摘要
因脑电信号更能直接反映大脑皮层疲劳状况,论文提出了一种基于共空间模式的脑电信号疲劳检测方法。该方法首先对数据集进行滤波等预处理操作,然后应用共空间模式提取特征,最后用支持向量机对提取到的有效空间特征二分类。此外,实验还采用了5折和10折交叉验证法进行评估;探索了脑电疲劳特征阶数相关系数m的取值;划分了脑区并对各区域疲劳识别准确率进行了比较。研究结果表明:论文方法的识别率高于基于样本熵、模糊熵等方法的识别率,疲劳检测准确率均值可达98.54%,全头皮疲劳识别率最高,额区疲劳识别率优于其他区域,可达92.54%。论文研究可为疲劳检测设备的研发提供更简单准确的检测方法,有助于促进可穿戴脑机接口在疲劳驾驶预警中的应用。
Because EEG can directly reflect the fatigue state of cerebral cortex,this paper proposes a fatigue detection method based on common spatial pattern.Firstly,the data set is preprocessed by filtering,and then the common spatial pattern is used to ex-tract features.Finally,the effective spatial features are classified by support vector machine.In addition,the experiment also uses 5 fold and 10 fold cross validation method to evaluate the method.It explores the value of correlation coefficient m of EEG fatigue char-acteristic order,divides the brain regions and compares the accuracy of fatigue recognition in each region.The results show that,the recognition rate of this method is higher than that of the methods based on sample entropy and fuzzy entropy,the average fatigue de-tection accuracy rate can reach 98.54%,the whole scalp fatigue recognition rate is the highest,and the frontal fatigue recognition rate is better than other regions,up to 92.54%.This study can provide a more simple and accurate detection method for the develop-ment of fatigue detection equipment,and help to promote the application of wearable brain computer interface in fatigue driving ear-ly warning.
作者
刘燕
郑威
龙佳伟
LIU Yan;ZHENG Wei;LONG Jiawei(College of Electronic Information,Jiangsu University of Science and Technology,Zhenjiang 212000)
出处
《计算机与数字工程》
2024年第1期195-200,共6页
Computer & Digital Engineering
关键词
脑电信号
疲劳检测
共空间模式
支持向量机
交叉验证
模糊熵
EEG signal
fatigue detection
common spatial pattern
support vector machine
cross validation
fuzzy entropy
