Content uploaded by Vishal Kummetha
Author content
All content in this area was uploaded by Vishal Kummetha on Apr 15, 2019
Content may be subject to copyright.
*This is an Accepted Manuscript of an article published by Taylor & Francis in the JOURNAL
OF TRANSPORTATION SAFETY & SECURITY on 05 December 2018, available
online: https://www.tandfonline.com/doi/abs/10.1080/19439962.2018.1518359
Analysis of the Effects of Adaptive Cruise Control on Driver Behavior and
Awareness Using a Driving Simulator
Vishal C. Kummethaa*, Alexandra Kondylia, and Steven D. Schrocka
aDepartment of Civil, Environmental, and Architectural Engineering, University of Kansas,
Lawrence, Kansas 66045, United States.
*corresponding author, e-mail: kummetha@ku.edu
Vishal C. Kummetha is a Ph.D. student at the University of Kansas and the principal investigator in
this research. This article is derived from the methodology and results obtained from the
corresponding author’s Master’s thesis.
Address: 2160 Learned Hall, 1530 W. 15th Street, Lawrence, KS, 66045
Dr. Alexandra Kondyli is an assistant professor in the Department of Civil, Environmental, and
Architectural Engineering, and the lead supervisor for this research project.
Address: 2159A Learned Hall, 1530 W. 15th Street, Lawrence, KS, 66045
Email: akondyli@ku.edu
Dr. Steven D. Schrock is an associate professor in the Department of Civil, Environmental, and
Architectural Engineering. His knowledge in driving simulation and study design provided key
insights for this research.
Address: 2159B Learned Hall, 1530 W. 15th Street, Lawrence, KS, 66045
Email: schrock@ku.edu
2
Analysis of the Effects of Adaptive Cruise Control on Driver Behavior and
Awareness Using a Driving Simulator
Adaptive cruise control (ACC) systems assist drivers by automatically adjusting
vehicle speed with respect to the driver selected time-gap to the lead vehicle and
maximum desired speed. Although ACC systems have been around for 20 years,
drivers are not always fully aware of their limitations. The main focus of this research
was to investigate the limitations of the ACC system and the resulting effects on driver
performance, taking into account variations of mental workload. Thirty participants
drove two simulated scenarios, with and without the ACC system active, in a
fixed-base driving simulator. The variability of the observed parameters between the
drives with and without the ACC provided insights into individual driver performance,
impact of age and gender, and behavioral trends resulting from changes to mental
workload. Data on cognitive workload were measured using a detection response task
(DRT) device to monitor availability of mental resources during the various tasks. The
statistical analysis showed that with an active ACC system, participants reached lower
maximum speeds, maintained longer headways, and reacted slower and more abruptly
to sudden events. The data from the DRT showed a significantly lower cognitive load
when participants were subject to secondary tasks during the ACC scenarios.
Keywords: Adaptive cruise control; driver behavior; driver awareness; detection
response task; cognitive workload
1. Introduction
ACC systems are becoming increasingly available in new vehicle models. These systems
automatically adjust longitudinal position based on the driver selected headway and driver
selected speed (Pauwelussen & Feenstra, 2010). ACC systems are intended to increase
overall roadway safety, especially on highways and rural arterials with higher traveling
speeds, by decreasing the overall mental and physical workload associated with the driving
task (Vollrath et al., 2011). ACC is marketed by automobile manufacturers as an advanced
3
guidance system that enhances convenience, comfort, and safety by reducing the stress of
driving (Heritage Ford, 2016, Hyundai, 2017, and Strand et al., 2014).
While researchers and developers strive to make autonomous vehicles a reality,
achieving full automation requires introduction of partial automation with a gradual increase
in the level of automation and supporting infrastructure to allow users to adapt accordingly.
This research mainly focuses on two levels of automation, defined by the Society of
Automotive Engineers (SAE) as level 0 (driver only) and level 1 (driver and ACC system)
(SAE-J3016, 2016). The research was undertaken to assess the safety and user experiences
associated with level 1 automation compared to level 0, as these experiences can also be
referenced during the implementation of higher automation levels. This is critical, as drivers
are not always fully aware of the limitations of their system (Jenness et al., 2008). In addition,
literature findings show contradictory effects of ACC regarding headways, time-to-collision,
standard deviation of lateral positions, etc. As such, the main objective of this research was
to investigate limitations of the ACC system and the resulting effects on driver performance
arising from these limitations across different age groups, taking into account variations of
mental workload.
Although the ACC system is controlled by an onboard network of modules and
sensors, the driver is still required to control the steering, set the maximum speed, select the
preferred time-gap, and apply brakes when greater deceleration rate than the pre-programmed
system is required. Comparing performance variables such as average distance headway,
time to collision, standard deviation of lateral position, maximum deceleration, maximum
braking force, and maximum speed, for a variety of scenarios involving driving with and
without ACC provided an insight into individual driver performance with respect to the above
4
variables, time-gap preferences, impact of age and gender, and behavioral trends resulting
from changes to mental workload.
2. Literature Review
ACC systems have been available in vehicles for the last 20 years. Automobile manufacturers
such as Mitsubishi, Toyota, and Mercedes Benz were the first to incorporate this technology
into their luxury (high-end) vehicles, although nowadays they are becoming more affordable
and readily available from most car manufacturers.
Engaging and disengaging the ACC is similar to that of the conventional cruise
control. The driver activates the cruise and sets it at a desired speed and uses the brake,
accelerator pedal, or cancel button to deactivate the system. In addition, drivers can usually
select from a choice of three or four time-gap settings to control the time-gap to the lead
vehicle (Vollrath et al., 2011). ACC is usually equipped with a warning system to alert the
driver in case the system fails or cannot brake in time and requires manual input.
Several studies have evaluated the influence of ACC on driver behavior. Ohno (2001)
carried out a simulator study on the adaptation process of driving behaviors using ACC.
Drivers were observed to maintain longer headways with the ACC compared to manual mode
(Ohno, 2001). Rudin-Brown and Parker (2004) carried out a similar study to Ohno (2001)
using a luxury sedan on a closed 6.9 km test track. The lead vehicle was attached to a
polyurethane trailer to avoid injury to the participant in case of a crash. The collected data
included braking times, lane keeping, sleepiness, trust, and subjective workload. The study
showed that drivers reacted slowly when braking in critical situations and had reduced lane
keeping ability when driving with the ACC.
