Accident Analysis and Prevention 45 (2012) 759–765

Contents lists available at SciVerse ScienceDirect

Accident Analysis and Prevention

journal homepage: www.elsevier.com/locate/aap

The effects of on-street parking and road environment visual complexity on

travel speed and reaction time
Jessica Edquist ∗ , Christina M. Rudin-Brown, Michael G. Lenné
Human Factors group, Monash University Accident Research Centre, Australia

a r t i c l e i n f o a b s t r a c t

Article history: On-street parking is associated with elevated crash risk. It is not known how drivers’ mental workload
Received 15 July 2011 and behaviour in the presence of on-street parking contributes to, or fails to reduce, this increased crash
Received in revised form 14 October 2011 risk. On-street parking tends to co-exist with visually complex streetscapes that may affect workload and
Accepted 17 October 2011
crash risk in their own right. The present paper reports results from a driving simulator study examining
the effects of on-street parking and road environment visual complexity on driver behaviour and surro-
Keywords:
gate measures of crash risk. Twenty-nine participants drove a simulated urban commercial and arterial
Driving simulation
route. Compared to sections with no parking bays or empty parking bays, in the presence of occupied
Driving task difficulty
Risk compensation
parking bays drivers lowered their speed and shifted their lateral position towards roadway centre to
Road safety compensate for the higher mental workload they reported experiencing. However, this compensation
was not sufficient to reduce drivers’ reaction time on a safety-relevant peripheral detection task or to
an unexpected pedestrian hazard. Compared to the urban road environments, the less visually complex
arterial road environment was associated with speeds that were closer to the posted limit, lower speed
variability and lower workload ratings. These results support theoretical positions that proffer workload
as a mediating variable of speed choice. However, drivers in this study did not modify their speed suf-
ficiently to maintain safe hazard response times in complex environments with on-street parking. This
inadequate speed compensation is likely to affect real world crash risk.
© 2011 Elsevier Ltd. All rights reserved.

1. Introduction
On-street parking is frequently found in urban retail road envi-
ronments, as it is convenient for shoppers and does not require
the space used by a dedicated parking lot. Unfortunately, on-street
parking is also associated with increased crash risk compared to
roads of the same category without on-street parking (Greibe,
2003; Pande and Abdel-Aty, 2009; Roberts et al., 1995). It is impor-
tant to understand the reasons behind this elevated crash risk in
order to design appropriate countermeasures.
At a practical level, one potential reason for increased crash risk
is narrowed road width. Parked vehicles leave less space for travel-
ling vehicles, forcing them to drive closer to vehicles in the next lane
(which may be travelling in the opposite direction). Similarly, nar-
row roads are associated with shifts in lateral position closer to the
centreline (Lewis-Evans and Charlton, 2006), as well as higher crash
rates than standard roads (Greibe, 2003). Greibe’s (2003) modelling
research found that both parking and road width were significant
predictors of crash rates on road links, and that the contribution of
∗ Corresponding author at: Building 70, Monash University, Victoria, 3800,
Australia. Tel.: +61 3 9905 5815; fax: +61 3 9905 4363.
E-mail address: Jessica.edquist@monash.edu (J. Edquist).
parking to increased crash rates was larger than that of road width.
This implies that factors other than road width must underlie the
higher crash risk when on-street parking is present.
In complex urban environments, drivers must monitor move-
ments of both pedestrians and vehicles. Parked cars may obstruct
the view of the road ahead, making it more difficult to see crossing
pedestrians. Parked cars may also suddenly become moving cars,
and rejoin the traffic stream. It has been found that the number
of (moving) vehicles in a scene negatively affects situation aware-
ness and hazard avoidance (Gugerty, 1997), and parked vehicles
may contribute to this effect, as they require monitoring to deter-
mine whether or not they are moving. The presence of parked cars
therefore increases the uncertainty, mental load and potential risk
associated with the road environment.
It would be surprising if drivers were oblivious to these changes
and did not adapt their behaviour accordingly. In fact, traffic obser-
vations in residential areas have found that high parking densities
correlate with slower speeds (Daisa and Peers, 1997). Psychological
research offers further insight into this adaptation. Chinn and Elliot
(2002) asked drivers questions about sketches of road scenes with
various features, and found that on-street parking increased ratings
of tension and decreased the perceived safe speed. This is consis-
tent with theories of driver behaviour that maintain drivers reduce
their speed when they perceive a road environment as being risky,

