基于小波熵的驾驶疲劳脑电信号特征提取

Feature Extraction of Fatigued Driver's Electroencephalogram Signals Based on Wavelet Entropy

为寻找衡量驾驶疲劳的可靠的生理指标,通过实验采集了模拟驾驶前、模拟驾驶疲劳后和休息后的脑电信号,并对其进行预处理和特征提取。应用小波变换对脑电信号进行降噪;并求得降噪后的脑电信号H和R小波熵比值。结果表明,驾驶疲劳后脑电信号H和R值有增大的趋势;再经充分休息后,脑电信号的H和R值有下降趋势,两者均具有显著性差异(P<0.05),说明小波熵值是驾驶疲劳脑电特征提取的有效方法。

For finding a reliable physiological indicator representing the level of driving fatigue, the electro-parison show that H and R values increase with fatigue after simulated driving and reduce after rest both with significant difference （P〈0. 05）, indicating the effectiveness of wavelet entropy in EEG feature extraction for driving fatigue.