Drivers’ Glance Behavior During Driver-Initiated Overrides of Level 2 Driving Automation: A Comparison Between Intentional and Unintentional Overrides

During the use of partial driving automation, transfers of control (TOC) between the driver and automation happen routinely while the driver supervises and manages the driving environment. Although driver-initiated TOCs are more prevalent, they have not been explored to the same extent as system-initiated TOCs due to their low criticality. Drivers can initiate TOCs intentionally and unintentionally, and unintended driver-initiated overrides may lead situations where a driver resumes manual driving before they are prepared. This study examines glance behavior during driver-initiated overrides, where a driver is an initiator and a receiver of control during the TOC, including unintended overrides. Drivers’ glance data from 10 participants driving a Tesla Model S were analyzed. The authors examine differences in glance behavior (a) during overall driving with Tesla Autopilot (AP) and driving without it (non-AP), (b) 30-20s, 20-10s, and 10-0s before and after overrides, and (c) after unintended overrides. Results showed that (a) glance frequency to the instrument panel (IP) was higher during non-AP driving compared to AP driving, (b) glance frequency to the IP was higher near overrides (0-10s before and after overrides), and (c) glance frequency to the IP was higher after unintentional overrides compared to intentional overrides. Drivers’ increased glance frequency to the IP before and after overrides, especially after unintentional overrides, may indicate drivers’ information needs. The driver-system interface should provide the information that is needed during overrides (e.g., confirming a successful override, checking the status of the automation). This will enable safe and more efficient information extraction for drivers.

• Record URL:
  • https://annualmeeting.mytrb.org/OnlineProgram/Details/19456
• Supplemental Notes:
  • This paper was sponsored by TRB committee ACH30 Standing Committee on Human Factors of Vehicles.
• Corporate Authors:

  Transportation Research Board

  ,    
• Authors:
  • Lee, Joonbum
  • Kamaraj, Amudha V
  • Li, Mengyao
  • Domeyer, Joshua
  • Toyoda, Heishiro
  • Lee, John D
• Conference:
  • Transportation Research Board 102nd Annual Meeting
  • Location: Washington DC, United States
  • Date: 2023-1-8 to 2023-1-12
• Date: 2023

Language

• English

Media Info

• Media Type: Digital/other
• Features: Figures; References;
• Pagination: 13p

Subject/Index Terms

• TRT Terms: Alertness; Autonomous vehicle handover; Behavior; Drivers; Eye movements; Level 2 driving automation
• Identifier Terms: Tesla automobile
• Subject Areas: Highways; Safety and Human Factors; Vehicles and Equipment;

Filing Info

• Accession Number: 01874945
• Record Type: Publication
• Report/Paper Numbers: TRBAM-23-04647
• Files: TRIS, TRB, ATRI
• Created Date: Feb 27 2023 9:32AM