显隐性危险源场景下驾驶员避险能力评价研究

Study on Evaluation of Driver Aggression Ability under Scenes of Obvious and Recessive Danger Sources

为研究显隐性危险源场景下驾驶员避险能力的差异性,借鉴中国交通事故深度调查数据库中的事故场景,利用Prescan软件构建测试场景并招募驾驶员进行避险能力驾驶模拟实验.采集7种驾驶参数数据用于表征驾驶员的避险能力,利用主成分分析法对参数进行降维,并用Ward法系统聚类对降维后得到的主成分因子进行聚类,实现对显隐性危险源场景下驾驶员避险行为的特征提取和避险能力的评价建议.

In order to study the differences of driver avoidance abilities under the scenarios of explicit and hidden hazards,the accident scenarios from China in-depth accident study are utilized,test scenarios are constructed with the software Prescan,and drivers are recruited to conduct driving simulation experiments of avoidance abilities.Then,seven kinds of driving parameter data are collected to characterize the driver risk aversion ability.PCA method is used to reduce the dimension of the parameters,and Ward method is used to cluster the principal component factors obtained after dimensionality reduction.The results show that the feature extraction of driver risk aversion behavior and evaluation suggestions on their risk aversion ability under the scenes of explicit and hidden dangers are achieved.