5
Ma and Kaber (2005) carried out a series of workload experiments using a
low-cost virtual reality simulator. Eighteen participants drove the scenarios with and without
ACC. The study also collected data on changes to situational awareness and mental demand
when engaged in a secondary task (cell phone usage) while driving with and without the ACC
system. This study measured workload on a subjective scale, with questionnaires requesting
feedback about the intensity of the task performed. A reduction in overall mental demand
was observed with the ACC. Participants also showed more consistent following speeds and
headways during the drive with the ACC.
Cho et al. (2006) performed a driving simulator study where they recorded the
headway and lateral position of forty participants with and without the ACC. The researchers
established that participants preferred a 1.5 second time-gap setting in the ACC system.
Similar to Rudin-Brown and Parker (2004), drivers demonstrated reduced lane keeping
ability with the ACC, implying decreased attentiveness to the roadway and surroundings,
although the impact of ACC on lateral driving performance is not as straightforward.
Vollrath et al. (2011) carried out a driving simulator study to determine the influence
of cruise control (CC) and ACC on driving behavior. The study required participants to safely
engage in as many secondary tasks as possible while driving a simulated scenario. Vollrath
et al. (2011) found that compared to cruise control, ACC did not cause significant delays to
reaction times due to increased engagement in secondary tasks when intervening in critical
situations. Larsson et al. (2014) compared driving without any automation to user experience
with the ACC. The results obtained showed experienced ACC users were half-second quicker
during braking events when compared to novice users. However, when compared to no
automation, drivers were on average 2 seconds slower.
6
Driving requires a significant amount of cognitive resources to complete a particular
task. Several methods are available to measure the cognitive resources being used during
driving. These methods can be grouped into: principal measures (e.g., difference in speeds),
subjective measures (e.g. NASA task load index (NASA-TLX) and rating scale mental effort
(RSME)), psychophysiological measures (e.g. electroencephalogram (EEG),
electrocardiogram (ECG), pupilometer), and detection response tasks (DRTs). Each of the
above methods has pros and cons. The NASA-TLX is one of simplest and the most widely
used subjective measure; however, it has been shown that the answers to the questionnaire
are strongly influenced by the last task performed (Stojmenova & Sodnik, 2015). Also, the
NASA-TLX does not provide time-varying data. On the other hand, psychophysiological
measures can provide time-varying data of workload but tend to be expensive and require
operation by well-trained staff (Stojmenova & Sodnik, 2015). The DRT method is cheaper
compared to psychophysiological measures, but presenting visual stimuli tends to introduce
added workload on the subjects (Stojmenova & Sodnik, 2015). Overall, all methods provide
similar workload results when used correctly (Strayer et al., 2013). However, some methods
are more sensitive to the task at hand than others (Cooper et al., 2014; Strayer et al., 2013).
De Winter et al. (2014) summarized the results obtained from various studies on
reaction time to visual stimuli when driving with and without the ACC using a DRT. The
DRT device presents frequent visual stimuli in the form of a red square or blue LED lamp.
Participants were required to respond by either pressing the horn or steering wheel buttons,
and their response times (RT), hit rates (HR), and miss rates (MR) were recorded. It was
observed that participants had a 14.4% greater hit rate when driving with the ACC than
7
without the ACC. Reaction time to the visual stimuli was also observed to be quicker, by up
to 15%, for participants driving with ACC.
In summary, literature is not conclusive on the effect of ACC on driver behavior and
awareness. Some studies demonstrated delays in braking time, reduction in lane keeping
ability, and decrease in attention towards the surroundings while others showed no significant
impact on reaction times and alertness. However, the variations might have been caused due
to different task instructions, lead vehicle behavior, scenario specifics, and participant
instructions. This research aims to validate some of the results stated above as well as provide
an insight into how different age categories of drivers perform with the ACC, using
non-subjective measures of both workload and awareness. In addition, investigating the
effect of ACC on varying mental workload using the DRT provides a better understanding
of driver behavior and its relationship to driving performance, especially when engaged in a
secondary task, providing supplementary data to existing naturalistic driving studies.
3. Methodology and Data Collection
The methodology followed in this research consisted of six main tasks as shown in
Figure 1.
(Figure 1 goes here)
3.1. Participants Recruitment
The initial proposal, consent forms, and methodology of the study was submitted to the
Human Subjects Committee – Lawrence (HSCL) for approval. After approval from the
HSCL, the study was advertised using bulletin boards in public locations around Lawrence,
Kansas and social media platforms such as Facebook and craigslist. Participants were
8
prescreened using questionnaires to determine personal and driving information such as
contact information, age, gender, possession of a valid U.S. driver’s license, model/year of
current vehicle, experience with ACC systems, estimate of a safe car following distance,
existing medical conditions, willingness to use ACC systems, willingness to participate in a
simulator based study, and history of motion sickness (Kummetha, 2017).
44 potential participants responded to the posted advertisements and expressed
interest in participating in the study. From those, thirty participants between the age of 20
and 65 years, with at least one year of driving experience and valid U.S. driver’s license, in
good health (free from heart conditions, seizures, inner ear/balance problems, and possibility
of pregnancy), and low motion sickness severity, were selected. None of the participants had
previous experience with the ACC.
The average age of the thirty participants was 37.8 years (SD = 17.1 years), and they
were equally split between males and females. Participants were categorized into three age
groups: 18-24 years, 25-49 years, and 50-65 years, each consisting of ten participants
(Kummetha, 2017).
3.2. Setting up the Simulator
The University of Kansas (KU) driving simulator is a fixed-base simulator with both the
Acura MDX vehicle chassis and the display screens mounted to the laboratory floor. The
scenarios are displayed onto the screens using overhead projectors. The three front screens
provide a 170o horizontal field of view (FOV) as seen in Figure 2 (a). A rear screen is also
available to render images in the rear-view mirror and side mirrors. This allows the simulator
to deliver an all-round display, providing a more immersive driving experience. The cab also
consists of a digital instrument panel that is activated when a scenario starts. The panel
9
interface is shown in Figure 2 (b), consisting of the speed, cruise control status, ACC time-
gap, gear shift, and turn signals.
(Figure 2(a) and 2(b) go here)
3.2.1. Configuring the ACC system
The ACC model used in this research is the same as the one implemented in the National
Advanced Driving Simulator (NADS) 1 (Moeckli et al., 2015). The ACC model operates in
either cruise control or following mode as shown in Equation 1.
