Regaining Situational Awareness as a User in Charge: Responding to transition demands in
automated vehicles
5
Non-Technological Activities:
• Reading a Magazine: Participants chose from a selection of magazines (e.g., BBC
Top Gear, National Geographic) and were instructed to pick up the magazine from
the passenger seat and start reading once the trial began.
• Completing a Wordsearch or Sudoku: Participants were given the choice between a
word search puzzle book or a Sudoku book, along with a pen. The materials were
placed on the passenger seat, and participants were instructed to start solving the
puzzle once the trial began.
Motoric Activities:
• Drinking Water: Participants drank water from a disposable coffee cup with a lid,
placed in the cup holder beside the driver. They were instructed to take frequent small
sips throughout the drive.
• Simulated Eating of “Popcorn”: Participants mimicked the action of eating by
transferring cotton balls from a packet into a cup holder attached to their chest. This
simulated the action of eating popcorn, avoiding the mess and potential distractions
of real food.
In each trial, participants received instructions on how to carry out the NDRA (or were
informed there would be no NDRA). The system operated in self-driving mode, and
participants were told to take manual control after a takeover request, "as soon as you feel
ready and safe to do so."
Automated drives lasted 2-4 minutes, ending with a takeover request. Takeover requests
were signalled by an auditory beep and a visual alert on the Human Machine Interface (HMI).
Participants had 30 seconds to take control. At UCL, manual control was regained when the
participant used the steering wheel or pedals, while at LU, participants were required to say,
“Ready to drive” and a researcher gave manual control. Participants then drove manually
for 30 seconds before completing a questionnaire on workload and self-perceived SA.
Data Collection and Analysis
A comprehensive set of measures were employed to assess SA following a takeover request
during automated driving. The analysis was designed to investigate whether participants
fully disengaged from NDRAs, their visual attention patterns post-takeover, particularly
focusing on whether they looked at mirrors or other critical areas of the driving environment,
and whether participants took appropriate behavioural actions, such as adjusting speed
based on the scenario (which would indicate their comprehension of the reasons behind the
takeover request). Additionally, the quality of the takeover was assessed by observing any
signs of erratic steering, such as swerving or crashing, which could suggest either poor
control or insufficient SA.
GoPro cameras captured in-cab behaviours, allowing for detailed analysis of how
participants managed the transition from NDRA to driving. Participants’ interactions with the
NDRA, their disengagement process, and their subsequent driving performance were all
recorded. This comprehensive data collection approach enabled the analysis of how
