能表征驾驶人对交通拥堵的主观感受，论文提出交通拥堵压力系数来表征城市道路的交通拥堵程度，并从驾驶人主观感受角度将城市道路交通流划分为畅通状态、轻度拥堵、中度拥堵和重度拥堵四个等级，基于实测数据标定各型交通流状态的压力系数分级阈值，为开展不同交通拥堵程度下驾驶人的视觉、生理和行为特性的测试研究提供了前提条件。针对交通畅通、轻度拥堵、中度拥堵和重度拥堵四种状态，设计实验测试方案，测定拥堵状态下驾驶人的视觉特征(注视、扫视、眨眼)、生理特征(心率、心率变异性)和换道构成比例数据。实验人员的选择兼顾了性别、年龄、驾龄等多个因素，实验路径包括快速路、主干路、次干路和支路等多种类型，仪器设备主要采用进口的IView XTM HED型头盔眼动仪和Spirit-10C型多通道生物生理记录仪，以确保数据的客观性和准确性。在实测实验的基础上研究交通畅通和交通拥堵状态下驾驶人的注视、扫视和眨眼特征的差异，采用非参数检验方法分析驾驶人各项视觉特性指标在畅通和拥堵状态下差异的显著性，研究结果表明交通拥堵状态下驾驶人的注视点分布范围更广，更集中于中央主视区(近距离)，注视持续时间变短，扫视速度和扫视加速度增大，眨眼持续时间和眨眼频率降低。这些变化的主要原因是交通拥堵状态下驾驶人周围的车辆、行人、实时信息板等交通元素的信息量增大，导致驾驶人单位时间内需要处理更多更复杂的交通环境信息，且驾驶人对自身行车安全性的忧虑增加。在实测数据基础上研究交通拥堵对驾驶人生理特性的影响机理，并采用心率指标和心率变异性指标来表征驾驶人的心电特性，对交通拥堵状态下驾驶人的心率均值、心率变异性时域和频域指标展开深入研究。驾驶人的心率均值随着交通拥堵程度的增加而增大，因此构建了心率均值和压力系数间的关系模型，该模型可以清晰的表达交通拥堵压力下驾驶人心率的波动规律。拥堵状态下的心率变异性分析结果表明，交通拥堵显著影响了驾驶人交感神经和迷走神经对心脏的调节功能，随着交通拥堵程度增加，交感神经兴奋加强，迷走神经兴奋降低，心率调节的平衡向交感神经倾斜。交通拥堵对驾驶人眼动和心率的影响最终会以具体的驾驶行为表现出来，车道变换行为发生变化是交通拥堵压力作用的直接结果，因此选择交通拥堵压力系数为自变量，斜插型、挤压型和并行型换道的构成比例为因变量，在充分的观测数据基础上，回归分析得到各类型换道比例数与交通拥堵压力系数间的关系模型，该模型客观的反映了交通拥堵程度对驾驶人换道行为的影响，可以为评价道路交通拥堵的危害性提供新的视角。在研究拥堵状态下驾驶人的心电特性和换道行为变化规律基础上，引入BP神经网络理论，将压力(交通拥堵压力系数)和状态(驾驶人心率均值)作为输入变量，响应(驾驶人换道行为构成比例)作为输出变量，构建交通拥堵状态下驾驶人换道的PSR模型，该模型揭示了交通拥堵压力系数和驾驶人心率状态对驾驶人换道构成比例的联合影响，并运用实测数据与模拟数据对照的方法检验了该模型的实用性。最后，基于驾驶人换道的PSR模型，提出了拥堵状态下斜插型换道的警示方法，以提前警醒驾驶人在道路拥堵时注意放松心情，尽量避免实施危险的斜插型换道行为。本文旨在揭示“道路交通拥堵→驾驶人焦虑感→驾驶人视觉特征→驾驶人生理特征→驾驶人行为特性→车道变换警示”之间的内在联系，丰富当前对交通拥堵状态下驾驶人生理和换道行为特性研究的理论体系，并提出一种及时预测高风险换道行为的方法和危险换道行为的警示思路，为今后提出缓解交通拥堵状态下驾驶人焦虑的策略奠定基础。更多还原

【Abstract】 The traffic congestion will be in a long memory in big cities of China, due tothe contradiction between traffic supply and traffic demand. A tremendous amountof studies have conducted to investigate the effects on traffic operation and pollutionof traffic congestion. However, very little is known about the influences of trafficcongestion on driver’s physiological and psychological health. The objectives of thisstudy are therefore to examine the change pattern of driver’s physiological andbehavioral characteristics under different traffic congested conditions, and toquantify the negative impacts of traffic congestion on driver’s physiology andbehavior via uncovering the formation mechanism. The PSR (pressure-state-response) model of driver’s lang changing is proposed to realize a warningtechnology of risky lane changing in the congested condition.The concept of pressure coefficient has been introduced to describe the trafficcongestion level. Based on driver personal subjective perception, traffic congestionwas categorized into four levels: unblock, light congestion, medium congestion, andsevere congestion. The threshold value of the pressure coefficient at each level wasestimated by using the field observation data. This provides a theoretical basis forcarrying out the experiment on driver’s visual, physiological and behavioralcharacteristics under different traffic congestion states.Advanced equipments were utilized to conduct an experiment on driver’s visual,physiological and behavioral characteristics under four traffic congestion scenarios.The indices of driver vision (i.e., gaze, glance, and wink), physiologicalcharacteristics (i.e., heart