摘要
精神疲劳影响驾驶员的警觉性和安全驾驶能力,引发的交通安全问题不容忽视。将脑电图识别与车辆操纵特性相结合来检测驾驶员的疲劳状态,预期为搭建疲劳驾驶检测系统提供理论及实验依据。设计了模拟驾驶实验,采集被试者的脑电图(EEG)信号和对应的方向盘操纵数据;针对疲劳程度三分类问题,利用小波包变换和共空间模式算法对EEG信号进行特征提取;依据车辆操纵特性评估驾驶员疲劳程度来确定EEG信号的分类标准;并选择支持向量机对EEG信号进行分类以完成对驾驶员精神疲劳状态的定性分析,分类准确率可达94.259%。
Mental fatigue affects driver' s alertness and safe driving ability, which is easy to cause traffic safety prob- lems. This paper intends to provide theoretical and experimental basis for building a driving fatigue detection system based on EEG recognition combining vehicle handling characteristics. Firstly, a driving simulation experiment was de- signed to collect the EEG signal and steering wheel handling data of the subject;then, aiming at the three-classifica- tion problem of fatigue degree, the features of EEG signal were extracted with wavelet packet transform and common spatial pattern methods. Moreover, the vehicle handling characteristics were used to estimate the degree of driving fa- tigue, and determine the criterion for EEG signal classification ; Finally, support vector machine (SVM) was employed to discriminate the EEG signal and realize the qualitative analysis of the mental fatigue of drivers. The classification accuracy of up to 94. 259% is achieved.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2014年第2期398-404,共7页
Chinese Journal of Scientific Instrument
基金
中央高校基础科研基金(N120204002)
广东省汽车工程重点实验室开放基金(GDAEL2012002)
辽宁省自然科学基金(2013020040)资助项目
关键词
驾驶疲劳
脑电图
操纵特性
共空间模式
小波包变换
支持向量机
driving fatigue
electroencephalogram (EEG)
handling characteristic
common spatial pattern
wavelet packet transform
support vector machine (SVM)