Investigating driver’s attention while monitoring an autonomous vehicle


At a glance

The recent advent of autonomous vehicles in the last decade has changed the traditional role of drivers, from being fully responsible of all driving functions to adopting a supervising role in an automated vehicle. As of today, autonomous vehicles are not perfect systems, they may make mistakes and require the driver to take control. Drivers must then monitor the automated vehicle behaviors to ensure safety of such technology. The goal of this proposed project is to investigate whether drivers can monitor efficiently the actions of an automated car. To this purpose, we plan on analyzing drivers gaze and attention maps in a driving simulator in different degrees of car automation, from fully manual to fully autonomous vehicle. Distribution of attention of a driver in fully manual mode can be used as ground truth of where an attentive driver looks at. Here, we propose comparing this baseline to attentional maps from drivers in the automated conditions and use any deviation as an indicator of driver’s inattentiveness. A driver monitoring system can be developed based on these findings to alert drivers to maintain attention on critical situations.

David Whitney
Peggy Wang
Teresa Canas-Bajo
Ye Xia
Driver attention, Human attention, Autonomous cars, Self-driving vehicles, Eye movements, Driver monitoring system