rate, heart rate variability) and risky lane changingcomposition ratio were selected to investigate the effects of traffic congestion ondriver’s physiology and behavior.According to experimental data, the differences of driver’s indicators wereexamined using non-parametric test method under different traffic congested states.Results show that the increasingly traffic congestion level results in driver’s gazepoint distribution becoming disperse, while more attention being focused on thecentral primary visual area (close-up). As the traffic congestion level increases,driver’s gaze duration shortens, saccade velocity and acceleration increases, andsaccade, blink duration and blink frequency decreases. This can be interpreted interms of information amount. If the driver is presented with traffic congestion, s/heshould to handle more traffic and more complex environmental information, such aspedestrians, adjacent vehicles and variable message signs.Heart rate and heart rate variability were adopted to analyze the effects of traffic congestion pressure on driver’s ECG characteristics. Time and frequencydomain of the two indexes were studied. Traffic congestion level and driver’s meanheart rate were taken into consideration to examine the relationship between meanheart rate and pressure coefficient. According to the analysis results, trafficcongestion significantly affects the regulatory function of driver’s nerve and vagusnerve. An increase in the level of traffic congestion associate with a reduction invagus nerve, sympathetic nerve strengthening, and the balance of heart rateadjustment changing.Congestion pressure coefficient was selected to be the independent variable.The composition ratio of driving behavior was chosen to be the dependent variable.A regression model was proposed to quantify the relationship of congestion pressurecoefficient and the composition ratio of driving behavior. Based on the study in theECG characteristics and lane changing behavior under congested conditions, a BPneural network model was developed. The pressure(congestion pressure coefficient)and status (driver’s mean heart rate) were regarded as the input variables, theresponse (the driver’s lane changing composition proportion) was chosen to be theoutput variable. Thus, the driver’s PSR model of lane changing is established underdfferent traffic congestion states to uncover the impacts of congestion pressurecoefficient and driver’s physiological status on the lane changing behaviorcomposition ratios. Field observation data were used to validate the accuracy andeffectiveness of the proposed model.The goal of this article is to investigate the inherent relationship among trafficcongestion, driver’s anxiety, driver’s visual characteristics, driver’s physiologicalcharacteristics, and driver’s behavioral characteristics. This aids understanding thevariation of driver physiological and behavioral characteristics under differenttraffic congested states. This work provides a solid basis for future research onconceiving strategies to alleviate the driver’s anxiety under traffic congestioncondition and improve traffic safety.更多还原