Where, r is the distance to the leading vehicle, v is the speed of the ACC equipped
vehicle,is the desired speed set with the cruise control, is the range rate (velocity of the
ACC equipped vehicle subtracted from the velocity of preceding vehicle), and rmax is the
maximum allowable distance between vehicles. The model caps acceleration of the vehicle
by using the global minimum acceleration (0.1g) when following a vehicle and a global
maximum acceleration (0.2g) when engaged in free driving mode (Moeckli et al., 2015). This
prevents the ACC vehicle from under- or over-using the throttle. The NADS model calculates
the time to collision with the lead vehicle and the rate at which the distance to the lead vehicle
is decreasing. It is restricted to brake with a maximum possible deceleration of 0.3 g. If a
deceleration rate is greater than 0.3 g, and the time to collision is less than three seconds, the
driver is alerted with an audible warning sound, requiring manual intervention to generate a
greater deceleration by stepping on the brakes in order to prevent collision (Moeckli et al.,
10
2015). The model also caps the maximum detection range of the lead vehicle to 122 m and
the minimum detection range to 5 m.
The MiniSim software (Minisim User’s Guide, 2015) links the ACC model to the
steering wheel from which ACC can be manually activated by the driver. The time-gap is
adjusted using repurposed volume control buttons present on the steering wheel. Participants
were given a choice of three time-gap settings, 3 seconds, 2 seconds, and 1.2 seconds. They
were free to toggle through these choices at any point during the ACC drive. Figure 3 shows
the modified steering wheel capable of adjusting the ACC time-gap settings.
(Figure 3 goes here)
3.2.2. Roadway Geometry
As ACC systems are typically used on roadways with higher speeds and preferably lower
number of interacting vehicles (Strand et al., 2011), a four-lane divided highway with a grass
median was selected consisting of four horizontal curves and three interchanges. The
interchanges were used as a hub to add and remove simulated traffic. The total length of the
highway was 20 km and the drive took approximately 13 to 16 minutes to complete
depending on the driver.
3.2.3. Configuring Events and Tasks
A total of three events lasting short durations and four tasks lasting longer durations were
configured, each aimed at collecting a specific set of variables. The events include: crossing
animal (deer), desk falling out of a van, and sudden merging vehicle. The tasks include: car
following, driving through a work zone, moving over for a stopped law enforcement vehicle,
and attempting a secondary task while driving. The scenarios were configured to have at least
11
five minutes of base driving between successive events or tasks. All events and tasks were
configured at 113 km/h speed limit, except for driving through a work zone which was
configured at a speed limit of 88.5 km/h. The traffic conditions were free-flowing throughout
all scenarios.
Crossing animal (deer): In this event, the participants were required to perform an
evasive maneuver to avoid hitting two deer running across the roadway. The deer were
programmed in such way that the collision was unavoidable. This was done to ensure that
most participants only use their brakes to avoid collision. Time to collision with the deer
when an evasive maneuver was performed was recorded along with maximum speed,
standard deviation of lateral position, and average distance headway for 600m before
encountering the deer.
Desk drop: This event comprised of two sequential sub-events. Participants were
required to attempt a secondary task simulating distracted driving (to emulate a high
workload situation with a greater task demand than crossing animal or sudden merging
vehicle events), during which a desk was dropped from the lead vehicle. Participant reaction,
in terms of time to collision, was measured based on their ability to perceive and perform an
evasive maneuver by applying the brakes, adjusting the steering wheel angle, and
speeding-up. Three main variables were collected during this event. The average distance
headway and standard deviation of lateral position were collected 500m before encountering
the desk, and time to collision with the desk was collected at the instance of the evasive
maneuver. Speed was not used as a dependent variable during this event because the desired
speed of the participants was not met as two vehicles were programmed to merge ahead in
12
order to drop a desk. Figures 4a and 4b show the event from a driver and design perspective,
respectively.
(Figure 4(a) and 4(b) go here)
Sudden merging vehicle: In this event, participants were required to react to a
vehicle merging suddenly from the right lane to the left without using turn signals. The length
of the event was 300m and participants had to react by applying brakes to avoid the collision.
Variables such as maximum braking force, maximum deceleration, and time to collision at
the instance when brakes were applied were collected.
Car following: This task required the participants to maintain a preferred headway
to the lead vehicle with and without the ACC system. The lead vehicle was programmed to
maintain a constant velocity of 113 km/h for a length of 4800m. Average distance headway,
standard deviation of lateral position, and maximum speed were collected.
Driving through a work zone: In this task, the awareness of the participants was
measured based on their ability to read and process traffic signs. The left lane and shoulder
were closed using traffic channelizers and the work zone was 3000m long. The speed limit
in the work zone was set at 88.5 km/h while a lead vehicle was programmed to violate the
set speed limit by travelling at 113 km/h. Participants’ ability to navigate the roadway based
on the speed regulations versus lead vehicle influence was measured. An average speed
above 104.6 km/h in the work zone was considered a violation.
Moving over for a stopped law enforcement vehicle: During this task, participants’
ability to successfully move over when a stationary law enforcement vehicle was encountered
alongside the curb was monitored. Three such tasks were placed in random locations.
13
Attempting a secondary task while driving: Participants were required to drive the
vehicle while using the touch screen interface (perform a secondary task), shown in Figure
6, designed to simulate in-vehicle distractions. Three distraction tasks were present in every
scenario of the drive, each with a duration ranging from 10 to 15 seconds depending on the
speed of the driver. The main variable recorded during the event was the number of hit
attempts on the touch screen interface as a measure of secondary task performance. Hit
attempts were analyzed together with the DRT data collected during the task, to determine
whether the reduction of mental workload as a result of using the ACC system significantly
altered the number of hit attempts recorded on the application.
Six highway scenarios were created, three intended for driving without the ACC and
three with the ACC system. The three variations of the highway scenario were created to
counterbalance the sequence of events and tasks. Each variation had at least one male and
one female participant per age group.
A tutorial scenario consisting of a two-lane undivided highway was created in order
to familiarize the participants with the DRT equipment, distraction interface, and in-vehicle
systems. Participants were allowed to proceed to the actual scenario if they showed proper
understanding of the use of ACC system. During the tutorial scenario, two participants
showed severe signs of motion sickness within 2-5 minutes and were replaced with other
candidates to fulfill the selected sample size.
3.3. Secondary Task and Workload
The distraction interface was used to test how the ACC affects the performance of a
secondary task. The DRT device was setup to measure changes to cognitive workload.
Participants responded by pressing a micro-switch attached to a finger of their choice (Figure
14
5). When a stimulus was presented, the RT to the stimulus was calculated and the total
number of successful hits in a specific duration of interest was recorded. Performance
changes were observed when cognitive workload increased or decreased depending on the
task.
