At the Institutes of Innovation for Future Society at Nagoya University, the HMI and Human Characteristics Research Division studies interaction design between humans and intelligent systems for novel mobility interfaces and services. These researchers are working to introduce new interfaces and services based on the analysis of driver behaviors behind the wheel. Eye tracking solutions empower their researchers to recognize crucial gaze motions of drivers without any intrusion.
“Driver gaze motions are one of the primary clues to understanding intentions behind the wheel and evaluating the acceptability and usability of HMI.”
In Japan, road accidents caused by elderly drivers are increasing. On top of this, cars are becoming ever more crucial for their continued independence in their daily lives. This is why the Institute for Innovation for Future Society, Nagoya University, used a study to propose a driver-agent system that supports elderly car users and encourages them to improve their driving in general.
In the eye tracking study, the Nagoya University laboratory gathered thirty male and female drivers between the ages of 50 and 76 to observe how they perform within a driving simulator when approaching a stop sign at an intersection. The data drawn from this analysis was able to show visual information processing abilities, awareness allocation function, and field of view.
Researchers experimented with the use of three distinct driver agents: a vocal, a visual, and a robot agent. These were applied in simulators, set-up with Tobii Pro X2-30 eye trackers, offering the same support to elderly drivers in order to determine the most effective version. The simulator presented a driving situation with a high collision rate amongst elderly drivers, and the different agents provided driving guidance through the danger zone. The eye trackers recorded gaze data for insight into attention driver distraction.
Overall, results of the study showed that, contrary to initial hypotheses the robot driver agent was the most effective at encouraging safe driving behaviors and preventing accidents. It was the most noticeable, least intrusive, and most acceptable to the majority of elderly drivers. Fixation points were most converged in the robot condition. and gaze time corresponded to a very short duration, implying that a similar agent does not necessarily lead to huge driving disturbance.
Voice condition Visual condition Robot condition
The vocal-only agent was the least helpful for elderly drivers despite the assumption that it would be the easiest to notice. It caused the most diverged gaze patterns, possibly indicating that drivers couldn’t understand information being provided and instinctively searched to find the source.
Takahiro Tanaka, Kazuhiro Fujikake, Takashi Yonekawa, Misako Yamagishi, Makoto Inagami, Fumiya Kinoshita, Hirofumi Aoki, Hitoshi Kanamori (2018).
Study on Driver Agent based on Analysis of Driving Instruction Data - Driver Agent for Encouraging Safe Driving Behavior (1)
Takahiro Tanaka, Kazuhiro Fujikake, Takashi Yonekawa, Makoto Inagami, Fumiya Kinoshita, Hirofumi Aoki, Hitoshi Kanamori (2018).
Takahiro Tanaka, Kazuhiro Fujikake, Yuki Yoshihara, Nihan Karatas, Hirofumi Aoki, Hitoshi Kanamori (2020)
Takahiro Tanaka, Kazuhiro Fujikake, Yuki Yoshihara, Nihan Karatas, Kan Shimazaki, Hirohumi Aoki, Hitoshi Kanamori (2020).