Journal
ACCIDENT ANALYSIS AND PREVENTION
Volume 154, Issue -, Pages -Publisher
PERGAMON-ELSEVIER SCIENCE LTD
DOI: 10.1016/j.aap.2021.106099
Keywords
Automated driving; Take-over; Visual distraction; Driving performance; Human factors
Categories
Funding
- Tsinghua University-Toyota Joint Research Center for AI Technology of Automated Vehicle [TTAD2020-05]
- National Natural Science Foundation of China [51965055]
Ask authors/readers for more resources
This study focuses on the issue of visual distraction during take-over in automated driving systems, proposing an evaluation method based on face orientation and implementing adaptive time budget adjustment through a regression model.
There are certain situations that automated driving (AD) systems are still unable to handle, preventing the implementation of Level 5 AD. Thus, a transition of control, colloquially known as take-over of the vehicle, is required when the system sends a take-over request (TOR) upon exiting the operational design domain (ODD). An adaptive TOR along with good take-over performance requires adjusting the time budget (TB) to drivers? visual distraction state, adhering to a reliable visual-distraction-based take-over performance model. Based on a number of driving simulator experiments, the percentage of face orientation to distraction area (PFODA) and time to boundary at take-over timing (TTBT) were proposed to accurately evaluate the degree of visual distraction based on merely face orientation under naturalistic non-driving related tasks (NDRTs) and to evaluate take-over performance, respectively. In order to elucidate the safety boundary, this study also proposed an algorithm to set a suitable minimum value of the TTBT. Finally, a multiple regression model was built to describe the relationship among PFODA, TB and TTBT along with a corrected coefficient of determination of 0.748. Based on the model, this study proposed an adaptive TB adjustment method for the take-over system.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available