【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， 2009， 硕士

【摘要】 据统计我国城市道路交通事故的80%是人为因素造成的,城市道路不良汇入与变道驾驶行为已成为城市交通安全的突出不良因素。如何选取不良汇入与变道驾驶行为特性指标,构建定量模型对该种驾驶行为进行界定,并基于充分的统计数据分析其特性,正逐步引起交通工程领域的重视。国内外研究表明,运用数理统计方法研究不良驾驶行为特性是预防和管理不良汇入与变道驾驶行为的关键。通过分析城市道路不良驾驶行为驾驶员的特征,构建不良驾驶行为严重度指标体系,引入熵权评价模型,从不良驾驶行为风险性、普遍性和可控性角度筛选出严重度显著的不良汇入和不良变道驾驶行为。在借鉴国内外研究成果的基础上,运用泊松分布和排队论推导车辆汇入行为的交通流参数表达式;基于视频实测数据,从累计频率分布角度标定不良汇入驾驶行为的界定指标,并构建汇入车辆驾驶员与主路驾驶员在汇入点处的混合战略博弈模型;依据不良汇入驾驶行为进入汇入点前的速度变化情况,把该种驾驶行为划分为直接型、缓冲型和等待型,并选取车辆、道路和交通流等指标参数,对城市道路不良汇入驾驶行为特性进行统计分析。基于大量的实测数据,从驾驶员、车辆组成和道路环境角度,分析了城市道路不良变道驾驶行为的影响因素;基于视频数据,考虑原始车道和目标车道跟驰车辆间的车头时距,从累计频率分布角度标定不良变道驾驶行为的界定指标;依据不良变道驾驶行为对交通运行安全的影响程度,把该种驾驶行为分为斜插型、挤压型和并行型,并选取车辆、道路和交通流等指标参数,对不良变道驾驶行为特性进行统计分析。综合比较不良汇入和不良变道驾驶行为特性的异同性,从驾驶行为判定、驾驶员教育、违章处罚和工程设计等角度,提出相应的不良汇入与不良变道驾驶行为联合管理和分离管理对策。鉴于本文研究的不完善之处,完善相应不良驾驶行为的界定理论和方法,定量化描述不良汇入与不良变道驾驶行为特性,并提出新的管理手段与处罚措施等均为后续研究的主攻方向和工作重点。更多还原

【Abstract】 According to the statistics, 80% of China’s urban road traffic accidents are caused by artificial factors. Risky merging and lane-changing have become the prominently adverse factors to the urban traffic safety. It attaches great importance in the traffic engineering area to problems such as how to select the characteristic indicators of risky merging and lane-changing, build a quantitative model to define these kinds of behaviors, and further analyze its characteristics based on the statistics. Both domestic and international studies have shown that, to apply statistical methods in analyzing the characteristics of risky driving behaviors has become a key factor to the prevention and control of the risky merging and lane-changing behaviors.The risky-driving-behavior extent index system was proposed based on the analysis of the characteristics of the risky drivers. The entropy evaluation model was applied to evaluate the degree of risky behaviors from the perspective of risk, universality and controllability, and further select the remarkable risky behaviors which include risky merging and lane-changing.Based on the research results at home and abroad, queuing theory is applied to calculate the expression of traffic flow parameters of merging. According to the video statistics, the indicators of risky merging behaviors were defined based on the cumulative probability distribution, and also the game model between merging drivers and the main road drivers at the point of the mixed was built. According the distribution of speed when the risky drivers merge, risky merging behaviors were divided into three types which were moving, buffering and waiting. The risky merging on urban roads was analyzed statistically with vehicles, road and traffic flow etc. as parameters.According to the full statistics, the factors influencing the risky lane-changing behaviors were analyzed from three points of view, the drivers, vehicles and road environment. Based on the video statistics, the indicators of risky lane-changing were defined based on headway of source and target lane from the perspective of the cumulative probability distribution. According to the impact of risky lane-changing behaviors to the traffic security, risky lane-changing behaviors were classified into three types including angled, extrusion and parallel, and they were analyzed statistically from the aspects of vehicles, road and traffic flow etc.Through comparing the similarities and differences between risky merging and lane-changing, several countermeasures were proposed both associatedly and dissociatively based on the behaviors, driver education, violation penalties and engineering design. Finally based on the limitation of the thesis, the author made some suggestions for further study in this area, such as improving the demarcations of risky behavior theories and methods, quantitatively describing the characteristics of risky merging and lane-changing, and proposing new management and penalties, etc.更多还原