A non-contact portable driving fatigue detection technique based on driver's physiological signals is proposed in this paper. Firstly the physiological signals of the biceps femoris of drivers in driving process are col-lected by sensors, electromyogram (EMG) and electrocardiogram (ECG) are separated from the original signals through fast independent principal components analysis and are denoised by empirical modal decomposition. Then on this basis,three characteristic parameters,i.e. the complexity of EMG,the complexity of ECG and the sample en-tropy of ECG are extracted,and the combination of these three characteristic parameters can apparently distinguish the normal and fatigue states of driver. Finally principal component analysis is adopted to reduce the dimension of characteristic parameters and obtain two principal components being able to effectively characterize the fatigue state of driver,with which as independent variables,a mathematical model for judging the fatigue state of driver is estab-lished. It is verified that the model can accurately distinguish the normal and fatigue states of driver with an accuracy rate of above 90%.%本文中基于驾驶员的生理信号提出一种非接触便携式的驾驶疲劳检测技术.首先通过传感器采集到汽车行驶过程中驾驶员的股二头肌的生理信号,经快速独立成分分析分离出肌电信号和心电信号,并采用经验模态分解进行去噪.接着在此基础上,提取出肌电信号复杂度、心电信号复杂度和心电信号样本熵3个特征参数.综合这3个特征参数能明显区分驾驶员的正常和疲劳两种状态.最后采用主成分分析法将特征参数进行降维,获得了2个能有效表征疲劳状态的主成分,以此为自变量建立了判定驾驶疲劳的数学模型.经验证,该模型能较准确地判别驾驶员在驾驶过程中的正常和疲劳状态,准确率达90%以上.
展开▼
