融合改进D-S理论与贝叶斯网络的机场旅客风险研究  

Research on airport passenger risk by fusing improved D-S theory and Bayesian network

作　　者：赵振武[1] 贾朋霖 ZHAO Zhenwu;JIA Penglin(School of Safety Science and Engineering,Civil Aviation University of China,Tianjin 300300,China)

机构地区：[1]中国民航大学安全学院,天津300300

出　　处：《安全与环境学报》2023年第12期4425-4434,共10页Journal of Safety and Environment

基　　金：天津市教委科研计划项目(2021SK025)。

摘　　要：为提高机场旅客风险评估准确性,实现分类安检的高效运行,提出了基于融合改进D-S理论和贝叶斯网络的旅客风险评估方法。首先,在识别旅客历史背景、面部情绪、异常行为的影响因素的基础上,建立旅客风险评价指标体系;然后,基于专家评价及文献统计方法,建立融合改进的D-S理论和贝叶斯网络模型以评价旅客风险;最后,通过实际案例计算各指标权重及风险值,在风险推理的基础上得出旅客风险等级,根据敏感性分析得到影响旅客风险的核心因素。结果表明,案例中的旅客为高风险等级,并且安检人员在对旅客风险分类时应重点审查托运行李违禁品记录和年飞行次数,同时,提高旅客风险评估准确性还应结合旅客情绪及异常行为,综合评估旅客风险。To improve the accuracy of airport passenger risk assessment and realize the efficient operation of classified security checks,a passenger risk assessment method based on fusion and improved D-S theory and Bayesian network was proposed.Firstly,a passenger risk assessment index system was established by identifying the influencing factors of passengers' historical background,facial emotions,and abnormal behavior.Then,based on the methods of expert evaluation and literature statistics,the qualitative indicators were evaluated with expert knowledge,and the final evaluation results and corresponding weights were obtained according to the expert evaluation language of fuzzy evaluation value quantification.The risk values of quantitative indicators and their corresponding weights were determined according to the law of literature statistics,and the conditional probability table was obtained with the weights.The improved D-S theory and Bayesian network model were established to evaluate the passenger risk.Finally,the weight of each index and the risk probability was calculated by the actual case,and the passenger risk level was obtained based on the risk reasoning.The research shows that the risk level of the passenger is a high risk,and the risk probability of the passenger's facial emotion is the highest,followed by the passenger's behavior and historical background of the passenger.Among the influencing factors of passenger risk,the historical background of passengers has a more significant impact on the risk of passengers.Among them,the record of the contraband quantity of checked baggage is the core factor affecting the risk of passengers,followed by the number of annual flights of passengers.Similarly,the risk level of passengers can be accurately obtained by evaluating other aviation security incidents based on the method presented in this paper.The proposed fusion and improved D-S theory and Bayesian network model can effectively identify the passenger risk level,and provide decision-making reference for securit

关 键 词：安全社会工程 旅客风险 D-S理论 贝叶斯网络 分类安检 

分 类 号：X949[环境科学与工程—安全科学]