基于实车仿真的不同风险等级下新手驾驶人跟车行为研究

发布时间：2018-11-11 00:05

【摘要】：公安部统计数据显示,尾随相撞的发生率占所有事故形态的40.39%，死亡、受伤人数和财产损失也都在40%以上，所以追尾事故可以说是危害最大的事故形态，而变道和追尾事故的发生大多是驾驶人的跟车决策出现偏差造成的，新手驾驶人作为此类事故的易发群体，对其跟车行为的研究具有重要意义。本文在国内外已有研究的基础上，针对不同风险等级的场景，对新手驾驶人的跟车行为进行研究，，探索新手驾驶人在不同风险等级下的跟车行为绩效和注视特征。通过前期访谈和资料查阅先总结出了22种跟车过程中常见的场景，构成跟车场景库。设计调查问卷，考察驾驶人对这些跟车场景风险等级的评估，统计分析调查结果，将这22个场景按照风险等级的高低分成低风险、低中风险、中风险、中高风险和高风险场景。分别从低风险、中风险和高风险场景中选取两个典型的不同风险下的跟车场景作为实验场景。利用Creator，Vega，C++等工具将这些典型的跟车场景在仿真软件中设计实现，为后面的仿真实验打下软件基础。对实验用真车进行选型和改装，在轿车上安装传感器和数据采集卡设备，设计信号的处理方式和传输方式，并在安装过程中设计相关电路，为后面的仿真实验打下硬件基础。进行仿真实验，分析新手驾驶人在不同风险等级下驾驶绩效的特征和视觉搜索模式的特征。结果表明：驾驶绩效方面，在高风险场景下，新手驾驶人的事故率是经验驾驶人的5倍左右；在任何风险场景下，经验组能够更早地感知到危险从而提早采取减速措施，而新手组做出减速反应更晚；在坡道起步和慢速跟车时，新手驾驶人对速度的把握很差，经常出现溜车现象，险情点平均车速在-2km/h；经验驾驶人基本可以遵守限速标准，而有些新手驾驶人在会出现超速行驶的情况；在高风险场景下险情突然出现时，新手驾驶人时常会采取过激的行为。在注视特征方面，不管是新手驾驶人还是经验驾驶人，扫视时间远远大于注视时间，是注视时间的3倍以上，这说明驾驶人的视觉搜索都是以扫视运动为主，注视运动较少；经验驾驶人在遇到可能会出现潜在危险的环境时会减少注视时间，增加扫视时间，来对可能出现的危险做预警，而新手驾驶人的扫视时间和注视时间相对稳定，不管高中低风险都保持一定的注视时间；新手驾驶人无论在任何风险下，对仪表板特别是车速的关注都高于经验驾驶人；不管在任何风险等级的场景下，新手驾驶人的注视区域都比经验驾驶人的要集中，新手驾驶人的注视区域一般集中在某些区域，而对其他区域的关注较少，从而会忽视后视镜，交通标识等。本文的研究成果对于新手驾驶人的筛选与培训，提高新手驾驶人驾驶绩效和能力，设计和开发新手驾驶人驾驶培训平台具有重要意义。
[Abstract]:Statistics from the Ministry of Public Security show that the incidence of trailing collisions accounts for 40.39 of all accident patterns, and the number of deaths, injuries and property losses are also above 40%. Therefore, rear-end accidents can be said to be the most harmful form of accidents. However, the accidents of changing track and rear-end are mostly caused by the deviation of the decision of the driver. As the vulnerable group of such accidents, it is of great significance for the novice driver to study the behavior of following the car. Based on the existing research at home and abroad, this paper studies the behavior of the novice driver under different risk levels, and explores the behavior performance and gaze characteristics of the novice driver under different risk levels. First, 22 common scenes in the process of following car are summarized through the previous interviews and the reference of data, which constitute the scene library of the car following. A questionnaire was designed to investigate the drivers' assessment of the risk levels of these car-following scenarios. The survey results were statistically analyzed. The 22 scenarios were divided into low risk, middle risk and high risk scenario according to the risk level. Two typical vehicle following scenarios with different risks were selected from low risk, medium risk and high risk scenarios as experimental scenarios. These typical car following scenes are designed and implemented in the simulation software by using Creator,Vega,C and other tools, which will lay the foundation for the simulation experiment in the future. The model selection and modification of the real vehicle for the experiment, the installation of sensors and data acquisition card equipment on the car, the design of the signal processing and transmission mode, and the design of the related circuit during the installation process, which will lay the hardware foundation for the simulation experiment behind. A simulation experiment was carried out to analyze the characteristics of driving performance and visual search mode of novice drivers under different risk levels. The results show that the accident rate of novice drivers is about 5 times that of experienced drivers in high risk scenarios. In any risk scenario, the experience group was able to sense the danger earlier and take early deceleration measures, while the novice group responded later; In the ramp starting and slow following the car, novice drivers have a poor grasp of the speed, often slip through the car phenomenon, the average speed of the danger point in -2 km / h; The experienced driver can basically abide by the speed limit standard, while some new drivers will overspeed, and when the danger suddenly appears in the high risk scenario, the novice driver will often take excessive behavior. In the aspect of gazing characteristics, whether the novice driver or the experienced driver, the scan time is far longer than the fixation time, is more than three times of the fixation time, which shows that the driver's visual search is mainly based on the scanning motion, and the gaze motion is less; When experienced drivers encounter potentially dangerous environments, they tend to reduce their gaze time and increase their scan time to warn of possible hazards, while the rookie driver's scan time and fixation time are relatively stable. Maintain a fixed amount of time regardless of high school and low risk; Novice drivers pay more attention to dashboards, especially speed, regardless of risk than experienced drivers; In any risk level scenario, the rookie driver's gaze area is more concentrated than that of the experienced driver. The novice driver's gaze area is generally concentrated in some areas, while the other areas are less focused, thus neglecting the rearview mirror. Traffic signs, etc. The research results of this paper are of great significance to the selection and training of novice drivers, the improvement of the performance and ability of novice drivers, and the design and development of rookie driver training platform.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.255;U471.3

【参考文献】