Users’ response to critical situations in automated driving: rear-ends, sideswipes, and false warnings
Artikel i vetenskaplig tidskrift, 2020

Although an automated vehicle may operate for an
extended time, it may suddenly request a user’s intervention in
critical situations (i.e., beyond the system’s operational design
domain). Despite the proliferation of studies to understand how
users resume control in such critical situations, a systematic
analysis of the whole response process is necessary. We analyzed
the visual-motor response process of distracted users to front and
lateral vehicle conflicts. We also investigated the effect of false
warnings and expectations. In a driving simulator experiment
(high fidelity, fixed-based, within-subject design), 45 participants
performed a visual-manual distracting task until an audio
warning was issued. The response process was modeled with
Bayesian generalized linear mixed effects models. The models
incorporate the carryover effect (up to the 2nd order), typical
side effect of within-subjects experiments. Reaction times were
modeled with a shifted-Wald distribution; response choices with
a softmax regression. The warning was effective at capturing
visual attention and prompting the resumption of control, but
it did not directly initiate an intervention. Glance location
and the choice and timing of evasive maneuver depended on
driving context and on previous experience. Analysis of the whole
response process yields more relevant information on the effect
of warnings for transition of control than a single measure of
intervention time. Furthermore, the carryover effect should not
be discounted, because trial randomization can only partially
alleviate the problem. We provided results in a format that can
be used as a reference for future studies and for computational
models of driver behavior.

response chain

take over request

reaction time

collision warning.

Bayesian data analysis

Författare

Alberto Morando

Chalmers, Mekanik och maritima vetenskaper, Fordonssäkerhet, Olycksanalys och prevention

Trent Victor

Chalmers, Mekanik och maritima vetenskaper, Fordonssäkerhet

Marco Dozza

Chalmers, Mekanik och maritima vetenskaper, Fordonssäkerhet, Olycksanalys och prevention

IEEE Transactions on Intelligent Transportation Systems

1524-9050 (ISSN)

Human Factors of Automated Driving (HFAUTO)

Europeiska kommissionen (FP7), 2013-11-01 -- 2017-10-31.

Ämneskategorier

Maskinteknik

Psykologi

Styrkeområden

Transport

Mer information

Senast uppdaterat

2020-02-19