A study of analysis method for driver features extraction

Nowadays, research and development in vehicle safety system has been intensively conducted. However, research focusing on detection unsafe signal from driving behavior is much less defined and explored. Abnormal driving behavior is a result of instability of internal state of the driver that could lead to uncomfortable and dangerous situation in driving task. In this study we described a method for abnormal driving behavior detection and classification using statistical analysis. Two parts of driving data called as start and stop period were used in the analysis. Preliminary result from analysis shows that the driver control behavior can be characterize into normal and abnormal using the proposed method based on jerk information. Through questionnaire survey, the relationship between driving behavior and internal state of driver is verified.