首页 | 本学科首页   官方微博 | 高级检索  
     

基于熵技术-全决策树的驾驶员人因可靠性研究
引用本文:林露青,黄坤明,廖素娟,沈斐敏. 基于熵技术-全决策树的驾驶员人因可靠性研究[J]. 交通标准化, 2014, 0(19): 75-80
作者姓名:林露青  黄坤明  廖素娟  沈斐敏
作者单位:福州大学环境与资源学院,福建 福州,350108
摘    要:
人因失误是道路交通事故发生的主要原因。全决策树方法可以确定不同情境中各影响因子对人因可靠性的影响程度和不同场景中的人因失误概率,其强调情境影响的动态分析模式符合驾驶员可靠性分析的需求。结合某道路运输企业具体情况,采用熵权法确定主要影响因子,基于全决策树法分析主要影响因子对驾驶员人误概率的影响程度,确定该情境下驾驶员的人误概率。计算结果显示,根据各主要影响因子的不同品质描述等级组合,该道路运输企业人因可靠性影响最明显的因子是疲劳,其次为外部环境、监督管理及驾驶员的性格,鉴于此结果提出以“筛选-矫正-监控”为控制手段的“三道防线”全程化人因管理模型。

关 键 词:驾驶员  全决策树(HDT)  熵权  影响因子

Study on Human Reliability of Drivers Based on Entropy Weight and HDT
LIN Lu-qing,HUANG Kun-ming,LIAO Su-juan,SHEN Fei-min. Study on Human Reliability of Drivers Based on Entropy Weight and HDT[J]. Communications Standardization, 2014, 0(19): 75-80
Authors:LIN Lu-qing  HUANG Kun-ming  LIAO Su-juan  SHEN Fei-min
Affiliation:(College of Environment and Resources, Fuzhou University, Fuzhou 350108, China)
Abstract:
Human error is the main cause of road traffic accidents. By using HDT method, the relation-ship between significant influence factors and human error probability is analyzed and human error proba-bility in different scenarios is estimated. This dynamic reliability analysis model which emphasizes on sit-uational analysis meets the needs of driver reliability. Significant influence factors are identified for driv-ers by using entropy weight analysis, and the relationship between significant influence factors and human error probability is analyzed by using HDT method.Results show that"fatigue"is the most significant in-fluence factor in this enterprise based on the integration of different influence factor quality in different scenarios, and external environment, supervision and drivers′character are also important factors.Based on the results, the "Three Lines of Defense" model of the entire management system for human factors with the control means of"Screening-Correction-Monitoring"is suggested.
Keywords:driver  Holistic Decision Tree(HDT)  entropy weight  influence factors
本文献已被 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号