基于颈腰部肌电及脑电信号的疲劳驾驶检测

摘要 为了有效判别驾驶员的疲劳状态，结合生物力学分析提取了驾驶过程中的颈腰部肌电信号EMG和头部脑电信号EEG，并分析其特征参数在驾驶过程中的变化规律.结果表明:颈肌样本熵、颈肌复杂度、腰肌样本熵、腰肌复杂度、脑电样本熵、脑电复杂度这6个生理信号的特征参数值都随着驾驶时间的延长而逐渐降低，通过主成分分析可实现特征参数间的合理组合.基于多元回归理论，建立了能够有效预测疲劳驾驶的数学模型.状态验证结果表明，该模型对疲劳状态判别的正确率可达95%以上.
Abstract：In order to effectively test driver fatigue， the surface electromyography (EMG) and electroencephalogram (EEG) were collected in driving processes， and the characteristic parameters were extracted and analyzed combined with biomechanics. The experimental results indicated that the sample entropy (SampEn) and complexity of EMG and EEG gradually decrease with the driving time expends. These characteristic parameters can be reasonably combined by using the principal component analysis. Based on the multiple regression theory， the characteristic parameters at different positions of the body are reasonably combined， and a mathematical model to evaluate fatigued driving is built. The accuracy of the model is up to 95% by the state validation.