A study on driver fatigue recognition based on SVM method

Fatigue driving is one of the major reasons causing road traffic accidents. With the purpose to propose one recognition fatigue model, the measurements of vehicle control characteristics, physiological and psychological under fatigue have been studied through dynamic and static driving tests in the paper. ANOVA test results indicated that there is a significant difference between normal condition and fatigue driving status regarding the following four indicators: night vision, simple reaction time, choice reaction time and speed perception. Therefore, all these four measurements can be used as physiological and psychological parameters for driver's fatigue driving recognition model. With the aim to effectively identify fatigue driving, eight parameters, which reflects the state of the vehicle motion and the physiological & psychological status of the driver were selected to establish the driving fatigue identification model by using support vector machine algorithm. The model test showed that the accuracy rate of fatigue detection is up to 80.83%. In conclusion, the model proposed in this study can effectively identify the driver's fatigue status.