基于自然驾驶研究的直行追尾危险场景诱导因素分析

Analysis of Causation of RearEnd Incidents Based on Naturalistic Driving Study

摘要:

基于中国自然驾驶工况数据，筛选了直行追尾危险工况，通过修正的DREAM（driving reliability and error analysis method）方法进行了诱导因素分析.对直行追尾危险场景依据车辆行驶特点进行了场景细分，分析了不同场景细分类型下诱导因素逻辑图.结果表明,在直行追尾场景中驾驶员的“力度不足”和“距离过短”为占比最高的紧急事件（危险特征），并分析了追尾场景中4个细分类型对应的紧急事件特点.研究发现，中国驾驶员的驾驶习惯与欧美有较大差别，直行追尾危险场景中最主要的深层诱导因素为驾驶员习惯性期待他车特定驾驶行为等不良驾驶习惯，而不是在驾驶过程中与驾驶操纵无关的第二行为.

关键词:自然驾驶工况; 追尾危险场景; 驾驶可靠性及差错分析方法; 诱导因素分析

Abstract:

Based on the rear-end incidents extracted from the China naturalistic driving data, the DREAM (driving reliability and error analysis method) was modified and used to identify contribution factors and causation patterns in these incidents. Those rearend incidents were respectively classified in different driving scenarios, and the logic diagrams of the contribution factors with different detailed categories were obtained. The results showed that the main factors behind the rearend incidents was the adoption of “insufficient force” and “small safety margins”. The characteristics of critical events of four detail categories in the rearend scenarios were also analyzed. In contrast to the results obtained from previous studies in the US and Europe, the deepseated contribution factors was driver’s habitual expectation for the specific driving behavior of other traffic users, instead of the second driving behavior.

Key words:naturalistic driving; rear-end critical scenarios; driving reliability and error analysis method(DREAM); causation analysis