An active driver fatigue identification technique using multiple physiological features

A system that can actively monitor the driver's fatigue level in real time is urgently needed for the prevention of accidents. Support vector machine (SVM) technique is used to identify driver's fatigue based on psychological features, such as EEG and ECG. Driver's fatigue is expressed as alert, mild fatigue, deep fatigue and drowsiness, and they are used as output variables of SVM model. Field experiments are carried out in JiangYan freeway to collect the required data to validate the SVM model. Results show that the model can recognize driver's fatigue levels effectively and recognition precisions of all states are larger than 87.5%.