【作者】 咸化彩；

【导师】 金立生；

【作者基本信息】 吉林大学 ， 载运工具运用工程， 2014， 博士

【摘要】 道路交通事故已成为重大的社会和经济问题据统计，25%~50%的道路交通事故因驾驶人注意力分散引起，而在注意力分散引起的交通事故中，由驾驶人主动参与次任务行为引起事故的比例高达36.4%，成为导致道路交通事故的重要原因但由于驾驶行为分析技术手段和理论工具的缺乏，目前国内外针对次任务驾驶安全性的研究尚不成熟，内容也较为分散为此，本文将以次任务驾驶安全性为研究对象，综合考虑次任务类型影响因素，融合驾驶人视觉行为和车辆运行状态信息，以准确评价次任务驾驶安全性为目标，以寻求不同驾驶过程中各参数的变化规律为研究主线，对次任务驾驶安全性评价指标及评价模型进行深入研究研究成果将为驾驶安全研究提供科学的理论依据和技术支持，并具有重要的理论意义和实际应用价值具体研究内容如下：1.综述国内外次任务驾驶安全研究进展从次任务类型次任务驾驶对不同表征参数的影响次任务驾驶模型和次任务驾驶行为研究方法四个方面对次任务驾驶安全研究现状进行详细介绍基于现状分析，发现目前国内外研究存在的问题，并针对评价参数系统性差视觉行为研究不完善及评价体系缺乏等不足开展进一步的研究工作2.驾驶行为信息采集方案设计与技术实现为准确分析次任务驾驶安全性，本文提出采用视觉行为和车辆运行状态参数相结合的次任务驾驶安全性评价指标体系，并基于驾驶行为影响因素及选择原则分析，对驾驶实验人员构成实验道路环境次任务类型及操作方式等进行选择设计通过集成Smarteye眼动仪驾驶模拟器等设备，搭建驾驶行为信息采集系统，同时研究视觉行为与车辆运行状态参数采集的时间一致性，解决数据流同步问题对数据提取方式异常数据剔除和数据分析方法进行设计，为后续研究奠定基础3.驾驶人视觉行为分析系统分析视觉行为表现形式，结合眼动仪采样频率确定注视行为与扫视行为的区分标准比较研究驾驶人注视区域划分方法，立足于本实验特点，提出一种基于驾驶人特性的注视区域划分方法利用方差分析法，分别从注视行为扫视行为及眨眼行为三个角度，重点解析不同驾驶状态及不同次任务类型下的驾驶人视觉行为变化规律，提取受次任务影响显著的参数，并确定出次任务驾驶视觉行为安全性评价指标4.车辆运行状态参数分析从车辆纵向和横向运行入手，对车辆运行状态参数进行细化分析，重点研究次任务驾驶对速度均值速度标准差加速度标准差方向盘转角及方向盘转角熵值的影响，并利用方差分析法对不同驾驶条件下的车辆运行状态参数差异显著性进行分析，进而确定出次任务驾驶车辆运行状态安全性评价指标5.次任务驾驶安全性评价模型构建方法研究针对次任务驾驶安全性评价目标评价对象及关系，对比指标权重确定方法，提出利用模糊网络分析法建立次任务驾驶安全性评价模型，并对模糊网络分析法基本原理评价体系构建权重求取过程等理论和方法进行梳理6.次任务驾驶安全性评价模型建立基于Pearson相关系数分析方法，分析次任务驾驶视觉安全和车辆运行状态安全评价指标间的相关性，构建次任务驾驶安全性评价指标体系对指标类型一致化指标无量纲化处理驾驶安全评语集建立及指标权重求取方法等进行深入研究，建立评价指标的单因素模糊判断矩阵，确定出评价指标权重向量，通过模型求解并结合最大隶属度原则判断不同驾驶过程中的驾驶安全性等级，最终实现次任务驾驶安全性评价模型构建7.次任务驾驶安全性评价模型适用性研究利用次任务驾驶安全性评价模型对次任务驾驶实验数据进行分析处理，从驾驶类型次任务类型和驾驶经验等方面分析次任务驾驶对驾驶安全性的影响，验证次任务驾驶安全性评价模型的适用性更多还原

