摘要
疲劳驾驶时,司机的脑电信号和眼电信号特征均发生显著变化,本文针对这两类信号进行分析研究,利用这两类数据综合分析判断司机是否处于疲劳驾驶状态.首先对采集的脑电信号进行小波包分解,提取信号中的α波,并计算其相对功率谱P;然后利用Pearson相关系数分析两路对称导联F7,F8中眨眼信号特征,去除干扰;最后利用BP神经元网络对眨眼信号进行识别,计算眨眼频率.结果表明,利用眼电信号和脑电信号特征综合分析司机眨眼动作,能准确识别出眨眼信号,并能正确检测人的驾驶疲劳状态的变化.
The EEG and EOG have significant changes when drivers are fatigued. So these two types of signals can be used to analyze fatigue driving. Firstly, the a rhythm was extracted from drivers＇ EEG signals using wavelet packet decomposition and its relative power spectrum P was calculated. Then the blinking characteristics of the EOG signals contained in F7 and F8 channels were analyzed and the interference signals were removed using the Pearson correlation coefficient. Finally, the blinking signals were identified using BP neural network and the blinking rate was calculated. The results show that using the comprehensive analysis of EOG and EEG can accurately identify the blinking signals and correctly detect the changes of driver fatigue state.
出处
《东北大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2014年第2期175-178,共4页
Journal of Northeastern University(Natural Science)
基金
国家自然科学基金资助项目(61071057)
关键词
疲劳驾驶
脑电信号
眼电信号
小波包分解
相对功率谱
眨眼频率
fatigue driving
EEG （ electroencephalogram ）
EOG （ electroencephalogram ）
wavelet packet decomposition
relative power spectrum
blinking rate