(Figure 5 goes here)
3.3.1. Secondary Task
To simulate distractions in the vehicle, a Windows-based application was designed using
VB.NET (Figure 5). The application was modeled to simulate in-vehicle distractions caused
when using devices such as the media controller, climate controller, GPS device, or cell
phone (Kummetha, 2017). The layout consisted of nine tiles with numbers varying randomly
between zero and eight as shown in Figure 6. Participants were asked to match the number
shown in the top box with one in the tile, while driving.
(Figure 6 goes here)
The application recorded the correct matches (hits) and total attempts (clicks), to
determine any changes in performance with respect to the secondary task.
3.3.2. DRT Device
For this study, a head-mount DRT device with a micro-switch was used. The DRT stimuli
were presented in accordance with the ISO-17488 2016. A red LED was presented at
intervals ranging between three to five seconds with a duration of one second. The RT was
collected in microseconds and only responses that occurred between 100ms to 2500ms were
considered as hits. Any responses earlier than 100ms were regarded as a premature hit while
15
responses that took longer than 2500ms or that never occurred, were considered a miss. The
average RTs for successful hits were then calculated per event/task. HRs were also calculated
based on the ratio of hits to total stimuli presented during the event/task.
3.4. Data Collection
Upon arrival to the driving simulator lab, participants were required to sign the informed
consent form describing the purpose of the study, potential side effects such as simulator
sickness, headache, sweating, etc., value of monetary compensation ($20), participant
confidentiality, and contact information for the principal investigator in case of any concerns
after the study. Next, they were familiarized with the simulator by adjusting the seats and
mirrors according to their preference. The DRT stimuli headband was attached to their
forehead while the micro-switch was attached to their left thumb. It was emphasized to the
participants that the DRT was not meant to measure their individual skill but to highlight any
differences in performance when using the ACC. The DRT device was then started and
participants were required to respond to ten stimuli to get used to the mechanism. The
distraction application stated in section 3.3.1 was explained and the voice command that
requires participants to start attempting the secondary task was played.
The ACC time-gap and cruise control switches on the steering wheel were then
explained to the participants. Participants were asked to drive the tutorial scenario that
required them to follow an ambulance at 88.5 km/h while responding to the DRT stimuli.
After two minutes of normal driving, participants were required to engage in ACC and try
all three time-gap settings while following the ambulance. This usually lasted between three
and ten minutes. Participants were then given an opportunity to ask any questions or concerns
they had with the system and also fill out a wellness questionnaire to screen for motion
16
sickness. After the tutorial scenario was completed, participants who demonstrated full
understanding of the controls were allowed to proceed to the actual scenarios. Participants
were instructed to drive as they would normally on a freeway. In the ACC scenario,
participants were free to engage and disengage as pleased. At the end of the study,
participants were kept engaged for five minutes in order to monitor any signs of
disorientation. Also, a questionnaire on the realism of the simulator was provided to identify
potential improvements to the system/simulation.
The collected dependent variables include: maximum speed, average distance
headway, time to collision, steering wheel angle, standard deviation of lateral position,
maximum braking force, and maximum deceleration. Variables were collected multiple times
through different events and tasks to establish a pattern.
3.5. Data Reduction and Statistical Analysis
A between-subjects (three levels of age groups by two levels of gender) analysis of
variance (ANOVA) was first carried out to determine if the data showed differences with
respect to the age groups, gender, or the interaction between age and gender. Since the same
participants drove both scenarios, the difference between the variable collected from the no
ACC drive to the same variable from the ACC was computed. This was done to satisfy the
independence criteria for the two-way ANOVA.
If a significant p-value (< 0.05) was obtained in the ANOVA, Tukey’s HSD post-hoc
test was performed to identify which age group or gender was different from the rest. If no
significance was found in the ANOVA, it implies that the mean of the variables collected
was not influenced by the age group or gender. A paired t-test was then performed on the
entire sample population to identify any significant differences.
17
The null hypothesis (H0) for the paired t-test stated that there is no difference in driver
behavior and awareness of individuals driving with the ACC and without the ACC. To test
this hypothesis, multiple two-tailed paired t-tests were performed at a 95% confidence level
using SPSS (IBM Corp., 2011). The dependent variables were categorized according to the
event/task and they include: collision count, application hit attempts, maximum speed,
average distance headway, standard deviation of lateral position, time to collision, maximum
braking force, maximum deceleration, successfully observed move over tasks, DRT response
time, and DRT hit rate.
4. Results
The ANOVA results in Tables 1 and 2 show that the average distance headway in the car-
following task was found to be significantly different (F(2, 24) = 6.032, p = 0.008,
= 0.335) between the age groups. Tukey’s HSD post-hoc test resulted in a significant
difference between the 18 to 25-year age group and the other two age groups (mean diff = -
60.28m, p = 0.013 for the comparison to the 25 to 49-year group and mean diff = -54.58m, p
= 0.025 for the comparison to the 50 to 65year group). However, no significant difference
(mean diff = 5.79m, p = 0.953) was observed between the 25 to 49-year and 50 to 65-year
age groups. The interaction term (age group*gender) of the standard deviation of lateral
position in the crossing animal event was found to be significant (F(2, 24) = 4.535, p = 0.021,
= 0.274). However, post-hoc tests confirmed no significant differences between the age
groups or gender. All remaining variables were not found to be statistically significant
between the two drives with and without the ACC.
(Table 1, 2 go here)
18
Out of the 30 participants, 76.7% preferred the 3-second time gap, 10% preferred the
2-second time gap, and 13.3% preferred the 1.2-second time gap as shown in Figure 7. This
was calculated by comparing the percentage of time each of the three time-gap settings were
used when ACC was active. The time-gap used predominantly during the drive was deemed
as the preferred setting for that participant. A one-way ANOVA for each time-gap setting
was also performed and resulted in no significant differences between the age groups
(1.2-sec: F(2, 27) = 1.676, p = 0.206; 2-sec: F(2, 27) = 0.700, p = 0.506; and 3-sec: F(2, 27)
= 1.057, p = 0.361).
(Figure 7 goes here)
Boxplots of average distance headway, maximum speed, time to collision, and
standard deviation of lateral position are shown by event in Figures 8, 9, 10, and 11. The
plots seem to show relatively normal data with equal variance satisfied.
(Figure 8, 9, 10, 11 go here)
A total of thirteen significant variables, grouped by events/tasks, were obtained from
the paired t-test. A summary of the results is shown in Tables 3 and 4.
(Table 3, 4 go here)
These tables show statistically significant differences between driving with ACC and
without ACC for a 95% and a 99% significance level for these 13 variables (p-value) as well
as the size of the effect (Cohen’s d value). Cohen’s value of greater than 0.80 suggests that
the effect of ACC on that particular variable is large, whereas a value between 0.40 and 0.80
suggests that the effect is medium. The analysis shows that participants have a 21.9% greater
number of hit attempts on the secondary task interface when driving with the ACC, indicating
more room and better performance for secondary tasks when using ACC.
