基于风险驾驶情境的驾驶适宜性机理及识别方法研究

The causes of driving adaptation is very complicated, varied from individuals, sites and time, and influenced by the driver's individual characteristics and traffic environment greatly. Most of current researches of the driving adpatation focused on driving ability under static conditions, ignoring the researches of the driving adaptation abbility under dynamic driving situations, especially the risk driving situations. This project will choose a new angle of drivers' risk perception trait of risk driving situations to study. By applying the psychometric paradigm and simulation technologies, the drivers' risk perception psychology experiment will be constructed. By the collection of subject's experimental data, the drivers' risk perception properties will be analyzed and identified in quantitative and qualitative analysis methods. By using the risk perception theory and the behavioral decision theory, the forming and affecting mechanism of driving adaptation will be researched, and the behavioral prodiction model, the behavioral performance model and the conceptal model of multi-dimensional structure of driving adaptation will be constructed. Based on the extraction of hidden factors of driving adaptation, the classification standards of the comprehensive dynamic/static driving adaptation level will be proposed, and then an effective identifying method of driving adaptation will be produced with the establishment of its comprehensive discrimination model. The project will improve the comprehensive discrimination capacity of driving adaptation, provide the basis for drivers' safety management, and promote the development of the theory of driving adaptation.

驾驶适宜性成因十分复杂，因人因时因地而异，与驾驶人个体特性及道路交通环境相关极大。已有的驾驶适宜性研究多着眼于静态条件下的驾驶能力，忽略了动态驾驶情境，尤其是风险驾驶情境下驾驶适宜性能力的研究。本项目以风险驾驶情境下驾驶人的风险感知特性为切入点，应用心理测量范式及仿真技术，构建驾驶人风险感知特性心理实验；通过被试的实验数据采集，定性、定量分析驾驶人风险感知特性，实现风险感知特性的辨识；应用风险感知理论与行为决策理论研究驾驶适宜性形成及作用机理，建立行为预测模型、行为绩效模型、驾驶适宜性多维度结构的概念模型等；在析取驾驶适宜性隐性因子的基础上，提出动、静综合驾驶适宜性水平的分级标准；建立驾驶适宜性综合判别模型，形成一种有效的驾驶适宜性识别方法。项目预期成果可提高驾驶适宜性的综合评判能力，为驾驶员安全管理提供依据，促进驾驶适宜性理论的纵深发展。

驾驶适宜性检测不仅需要评价驾驶人是否具备正常驾驶工作的能力，更需要评价驾驶风险条件下驾驶人是否具备良好的风险感知和行为决策能力，从人因的角度保障道路交通安全。本项目对风险驾驶情境下的驾驶适宜性影响机理及识别方法进行了深入的研究和探讨，按计划完成了研究内容，并取得了预期的成果。首先，采用专家咨询法及文献综合法，编制了基于风险驾驶情境的驾驶人风险意识主观评价问卷，对驾驶人风险意识的心理状态进行研究，构建了以风险感知、风险态度、风险容忍为自变量的驾驶行为回归模型。其次，采用驾驶适宜性测试系统对深度知觉、暗视力、动视力等六项指标进行测试，并结合主观心理问卷评价和专家评价，定性、定量分析了被试的静态驾驶适宜性水平。第三，应用驾驶模拟系统平台（KMRTDS）构建了典型的交通风险驾驶模拟情境，进行风险感知测试实验及其研究。首次提出了“风险感知效用”的概念及其量化，分析不同因素对驾驶人风险感知效用的差异性影响。采用K均值聚类方法将驾驶人风险感知分为谨慎型、中间型、冒险型、复合型四种类型；采用Fisher判别方法，构建了驾驶人风险感知判别模型。第四，对不同风险感知效用类型驾驶人的视觉行为及特性进行研究，应用计划行为理论构建了驾驶人驾驶态度、主观规范、知觉行为控制、眼动特性等变量的风险感知效用类型模型框架；选择结构方程模型构建驾驶人眼动行为与风险感知效用类型的关系模型，用于解释驾驶人眼动行为、行为意向及风险感知效用类型之间关系。最后，运用聚类分析并构建BP人工神经网络模型，对驾驶人的风险感知类型进行识别分类，得到的结论一致并可以相互验证；采用主成分分析，建立了动、静综合驾驶适宜性条件下驾驶人风险感知特性的评价模型，为综合驾驶适宜性检测提供依据。