基于复杂网络的营运货车交通事故风险因素识别

Driving Risk Identification of Commercial Trucks Based on Complex Network Theor

为明晰营运货车各交通事故风险因素间的关联性,通过分析营运货车交通事故数据,从人、车、环、管四方面确定29个风险因素,并运用共现分析法计算邻接矩阵和Jaccard指数确定各风险因素间的关联性,据此基于复杂网络理论建立网络节点多属性决策综合评价网络模型,判断各风险因素的重要程度。结果表明:营运货车交通事故风险因素网络具有复杂特性、小世界特性及无标度特性。超速、超载、操作不当、制动系统不合格、路面湿滑、交叉路口、驾驶员及车辆监管疏漏是关键链接风险因素。整体上,驾驶员风险因素异常会直接导致交通事故发生;道路环境因素属于较深层次的间接因素;企业风险因素是本质因素,具有高驱动力。

In this study,29 risk factors were identified from four aspects,that is,human,machine,environment,and management,by analyzing truck traffic accident data to clarify the correlation between the driving risk factors of commercial trucks.The co-occurrence matrix and Jaccard index were calculated to determine the correlation between the risk factors,and a complex network of driving risk of commercial trucks was constructed.Based on the network statistical indicators,a comprehensive evaluation model of network node multi-attribute decision-making was established to determine the degree of influence of various risk factors on the commercial truck-driving system.The results show that the risk network of commercial trucks exhibits complex,small-world,and scale-free characteristics.Speeding,overloading,improper operation,inadequate braking system,slippery road surface,intersections,driver,and vehicle supervision omissions are key link risk factors.In addition,road and environmental factors are significant indirect factors,and corporate risk factors with high driving force are essential factors.