Recognition method of driver's visual distraction based on pupil diameter

• Published in: 2018 33rd Youth Academic Annual Conference of Chinese Association of Automation (YAC) pp. 57 - 64
• Main Authors: Guo, Weiwei, Li, Yangyang, Tan, Jiyuan, Li, Yinghong, Yang, Shaohui, Ma, Xin, Jiao, Chengwu
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.05.2018
• Subjects: abnormal-data disposal; Automobiles; Information systems; pupil diameter; Roads; Safety; Traffic control; traffic safety; visual distraction; Visualization
• DOI: 10.1109/YAC.2018.8406347

In order to recognize the driver's visual distraction in the driving process, this paper chooses the information of pupil diameter as an indicator to study. At first, it designs three typical experimental scenarios of following, overtaking, and target search according to the research goal, selects 12 subjects and does the experiment of distracted driving using the driving simulation platform and the eye tracker. Then, the data is processed and repaired. After the static experiment verification threshold, the difference of the pupil diameter value was treated. The method of sliding standard deviation is used to automatically identify the abnormal mutation of pupil diameter, and use linear interpolation to repair. Finally, by the methods of variance, mean square root and recursive graph, this paper analyzes the changes of pupil diameter in three scenes such as gallop, overtaking and target search, and realizes the recognition of driver's visual distraction. The research results show that the changes of drivers' pupil diameter is more stable and small discrete degree in normal driving condition, on the contrary it is unstable and large discrete degree in distracted driving. At all, it can analyze the pupil diameter using means of variance, mean square root and recursion graph, and identifies the driver's visual distraction behavior.