道路交通安全风险辨识与分析方法综述

A Review of Identification and Analysis Methods for Road Safety Risk

道路交通安全风险辨识及分析的准确性、全面性,是实现风险主动防控的基础和关键环节,直接影响道路交通安全管理的精细化水平。从影响因素和分析方法2个方面对道路交通安全风险相关研究进行综述和评论。针对人的不安全行为、车辆的不安全状态、道路的不安全条件、外界环境刺激等单因素风险,以及多因素间的关联耦合风险辨识,梳理了安全风险理论分析法、系统安全分析法、大数据与人工智能分析方法等道路交通安全风险分析方法。研究表明:安全风险理论分析法、系统安全分析法等以定性分析为主的方法侧重于对道路交通安全风险要素的全面、系统梳理,具有简单、直观、易操作等优势,但在多因素交织影响下的道路交通事故定量化剖析和事故成因深度挖掘方面存在较多局限性;基于多源数据挖掘技术的大数据与人工智能分析方法在海量信息感知、高效计算处理等方面优势明显,可基于多元数据对交通安全风险进行综合分析、精准挖掘,刻画多因素耦合下的事故风险特征、探究事故发生规律,是当前较为主流的研究方向。并提出道路交通安全风险研究领域存在的不足之处及未来研究发展方向,主要包括多源异构数据的动态采集与融合、智能网联环境下的道路交通安全风险辨识、考虑时空异质性的可移植的道路交通安全风险识别模型研究等。

The accuracy and comprehensiveness of road traffic safety risk identification and analysis is the basis and key link to achieve active risk prevention and control,and directly affects the refinement level of road traffic safety management.This paper summarizes and comments on the studies related to road traffic safety risk from two aspects of influencing factors and analysis methods.In a view of the single factor risk such as unsafe behavior of drivers,unsafe state of vehicles,unsafe conditions of roads,and external environmental stimulation,as well as the correlation and coupling risk identification among multiple factors,the road traffic safety risk analysis methods such as safety risk theoretical analysis method,system safety analysis method,big data and artificial intelligence analysis method are sorted out.The study shows that the qualitative analysis methods such as the safety risk theoretical analysis method and the system safety analysis method focus on the comprehensive and systematic analysis of the road traffic safety risk factors,and have the advantages of simplicity,directness,and ease of operation,but there are many limitations in the quantitative analysis of road traffic accidents and the deep excavation of accident causes under the influence of multiple factors.Big data and artificial intelligence analysis methods based on multi-source data mining technology have obvious advantages in massive information perception,efficient computing,and processing,and can comprehensively analyze and accurately mine traffic safety risks based on multiple data,depict accident risk characteristics under the coupling of multiple factors,and explore the rules of accident occurrence,which is the current mainstream research direction.It also points out the shortcomings in the field of road traffic safety risk research and the direction of future research and development,mainly including the dynamic collection and fusion of multi-source heterogeneous data,road traffic safety risk identification under the intelligent network environment,and the research of transplantable road traffic safety risk identification model considering space-time heterogeneity.