驾驶人风险感知类型预测模型研究  

A predictive and forecast model of the risk perception type of the vehicle drivers

作　　者：秦雅琴[1] 梁锺月 贾现广[1] 郭凤香[1] 熊坚[1] QIN Ya-qin;LIANG Zhong-yue;JIA Xian-guang;GUO Feng-xiang;XIONG Jian(Faculty of Transportation Engineering,Kunming University of Science and Technology,Kunming 650500,China)

机构地区：[1]昆明理工大学交通工程学院

出　　处：《安全与环境学报》2019年第4期1266-1273,共8页Journal of Safety and Environment

基　　金：国家自然科学基金项目(71861016)

摘　　要：研究风险驾驶情境中驾驶人感知与行为的关系,应用KMRTDS驾驶模拟系统平台,试验得到56名被试在风险驾驶情境中的驾驶行为数据。利用风险驾驶情境风险度量化评价方法计算被试的风险感知效用值,用聚类分析法将被试分为3种风险感知类型。通过相关性分析得到与风险感知效用显著相关的13个驾驶行为表征参数,并提取出可有效表征驾驶行为特性的主因子:侧向控制因子、纵向控制因子、纵向姿态因子和稳定性因子。基于多项Logistic回归方法构建以客观情境风险度和驾驶行为特性主因子为输入的驾驶人风险感知类型预测模型,发现纵向控制因子、纵向姿态因子和稳定性因子更能反映驾驶人的风险感知类型。This paper is aimed to trace and describe the relationship between the drivers'driving behavior and their risk perception under different contexts of the likely to be faced with.In the paper,we have laid out 6 typical driving risk scenarios(3 types of risk driving scenarios with the motor vehicles,non motorized vehicles and pedestrians as the objects of conflict)and try to conduct our simulated experiments by using the KMRTDS driving simulation system as a platform.And,then,the average results of the subjective risk-making have been taken as the degree of the expert drivers while facing the objective risk scenarios,so as to take the average evaluation index as the objective situation risk degree of the situations.In the experiment,we have invited 56 drivers to join us as subjects.From the experiment,we have also measured the data related to the driving behavior of the subjects,including their operation data and the data of the vehicle's actual running state.In the experiment,the subjects have also been asked to fill in a subjective risk assessment questionnaire,which can help us to quantify their risk perception degrees and differences so as for us to assess their actual risk perception liabilities.Based on the testing results of the risk perception ability tests and the traffic conflict analysis theory,we have managed to divide the testing subjects into 3 risk perception types,that is,the cautious type,the radical type and the compound one.Through the correlation analysis,the 13 driving behavior parameters which are most significantly related to the relevance of the risk perception utility have been contributed to the 4 main factors that can effectively influence and decide the driving behaviors,that is,the lateral control factor,the longitudinal control factor,the longitudinal posture factor and the stability control one.And,furthermore,we have built up a prediction model of the driver's risk perception types based on the objective situational risk degrees and the driving behavior features as the input factors.

关 键 词：安全管理工程 风险感知效用 多项Logistics回归模型 驾驶模拟 驾驶行为 交通安全 

分 类 号：X951[环境科学与工程—安全科学]