19
The maximum speed was found to be significantly different (higher) without the ACC
than with the ACC, during the whole drive, car-following task, and animal crossing event.
The higher maximum speed without ACC can be attributed to the greater variability in overall
speed when compared to the ACC, resulting in high peak values. This can also be related to
the lower number of abrupt braking and acceleration interventions by the driver during the
ACC drive when compared to the no ACC drive. The effect size of the maximum speed
variable was also found to be large.
Participants were found to follow significantly longer headways when driving with
ACC during the car-following (by 21%) and the desk drop (by 29.6%) events. In addition, as
shown earlier, the average distance headway from the car following event resulted in
significant differences between drivers from 18 to 24-year age group and the others. The
effect of ACC on the average distance headway was medium.
Participants had larger standard deviation of lateral position (significant at a 95%
confidence interval) when traveling with ACC compared without ACC during car-following
(by about 9.3%). The overall effect of ACC on the standard deviation of lateral position was
small.
Time to collision was found to be significant in the crossing animal and sudden
merging events, but not in the desk drop event. The results showed a 30-35% decrease in
time to collision when driving with the ACC in the two events, which indicates an increase
in reaction times. The decrease in time to collision was a result of drivers applying brakes
later with ACC than without the ACC during sudden events. ACC was found to have a
medium effect on time to collision.
20
The maximum braking force and maximum deceleration were found to be
significantly different during the sudden merging event. Participants were observed to brake
with a 28.4% greater force and decelerate much quicker when using the ACC. ACC was
found to have a medium effect on these two variables.
Measurements of driver workload through the DRT showed significant difference in
the HR between the two drives during car-following and the secondary task distraction event.
A 5.1% decrease in HR was found during the ACC phase in car following and can be
attributed to engaging ACC and selecting the preferred time-gap. A 26.3% increase in HR
indicated that participants were less likely to miss a stimulus while performing a secondary
task with the ACC than without the ACC. It can be noted that the differences between the
ACC and no ACC drivers were quite small on an absolute scale.
5. Conclusions and Recommendations
This research established some significant conclusions regarding changes in driver behavior
between two levels of vehicle automation, SAE level 0 and level 1. Participants were
observed to reach lower maximum speeds when driving with the ACC. The average of the
three events resulted in a 5.2% decrease in the maximum speed attained by each participant
during the drive. This was consistent with the findings of Ma and Kaber (2005) who also
reported a decrease in the maximum speed.
The average distance headway was found to be significantly different in two out of
the three scenarios. However, the car following scenario indicated that age has an effect on
the headways maintained. Younger drivers (18-24 years) exhibited an increase in average
distance headway when using the ACC. This is a significant finding and more research is
suggested in this area. A 29.6% increase in the following distance was observed during the
21
ACC phase in the desk drop event. Ohno (2001) established similar results, where
participants were observed to maintain longer average distance headways with ACC.
Time to collision was found to be significantly different in two out of the three
configured events. A 32.5% decrease in time to collision during the ACC phase was
observed. This can also be interpreted as a 32.5% increase in reaction time as participants
took longer to perform an evasive maneuver. The increase in reaction time was consistent
with Rudin-Brown and Parker (2004). However, it is not clear whether the delay in reaction
time is attributed to using the ACC or due to an increase in distance/time to reach the brake
pedal from a resting foot position. This was also consistent with the results obtained by
Larsson et al. (2014), where participants were slower to brake when engaged in ACC than
without any automation by a mean of 2 seconds. The results also showed a significant
difference in the maximum braking force and maximum deceleration when reacting to the
sudden merging vehicle. Participants were observed to brake with a greater force and
decelerate more rapidly when using the ACC.
Standard deviations of lateral position were found to be significantly different during
the car following task. This was consistent with the findings of Cho et al. (2006) and
Rudin-Brown and Parker (2004). Participants driving with the ACC were observed to sway
more from the center of their lanes. This could be attributed to the process of engaging ACC
and setting the time-gap. However, more analysis is required for a better conclusion.
In the instance of the distraction, RTs increased during the ACC phase. However, this
could be a result of the compensation effect experienced by participants to account for not
being fully comfortable with the capabilities of the ACC system (Xiong, 2013) or a result of
added mental resources to initially monitor the ACC performance (Larsson, 2012). Although
22
the RTs increased slightly, the HRs during the distraction tasks were significantly higher
during the ACC phase, suggesting that participants were able to perform a secondary task
better with the ACC (participants attempted more hits on the distraction application when
using the ACC). This was consistent with the secondary task results obtained by
Rudin-Brown and Parker (2004).
During the work zone task, participants did not demonstrate a lack of situational
awareness. No participant was observed to blindly follow the lead vehicle, configured to
violate the speed limit in the work zone.
The most used ACC time-gap setting was found to be 3 seconds. This was
inconsistent with the results obtained by Cho et al. (2006) and Rosenfeld et al. (2015), Cho
et al. found the preferred time-gap to be 1.5 seconds, while Rosenfeld et al. concluded that
time-gap preferences vary between 0.8 seconds and 2 seconds and by personality traits.
However, the study by Rosenfeld et al. (2015) used results from drivers, in actual vehicles,
who regularly use ACC. Unfamiliarity with the ACC system could be attributed to the larger
time-gap setting selected by participants in this study.
In summary, it was found that ACC significantly affects driver performance, and
more specifically it affects the maximum speed, average distance headways, time to collision,
maximum braking force, standard deviation of lateral position, and maximum deceleration.
Drivers’ workload when driving with ACC was reduced, and therefore they performed
secondary tasks better, but this resulted in shorter time to collision and more abrupt braking.
To compensate for these effects, vehicles equipped with the ACC should also be equipped
with an active collision avoidance or emergency braking system.
23
Although the results of this research are promising, the sample size of 30 participants
is relatively small to draw any general inferences. Further studies need to be conducted to
provide a better understanding of the trust between the drivers and the ACC system, thus
decreasing ambiguities arising from compensation. Assessing the effects of ACC on
tired/fatigued participants or during complex environments can provide key insights into the
role of ACC in reducing physical and mental stress endured by the driver.
Lastly, comparing the conventional cruise control and ACC systems can provide
more details on how foot-resting positions and body posture alter reaction times. This can
also be used to quantify the compensation effect, especially to determine the amount of trust
drivers place in such systems.