【Abstract】 Traffic accident has become a significant social and economic problem. According tostatistics,25%~50%of the total traffic accidents were caused by driver distraction, andsecondary task driving is the primary cause leading to driver distraction. Generally, driversactively involved in distracting activities and accounts for as high as36.4%of crashes.However, owing to the lack of driving behavior analysis technology and theoretical tools,researches on secondary task driving safety are immature and the research achievements arerelatively scattered. Therefore, in this study, taking secondary task driving safety as researchobject, fusing the information of driver visual behavior and vehicle running status, targetingat accurately evaluating the safety of secondary task driving, finding the deviationcharacteristic of various parameters under different driving conditions as the main line, thesecondary task driving safety evaluation index system and evaluation model are deeplystudied. The results can provide theoretical and technical support for driving safety research,and have important theoretical significance and practical application value. The specificresearch contents are as follows:1. Review of existing research results. The dissertation proposes a research frame aboutsecondary task driving research, and includes4categories: secondary task type, effects ofsecondary task driving on different parameters, secondary task driving evaluation model, andsecondary task driving behavior. Based on the analysis of current situation, problemsexisting in these researches are founded. At last, considering the existent problems ofevaluation indexes lacking systematicness, poor visual behavior research, and lackingsecondary task driving safety evaluation model, further studies will be carried out in thisstudy.2. Design and implementation of information acquisition system for secondary taskdriving safety analysis. For accurate analyses the safety of secondary task driving, thesecondary task driving safety evaluation index system combining measures of driver visualbehavior and vehicle running status is proposed. In addition, experiment project is designed,including installing the collecting devices, developing the traffic scenes, standardizing theexperiment process, selecting the measures, selecting secondary task types, etc. Based on theequipments of Smarteye eye tracking system and driving simulator, information acquisitionsystem for secondary task driving safety analysis is developed. Subsequently, the technology of multi-threads synchronization is used to integrate multi-source information to achievetime synchronization. Finally, methods of data extract, removing abnormal data and dataanalysis are developed, which could provide data support for following research.3. Analysis of drivers visual behavior under different driving conditions. First, theforms of visual behavior are systematically analyzed and the distinguish criteria betweenfixation and saccade is designed. Second, after a thorough comparative analysis of methodfor driver gaze areas dividing, a new method based on driver character to divide gaze areas isdeveloped. Using analysis of variance, deviation characteristics of driver visual behaviorincluding fixation, saccade and blink are studied deeply. Parameters significantly affected bysecondary task are extracted and evaluation indexes for secondary task driving safetyevaluation are determined.4. Analysis of vehicle running status under different driving conditions. Analysis ofvehicle running status is conducted from two aspects, lateral and longitudinal running status.Parameters of average speed, standard deviation of speed, standard deviation of acceleration,steering wheel angle, and steering entropy are analyzed deeply. Based on essential featuresof the vehicle running status analysis, method of analysis of variance (ANOVA) is used toquantify the effects of secondary task on driving performance measures. At last, the indexsystem for evaluating vehicle running status is constructed.5. Method research for building secondary task driving model. On the basis of analysisof the objective for secondary task driving safety evaluation, evaluation objects, andrelationships between objects, an approach to construct secondary task driving safetyevaluation model is developed, that is Fuzzy Analytic Network Process (F-ANP).Subsequently, the theory and method of the F-ANP, evaluation index system constructing,and weight array of indexes obtaining are studied.6. Building secondary task driving safety model by using F-ANP. Based on the analysismethod of correlation coefficient, we analyze the indexes relationships between driver visualbehavior and vehicle running status, and construct the secondary task driving safetyevaluation index system. Furthermore, this study introduces a method that the evaluationindexs are selected and filtered, and establishes the driving safety remark set and constructsthe single factor assessment matrix for each driver. In addition, the global weight of eachindex is calculated and the driving safety levels are determined. Finally, due to the maximummembership grade principle, the driving safety level is determined and the secondary taskdriving safety evaluation model is completed.7. Study on the applicability of secondary task driving safety evaluation model. Data obtained from secondary task driving experiment are processed and analyzed by using thesecondary task driving safety evaluation model built in this study. Based on the evaluationresults, we classify and analyze driving safety according to driving condition, drivingexperience and the type of secondary tasks.更多还原