Modeling and Recognition of Driving Fatigue State Based on R-R Intervals of ECG Data

Abstract

Driving fatigue is an important contributing factor to traffic crashes. Developing a system that monitors the driver's fatigue level in real time and produces alarm signals when necessary, is important for the prevention of accidents. In the past decades, many recognition algorithms were developed based on multiple indicators. However, the relationship between R-R intervals of ECG and driving fatigue has not been studied. We develop a model to recognize the driving fatigue state based on R-R intervals. The cluster effect in the R-R interval sequence is found based on the stationary test and ARCH effect test. Then the AR (1)-GARCH (1, 1) model is developed to fit the time sequence of R-R intervals. The conditional variance of the residual R-R sequence is used to recognize whether there are changes in driving states. Field data was collected on the freeway between Baicheng and Changchun cities, where 10 drivers took part in the experiments. Validations are conducted to test the effectiveness of the developed model, and the results show that the recognitions of driving fatigue for the 10 drivers are correct in all cases. In addition, the recognition time delay is smaller than 5 minutes.