Driving fatigue detecting method based on temperature insensitive ECG parameters

Abstract

Driving fatigue has become one of the main causes in transportation accidents. Driving fatigue detection and early warning become very important research area to enhance road safety. ECG parameters, which could illustrate driving fatigue state directly, play an important role in corresponding research works. However, ambient temperature is normally not considered in these methods, while according to medical research, temperature may affect ECG signal considerably. In this paper, a significant analysis based temperature insensitive ECG parameter selection method is proposed. An integrated ECG parameter obtain method, which includes both questionnaire process and ECG signal extraction, is presented. Meanwhile, ROC curve is employed to find fatigue decision threshold. The validity of the proposed method is verified by a separate data set. Test results demonstrate that the developed model and threshold selection method are effective.