Acknowledgements
The authors would like to thank the staff at the National Advanced Driving Simulator at the
University of Iowa for their extensive support in configuring the simulator and the ACC
system.
24
References
Adaptive Cruise Control System Overview. (2005). 5th Meeting of the U.S. Software System
Safety Working Group, Anaheim, CA.
Cho, J. H., Nam, H. K., & Lee, W. S. (2006). Driver Behavior with Adaptive Cruise Control.
International Journal of Automotive Technology, 7(5), 603-608.
Cooper, J. M., Ingebresten, H., & Strayer, D. L. (2014). Mental Workload of Common Voice-
Based Vehicle Interactions across Six Different Vehicle Systems. Washington, DC:
AAA Foundation for Traffic Safety.
Guvenc, B. A., & Kural, E. (2006). Adaptive Cruise Control Simulator: A Low-Cost,
Multiple-Driver-In-The-Loop Simulator. IEEE Control Systems Magazine, 26(3),
42-55.
Heritage Ford. (2016). What is Ford’s Adaptive Cruise Control with Stop and Go
Technology? Retrieved from http://ford.heritagevt.com/blog/what-is-fords-adaptive-
cruise-control-with-stop-and-go-technology/
Hyundai. (2017). Advanced Driver Assistance System. Retrieved from http://brand.
hyundai.com/en/brand/technology/adas.do.
IBM Corp. (2011). IBM SPSS Statistics for Windows, Version 20. Armonk, NY.
ISO 17488: Road Vehicles – Transport Information and Control Systems – Detection
Response Task (DRT) for Assessing Attentional Effects of Cognitive Load in
Driving. (2016). International Organization for Standardization, Switzerland.
Jenness, J. W., Lerner, N. D., Mazor, S., Osberg, J. S., & Tefft, B. C. (2008). Use of
Advanced In-Vehicle Technology by Young and Older Early Adopters. Survey
Results on Adaptive Cruise Control Systems. National Highway Traffic Safety
Administration, Washington, DC.
Kummetha, V. C. (2017). Analysis of the Effects of Adaptive Cruise Control (ACC) on Driver
Behavior and Awareness Using a Driving Simulator. The University of Kansas,
Lawrence, KS.
Larsson, A. (2012). Driver Usage and Understanding of Adaptive Cruise Control. Applied
Ergonomics, 43(3), 501-506.
25
Larsson, A., Kircher, K., Andersson Hultgren, J. (2014). Learning from experience:
Familiarity with ACC and Responding to a Cut-in Situation in Automated Driving.
Transportation Research Part F, 27, 229-237.
Ma, R., & Kaber, D. B. (2005). Situation Awareness and Workload in Driving While Using
Adaptive Cruise Control and a Cell Phone. International Journal of Industrial
Ergonomics, 35(10), 939-953.
Mardsen, G., McDonald, M., & Brackstone, M. (2000). Towards an Understanding of
Adaptive Cruise Control. Transportation Research Part C, 9(1), 33-51.
MiniSim User’s Guide. (2015). The National Advanced Driving Simulator, Document
version 19, The University of Iowa, IA.
Moeckli, F., Brown, T., Dow, B., Boyle, L. N., Schwarz, C., & Xiong, H. (2015). Evaluation
of Adaptive Cruise Control Interface Requirements on the National Advanced
Driving Simulator. Washington, DC: U.S. Department of Transportation, National
Highway Traffic Safety Administration.
Ohno, H. (2001). Analysis and Modeling of Human Driving Behaviors using Adaptive
Cruise Control. Applied Soft Computing, 1(3), 237-243.
Pauwelussen, J., & Feenstra, P. J. (2010). Driver Behavior Analysis during ACC Activation
and Deactivation in a Real Traffic Environment. IEEE Transactions on Intelligent
Transportation Systems, 11(2), 329-338.
Rosenfeld, A., Bareket, Z., Goldman, C. V., Kraus, S., LeBlanc, D. J., & Tsimoni, O. (2015).
Learning Drivers’ Behavior to Improve Adaptive Cruise Control. Journal of
Intelligent Transportation Systems, 19(1), 18-31.
Rudin-Brown, C. M., & Parker, H. A. (2004). Behavioral Adaptation to Adaptive Cruise
Control (ACC): Implications for Preventive Strategies. Transportation Research Part
F: Traffic Psychology and Behavior, 7(2), 59-76.
SAE J3016. (2016). Surface Vehicle Recommended Practice -Taxonomy and Definitions for
Terms Related to Driving Automation Systems for On-Road Motor Vehicles. SAE
International. 15-23.
Stojmenova, K., & Sodnik, J. (2015). Methods for assessment of cognitive workload in
driving tasks. ICIST 2015 5th International Conference on Information Society and
Technology, 229-234.
26
Strand, N., Nilsson, J., Karlsson, I.C.M., & Nilsson, L. (2011). Exploring End-user
Experiences: Self-perceived Notions on Use of Adaptive Cruise Control Systems.
IET Intelligent Transport Systems, 5(2), 134-140.
Strand, N., Nilsson, J., Karlsson, I.C.M., & Nilsson, L. (2014). Semi-automated Versus
Highly Automated Driving in Critical Situations Caused by Automation Failures.
Transportation Research Part F, 27, 218-228.
Strayer, D. L., Cooper J. M., Turrill, J., Coleman, J., & Ortiz, E. V. (2014). Measuring
Cognitive Distraction in the Automobile II: Assessing In-Vehicle Voice-Based
Interactive Technologies. Washington, DC: AAA Foundation for Traffic Safety.
Strayer, D. L., Cooper J. M., Turrill, J., Coleman, J., Medeiros-Ward, N., & Biondi, F. (2013).
Measuring Cognitive Distraction in the Automobile. Washington, DC: AAA
Foundation for Traffic Safety.
Vollrath, M., Schleicher, S., & Gelau, C. (2011). The Influence of Cruise Control and
Adaptive Cruise Control on Driving Behavior-A Driving Simulator Study. Accident
Analysis and Prevention, 43(3), 1134-1139.
De Winter, J. C. F., Happee, R., Martens, M. H., & Stanton, N. A. (2014). Effects of Adaptive
Cruise Control and Highly Automated Driving on Workload and Situation
Awareness: A Review of the Empirical Evidence. Transportation Research Part F:
Traffic Psychology and Behavior, 27(Part B), 196-217.
Xiong, H., & Boyle, L. N. (2013). Drivers’ Selected Settings for Adaptive Cruise Control
(ACC): Implications for Long-Term Use. Human Factors and Ergonomics Society,
Inc., 57(1), 1928-1932.
27
Figures and Tables
Figure 1. Methodological framework.
