基于眼动仪的飞行员疲劳判定相关属性研究  

Research on Related Attribute of Pilot Fatigue Based on Eye Tracker

作　　者：樊玮[1] 鲁胜华[1] 

机构地区：[1]中国民航大学计算机科学与技术学院,天津300300

出　　处：《计算机技术与发展》2014年第6期15-18,23,共5页Computer Technology and Development

基　　金：国家自然科学基金资助项目(60979009);国家"973"重点基础研究发展计划项目(2010CB734105)

摘　　要：飞行员驾驶疲劳是影响飞行安全的重要因素之一。眼动仪已被尝试应用于驾驶人员的疲劳检测,但眼动仪采集数据属性较多,且缺乏明确的疲劳判定决策属性,故将PVT(精神运动警戒任务)测量所得疲劳程度组合进眼动仪测量数据中,作为决策属性,采用基于二元信道互信息的粗糙集属性约简方法,进行针对疲劳判定的眼动仪属性知识约简,并在约简前后,分别采用BP神经网络进行分类计算。结果表明,将二元信道互信息作为启发式信息,进行以疲劳判定为目标的眼动仪属性约简,能有效提取反映飞行员驾驶疲劳的主要属性。Pilot＇s driving fatigue is one of the important factors that affect flight safety.Eye tracker has been trying to apply to driver fatigue detection,but eye tracker collects too many data attributes,lacks of a clear decision attribute.It will make PVT（ Psychomotor Vigilance Task） into the data get by eye tracker,as decision attribute,use the rough set attribute reduction method based on binary channel of mutual information for attributes reduction of eye tracker in view of the fatigue test,and before and after the reduction,use BP neural network to classify respectively.Results show that use the binary channel mutual information as the heuristic information to test fatigue by attribute reduction,it can effectively extract the main pilot fatigue properties.

关 键 词：驾驶疲劳 粗糙集 属性约简 二元信道 精神运动警戒任务 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论]