基于压缩感知的疲劳驾驶脑电信号监测方法  

An EEG monitoring method based on compressed sensing for fatigue driving

作　　者：辛增念[1] 刘艳杰 Xin Zengnian;Liu Yanjie

机构地区：[1]江西科技学院协同创新中心,南昌330098 [2]江西科技学院理学教学部,南昌330098

出　　处：《科技创新与应用》2024年第36期47-50,共4页Technology Innovation and Application

基　　金：江西省教育厅科学技术重点研究项目(GJJ212003);江西科技学院协同创新中心开放基金重点项目(XTCX2104)。

摘　　要：脑电信号可用来有效地判断驾驶员是否疲劳驾驶,为减少驾驶人员驾驶过程中脑电信号的采集量,在信号采样端采用离散余弦基对驾驶员的脑电信号进行稀疏化,然后通过伯努利矩阵把稀疏的高维信号压缩采样成低维信号,最后在车上电脑端利用基追踪降噪法把压缩采样后的低维信号进行重构,还原出原脑电信号。在实验室进行模拟驾驶及脑电信号压缩采样的实验,结果表明,在压缩率小于80%时,重构后的脑电信号误差小于0.26,方法能保证疲劳监测系统所需的精确的脑电信号。Electroencephalogram(EEG)signals can be used to effectively determine whether the driver is tired or not.In order to reduce the amount of brain electrical signals collected by the driver during driving,the driver's brain electrical signals are sparse using discrete cosine bases at the signal sampling end,and then the sparse high-dimensional signals are compressed and sampled into low-dimensional signals through Bernoulli matrix.Finally,the compressed and sampled low-dimensional signals are reconstructed on the computer side of the vehicle using the base tracking noise reduction method to restore the original EEG signals.Experiments on simulated driving and compression sampling of EEG signals were conducted in the laboratory.The results showed that when the compression ratio was less than 80%,the error of the reconstructed EEG signal was less than 0.26.The method can ensure the accuracy required by the fatigue monitoring system.

关 键 词：脑电信号 疲劳驾驶 压缩采样 压缩感知 监测方法 

分 类 号：U471.15[机械工程—车辆工程]