28
(a) (b)
Figure 2. KU driving simulator setup (a) and instrument panel interface (b).
Cruise control
ACC time-gap
Speed
29
Figure 3. Steering wheel configuration.
Time-gap increase
Time-gap decrease
Cruise on/off
Set
Resume
Cancel
30
Figure 4a. Desk drop event - driver view.
Figure 4b. Desk drop event - design view.
Desk
Desk
31
Figure 5. Areas of focus for video recordings.
Foot Position
Distraction
Interface
DRT Micro-Switch
32
Figure 6. Interface of the application used to simulate distraction.
33
Table 1. ANOVA results summarizing the differences within the age groups and gender
SS
df
MS
F
Sig.
Whole
Drive
App Hit
Attempts
Age group
14.067
2
7.033
0.701
0.506
0.055
Gender
0.300
1
0.300
0.030
0.864
0.001
Age group*Gender
35.000
2
17.500
1.744
0.196
0.127
Error
240.800
24
10.033
Collision
Count
Age group
0.200
2
0.100
0.375
0.691
0.030
Gender
0.033
1
0.033
0.125
0.727
0.005
Age group*Gender
0.067
2
0.033
0.125
0.883
0.010
Error
6.400
24
0.267
Maximum
Speed
(km/h)
Age group
22.671
2
11.336
0.154
0.858
0.013
Gender
54.477
1
54.477
0.740
0.398
0.030
Age group*Gender
46.089
2
23.045
0.313
0.734
0.025
Error
1766.328
24
73.597
Car
Following
Avg
Distance
Headway
(m)
Age group
22117.08
2
11058.54
6.032
0.008**
0.335
Gender
687.46
1
687.46
0.375
0.546
0.015
Age group*Gender
6906.64
2
3453.32
1.884
0.174
0.136
Error
43997.65
24
1833.24
Maximum
Speed
(km/h)
Age group
7.300
2
3.650
0.054
0.947
0.005
Gender
.014
1
0.014
0.000
0.989
0.000
Age group*Gender
85.994
2
42.997
0.641
0.536
0.051
Error
1610.167
24
67.090
SD of
lateral
position
(m)
Age group
0.004
2
0.002
0.307
0.738
0.025
Gender
0.007
1
0.007
1.122
0.300
0.045
Age group*Gender
0.001
2
0.000
0.072
0.931
0.006
Error
0.145
24
0.006
Crossing
Animal
Avg
Distance
Headway
(m)
Age group
5333.34
2
2666.67
1.240
0.310
0.106
Gender
807.07
1
807.07
0.375
0.547
0.018
Age group*Gender
3777.79
2
1888.89
0.878
0.430
0.077
Error
45178.53
21
2151.36
Maximum
Speed
(km/h)
Age group
9.585
2
4.792
0.170
0.845
0.014
Gender
7.792
1
7.792
0.276
0.604
0.011
Age group*Gender
127.170
2
63.585
2.255
0.127
0.158
Error
676.615
24
28.192
SD of
lateral
position
(m)
Age group
0.209
2
0.104
1.259
0.302
0.095
Gender
0.085
1
0.085
1.022
0.322
0.041
Age group*Gender
0.751
2
0.376
4.535
0.021*
0.274
Error
1.988
24
0.083
Time to
Collision
(s)
Age group
0.105
2
0.053
1.559
0.231
0.115
Gender
3.333E-6
1
3.333E-6
0.000
0.992
0.000
Age group*Gender
0.036
2
0.018
0.540
0.590
0.043
Error
0.809
24
0.034
*Sig. < 0.05, **Sig. < 0.01
34
Table 1. (Cont.)
SS
df
MS
F
Sig.
Desk
Drop
Avg Distance
Headway (m)
Age group
15017.84
2
7508.92
2.706
0.091
0.213
Gender
2.36
1
2.356
0.001
0.977
0.000
Age group*Gender
4079.37
2
2039.68
0.735
0.492
0.068
Error
55508.28
20
2775.41
Time to
Collision (s)
Age group
10.019
2
5.009
2.046
0.158
0.185
Gender
2.222
1
2.222
0.908
0.353
0.048
Age group*Gender
3.301
2
1.651
0.674
0.522
0.070
Error
44.060
18
2.448
SD of lateral
position (m)
Age group
0.069
2
0.034
0.964
0.396
0.077
Gender
0.002
1
0.002
0.054
0.818
0.002
Age group*Gender
0.082
2
0.041
1.148
0.335
0.091
Error
0.817
23
0.036
Sudden
Merging
Vehicle
Time to
Collision (s)
Age group
20.886
2
10.443
2.484
0.107
0.184
Gender
7.745
1
7.745
1.842
0.188
0.077
Age group*Gender
0.949
2
0.474
0.113
0.894
0.010
Error
92.499
22
4.205
Max Braking
Force (N)
Age group
6435.77
2
3217.89
0.827
0.453
0.080
Gender
900.59
1
900.59
0.231
0.636
0.012
Age group*Gender
8447.08
2
4223.54
1.085
0.358
0.102
Error
73967.22
19
3893.01
Max
Deceleration
(m/sec2)
Age group
4.337
2
2.168
0.881
0.431
0.085
Gender
1.265
1
1.265
0.514
0.482
0.026
Age group*Gender
3.892
2
1.946
0.791
0.468
0.077
Error
46.747
19
2.460
Move
Over
Task
Total
Observed
Age group
1.400
2
0.700
2.800
0.081
0.189
Gender
0.033
1
0.033
0.133
0.718
0.006
Age group*Gender
1.267
2
0.633
2.533
0.100
0.174
Error
6.000
24
0.250
*Sig. < 0.05, **Sig. < 0.01
35
Table 2. ANOVA results summarizing the differences in mean RTs and HRs within the age
groups and gender
SS
df
MS
F
Sig.
