【作者】 张娟；

【导师】 刘浩学；

【作者基本信息】 长安大学 ， 交通环境与安全技术， 2012， 博士

【摘要】 近年来，公路连续长坡路段的安全问题越来越突出，出现了很多触目惊心的事故多发段，表明传统以“汽车动力学”为基础的道路线形设计理论，在指导山区连续长坡路段的设计方面存在局限性。本文的研究从驾驶员角度出发，分析连续长坡路段线形指标对驾驶员的影响，试图从“以人为本”的角度，提出更安全的、适宜的线形参数指标。研究成果对传统连续长坡路段的线形设计理论进行改进，可对现行的标准、规范的参数指标的重新评价和修改提供参考性建议；对即将建设和现有正在运行的连续长坡路段运用研究成果进行检查，以发现事故隐患点，提高用路安全性。基于以上目的，选择以实车试验作为获取研究数据的手段。随机选择9名在年龄和驾驶经验上具有代表性的驾驶员；采用多导生理记录仪采集被试驾驶员的生理指标；用GPS来采集实车试验过程的实时速度。依据实测数据，运用统计分析方法，对试验数据和道路线形参数指标进行分析研究，建立相关的回归分析模型，以心率增长率、速度增长率分别作为表征驾驶员心理生理反应和车辆运行速度的指标，并建立其与线形参数指标间的定量关系。平面线形参数研究：对选取的典型公路连续长坡路段平曲线进行了7个位置区段划分，对每个位置驾驶员心率增长率与曲线半径的关系进行研究，依据研究结果提出了连续长坡路段独立非连续曲线的半径建议值；以圆曲线长与曲线长的比值作为指标，研究其与驾驶员心率增长率的关系，提出了建议的取值；对缓和曲线长和驾驶员的心率增长率进行研究提出了缓和曲线长的建议值；以平曲线半径与平曲线长的比值与驾驶员的心率增长率的关系进行研究并建立相关模型，指出了最不利于安全的比值；对平曲线半径和驾驶员心率增长率的关系进行研究，提出了不建议连接的相邻曲线半径的参考值。纵断面线形参数研究：对公路连续长坡路段坡差与驾驶员心率增率进行研究，提出了坡差建议值；将纵坡分为缓坡和陡坡，对缓坡段坡长与心率增率的研究，提出了坡长建议值，陡坡段坡长与心率增率表明下坡方向驾驶员心理负荷较大；依据坡度与驾驶员心率增率及速度的关系，发现下坡方向驾驶员对坡度的敏感性较弱，上坡方向传统设计及规范建议采用两陡坡中间需加缓坡的设计在连续长坡路段有存在安全隐患的可能；得到了速度增长率和坡长之间的变化规律；依据竖曲线半径和速度增率的关系，提出了竖曲线半径的建议值。对平纵组合的参数指标与驾驶员心率增率及速度增率的关系运用二次多项式逐步回归方法进行了分析，并建立了拟合精度较好的模型。论文在最后以研究结论为依据，结合事故调查数据，建立了BP神经网络训练模型，并对该模型的可靠性进行了测试，也证实了本论文研究成果的可靠性；同时，该模型也为连续长坡道路的前期安全性评价提供了一个有效的方法，使连续长坡路段在设计阶段就可以对其安全性进行评定，以修改完善设计结果。更多还原

【Abstract】 In recent years, continuous long slope roads appear a series of segments where accidentsare facile to happen, the tradition road alignment design theory on the base of vehicledynamics is exposing more and more drawbacks for guiding the alignment design of thecontinuous long slope segments.This study, from the driver’s view, attempts to research the influence of alignment ofcontinuous long slope segments on drivers in order to propose more reasonable and safetycontinuous long slope roads alignment. The research results are valuable to improve thetraditional design theory of continuous slope roads alignment, provide the referencesuggestion to modifying the current standard; the research results are also useful in inspectingthe continuous long slope road which is on the point of construction or has existed for running,finding out the black point, so safety measures can be taken in time.Under the guide of above ideas, actual vehicle tests are used as a mean to obtain researchdata. Nine drivers, age and driving experience are representative, are random selected as theresearch object of study; Acquisition of the drivers’ physiological indexes is by thephysiological recorder; GPS is used to get the real-time speed of actual vehicle tests, andmark some special events. Based on the collected data, using the corresponding statisticalanalysis method, relevant regression analysis models are built. Heart rate increase rate andspeed increase rate are first used as the characterization factors for drivers’ physiologyreaction and vehicle speed.The results of horizontal alignment research: Every plane curve is divided into sevensections, research the relationship between driver heart rate increase rate of each curvesection and the radius, put forward suggested value of curve radius based on the researchresults for long slope sections; The ratio of circular curve length with the curve length asindexes, study the regulation of drivers’ heart rate increase rate,and put forward the suggestedratio; Study the regulation between drivers’ heart rate increase rate and easement curve, putforward the suggested value of easement curve; Study the relationship between driver’s heart rate increase rate and the ratio of radius with plane curve length, establish related model, andpoints out the most unfavorable ratio.The results of longitudinal alignment research: Research the relationship between thedrivers’ heart rate increase rate and the difference of slop, puts forward the suggesteddifference of gradient which is unsafely; Study the relationship among slope, drivers’ heartrate increase rate and speed increase rate, the research result shows drivers’ sensitivity toslope is weak for downhill direction; For uphill direction, the traditional design theory,whichsuggests that between two steep slope should have a gentle slope may not be unsafe; Theresearch also get the relationship between speed increase rate and the slope length,therelationship between vertical curve radius and speed increase rate.By applying the polynomial regression method, models think about the horizontalalignment, longitudinal alignment, drivers’ heart rate increase rate and speed increase rate areestablished.Based on the above research conclusion, combined with the accident investigation data, atraining model of BP neural network is built, and reliability the model is tested, whichconfirm the above research results are reliable. In addition, the model also provides a reliablemethod in road safety evaluation of long slope sections in designing phase, modifying thedesign results in time.更多还原