Using Driver State Detection in Automated Vehicles

The next several years will bring a large increase in automated vehicle capabilities. High levels of automation will require bi-directional transfers of control between the driver and vehicle. These control transfer situations pose one of the greatest potential safety shortfalls. One specific issue that arises is that drivers may be unfit or ill-prepared to retake control from the vehicle because of distraction, drowsiness, or intoxication. Driver state monitoring systems based on eye tracking, head tracking, and other measures may be useful in such situations. The goal of this project was to examine how driver state monitoring could be used in the context of an automated vehicle. Using data from a production driver monitoring system, the authors examined two approaches to using driver state information. One method provided feedback throughout the drive when drivers were classified as distracted (i.e., attentional maintenance). The other method utilized state-contingent takeover messages, which provided earlier warnings when drivers were distracted. These were compared against a baseline drive in which the automation did not use driver state information. The results indicated that providing attentional maintenance alerts throughout the drive increased drivers’ situational awareness and enhanced takeover during unexpected automation failures. Although state-contingent takeover requests improved some components of the takeover process, there was limited evidence that they improved takeover performance relative to baseline. This study highlights the potential utility of data from driver monitoring systems in the context of partial vehicle automation.

Language

  • English

Media Info

  • Media Type: Digital/other
  • Features: Figures; Photos; References; Tables;
  • Pagination: 28p

Subject/Index Terms

Filing Info

  • Accession Number: 01685329
  • Record Type: Publication
  • Contract Numbers: 69A3551747131
  • Files: UTC, NTL, TRIS, ATRI, USDOT
  • Created Date: Nov 14 2018 4:51PM