Without
Events or
Tasks
RT
(s)
Age group
0.013
2
0.006
0.535
0.593
0.046
Gender
4.443E-5
1
4.443E-5
0.004
0.951
0.000
Age group*Gender
0.006
2
0.003
0.242
0.787
0.022
Error
0.258
22
0.012
HR
(%)
Age group
136.937
2
68.469
0.875
0.431
0.074
Gender
15.726
1
15.726
0.201
0.658
0.009
Age group*Gender
120.200
2
60.100
0.768
0.476
0.065
Error
1720.827
22
78.219
Car
Following
RT
(s)
Age group
0.007
2
0.004
0.219
0.805
0.020
Gender
0.001
1
0.001
0.073
0.789
0.003
Age group*Gender
0.010
2
0.005
0.309
0.738
0.027
Error
0.358
22
0.016
HR
(%)
Age group
88.385
2
44.193
0.262
0.772
0.023
Gender
12.363
1
12.363
0.073
0.789
0.003
Age group*Gender
52.345
2
26.172
0.155
0.857
0.014
Error
3708.892
22
168.586
Attempting
a
Secondary
Task while
Driving
RT
(s)
Age group
0.154
2
0.077
0.503
0.612
0.044
Gender
0.222
1
0.222
1.452
0.241
0.062
Age group*Gender
0.010
2
0.005
0.032
0.968
0.003
Error
3.365
22
0.153
HR
(%)
Age group
53.435
2
26.718
0.029
0.971
0.003
Gender
1968.496
1
1968.496
2.138
0.158
0.089
Age group*Gender
586.021
2
293.011
0.318
0.731
0.028
Error
20256.933
22
920.770
Work Zone
RT
(s)
Age group
0.030
2
0.015
0.567
0.575
0.049
Gender
0.009
1
0.009
0.348
0.561
0.016
Age group*Gender
0.069
2
0.035
1.294
0.294
0.105
Error
0.589
22
0.027
HR
(%)
Age group
147.068
2
73.534
0.444
0.647
0.039
Gender
200.496
1
200.496
1.211
0.283
0.052
Age group*Gender
48.359
2
24.180
0.146
0.865
0.013
Error
3642.101
22
165.550
*Sig. < 0.05, **Sig. < 0.01
36
Figure 7. Mean percentage usage by time-gap for all 30 participants (Error bars represent
the standard deviation of mean usage)
0%
20%
40%
60%
80%
100%
Mean percentage usage
1.2-sec 2.0-sec 3.0-sec
Percentage Usage by Time-Gap
SD
37
Figure 8. Average distance headways with respect to the events (Error bars represent the
Variance).
Figure 9. Maximum speed with respect to the events (Error bars represent the Variance).
No ACC
ACC
No ACC
ACC
Mean
Median
Variance
Mean
Median
Variance
38
Figure 10. Time to collision by event (Error bars represent the Variance).
Figure 11. Standard deviation of lateral position by event (Error bars represent the
Variance).
Mean
No ACC
ACC
Variance
Mean
Median
No ACC
ACC
Variance
Median
39
Table 3. Paired t-test summary of the data variables collected by event/task.
Data Variable
Phase
Mean
St dev
t-stat
df
% diff
Cohen’s
d
Sig.
Whole
Drive
App Hit Attempts
No ACC
12.8
4.78
-4.906
29
-21.9%
0.90
<0.001**
ACC
15.6
5.38
Collision Count
No ACC
1.3
0.45
1.140
29
7.7%
0.21
0.264
ACC
1.2
0.38
Maximum Speed
(km/h)
No ACC
130.2
6.47
5.899
29
6.7%
1.08
<0.001**
ACC
121.5
6.00
Car
Following
Avg Distance
Headway (m)
No ACC
108.9
48.9
-2.425
29
-20.5%
0.44
0.022*
ACC
131.2
42.1
Maximum Speed
(km/h)
No ACC
124.4
6.10
4.593
29
5.2%
0.84
<0.001**
ACC
118.0
4.55
SD of lateral
position (m)
No ACC
0.30
0.08
-2.057
29
-9.3%
0.38
0.049*
ACC
0.32
0.08
Crossing
Animal
Avg Distance
Headway (m)
No ACC
102.0
46.6
-1.265
26
-11.0%
0.24
0.217
ACC
113.2
42.6
Time to Collision
(s)
No ACC
0.37
0.18
3.356
29
29.7%
0.61
0.002**
ACC
0.26
0.19
Maximum Speed
(km/h)
No ACC
118.1
5.81
4.710
29
3.8%
0.86
<0.001**
ACC
113.6
3.06
SD of lateral
position (m)
No ACC
0.27
0.24
0.189
29
3.4%
0.04
0.852
ACC
0.26
0.19
Desk
Drop
Avg Distance
Headway (m)
No ACC
94.6
68.1
-2.578
25
-29.6%
0.51
0.016*
ACC
122.6
50.5
Time to Collision
(s)
No ACC
2.09
1.40
-0.851
23
-13.4%
0.17
0.404
ACC
2.37
1.24
SD of lateral
position (m)
No ACC
0.37
0.20
-1.003
28
-8.9%
0.19
0.324
ACC
0.41
0.22
Sudden
Merging
Vehicle
Time to Collision
(s)
No ACC
3.88
2.65
3.373
27
35.1%
0.64
0.002**
ACC
2.52
1.77
Max Braking
Force (N)
No ACC
127.7
70.1
-2.942
24
-28.6%
0.59
0.007**
ACC
164.1
92.1
Max Deceleration
(m/sec2)
No ACC
-3.96
1.21
3.029
24
-23.5%
0.61
0.006**
ACC
-4.89
1.49
Move
Over
Task
Total Observed
No ACC
1.67
0.76
1.000
29
6.0%
0.18
0.326
ACC
1.57
0.82
*Sig. < 0.05, **Sig. < 0.01, % diff = (No ACC-ACC/No ACC) * 100%
40
Table 4. Paired t-test summary of the data collected using the DRT.
Data
Variable
Phase
Mean
St Dev
t-stat
df
% diff
Cohen’s
d
Sig.
Without
Events or
Tasks
RT (s)
No ACC
0.611
0.142
0.137
27
0.5%
0.03
0.892
ACC
0.608
0.125
HR (%)
No ACC
84.6
12.2
-0.935
27
-1.8%
0.18
0.358
ACC
86.1
13.2
Car
Following
RT (s)
No ACC
0.528
0.151
-1.452
27
-6.1%
0.27
0.158
ACC
0.560
0.129
HR (%)
No ACC
93.3
10.4
2.107
27
5.1%
0.40
0.045*
ACC
88.5
14.8
Attempting
a
Secondary
Task while
Driving
RT (s)
No ACC
0.775
0.306
-0.845
27
-7.7%
0.16
0.406
ACC
0.835
0.270
HR (%)
No ACC
55.3
27.8
-2.660
27
-26.4%
0.50
0.013*
ACC
69.9
25.1
Work Zone
RT (s)
No ACC
0.641
0.184
-0.092
27
-0.3%
0.02
0.927
ACC
0.643
0.213
HR (%)
No ACC
83.6
19.1
-1.391
27
-3.8%
0.26
0.176
ACC
86.8
19.7
*Sig. < 0.05, **Sig. < 0.01, % diff = (No ACC-ACC/No ACC) * 100%