0001-4575/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aap.2011.10.001
760 J. Edquist et al. / Accident Analysis and Prevention 45 (2012) 759–765
for example Wilde’s (1982) theory of Risk Homeostasis, Summala’s
(1988) Zero Risk theory and Fuller’s (2005) and Fuller et al.’s (2008)
Task-Capability Interface model.
Besides on-street parking, the roadside environment may also
contribute to driver workload and perceived risk, and hence affect
travel speed. Chinn and Elliot (2002) found that the height and type
(office vs. commercial) of buildings adjacent to the road lowered
perceived safe speeds, although they did not change drivers’ ratings
of tension. Factors such as building setback (Martens et al., 1997;
van der Horst and de Ridder, 2007), building height (Allsop, 1995;
Elliot et al., 2003), and visual complexity (Horberry et al., 2006)
have all been found to affect speed on urban roads, presumably by
increasing the visual/cognitive workload of the driver. It is possible
that crash risk may also be affected by these roadside factors: Harms
(1986) found that, compared to rural areas, driving on highways
through village areas was associated with slower speeds and higher
cognitive load, which was correlated with the number of reported
accidents.
Driver workload, the mediating variable between the difficulty
of the driving task and performance, reflects the driver’s current
ability to meet the demands of the task (De Waard, 1996). A driver’s
experience of increased workload would be expected to medi-
ate compensation to a more difficult driving environment. Greater
workload is reflected in lower vehicle speed and longer headways
to a lead vehicle, as drivers attempt to make the driving task less dif-
ficult. Variability in speed and headway has been found to increase
with workload (Parkes et al., 2007), as drivers are less able to con-
centrate on longitudinal vehicle control. Interestingly, variability in
lateral position has been found to both increase and decrease with
workload, depending on the source of the workload (De Waard,
1996; Engstrom et al., 2005). These variations may affect crash risk
if they result in lane excursions or very short headways. In addition,
if drivers’ adaptation to increased workload/task demand is inad-
equate (i.e. if they do not slow down sufficiently to compensate),
their reaction time to hazards will be slower, leaving less time to
avoid a collision.
The present study used a driving simulator to examine differ-
ences in driver behaviour which may help to explain crash risk
differences between different environments. In each road environ-
ment studied, the road geometry was held constant; however we
varied the complexity of the road environment, and the presence of
on-street parking. It was hypothesised that both road complexity
and on-street parking would lead to increased subjective workload,
decreased vehicle speed (and possibly increased speed variability),
increased headway to a slower-moving lead vehicle (and pos-
sibly increased headway variability), changes in lateral position
(and possibly increased lane position variability), increased reac-
tion time to safety-relevant peripheral detection task (PDT) targets
and more missed targets, and impaired responses to an unexpected
pedestrian event.
2. Method
2.1. Design
To examine the effects of road environment on driver behaviour,
we compared a low complexity environment (the Arterial No Park-
ing condition) with three conditions involving a combination of
high complexity and parking factors in a single-factor, repeated
measures ANOVA with four levels.
Within the higher complexity roadside environment, we exam-
ined the effects of on-street parking with and without the presence
of traffic ahead. This analysis included three levels of parking (None,
Empty bays, Full bays) and two levels of traffic (lead vehicle present,
lead vehicle absent – both levels contained oncoming traffic).
Dependent variables included subjective workload, vehicle con-
trol (speed/headway, lateral position and variability in each of
these), performance on a PDT, and response to a critical situation
(in which the driver would collide with a simulated pedestrian
unless they took preventive action). PDT performance can be con-
strued as an objective measure of both hazard perception ability
and situation awareness, two essential driver skills. Vehicle control
is also directly related to crash risk: a vehicle will crash if its posi-
tion coincides with another object either laterally or longitudinally,
therefore the driver must maintain lateral control (lane position)
and longitudinal control (speed, headway).
2.2. Participants
Twenty-nine drivers (15 male) aged between 20 and 53 years
(mean = 28.0 years, SD = 7.6 years) were recruited from the Univer-
sity community via an online advertisement. The average number
of years participants had been driving was 9.8 (SD = 8.0); all par-
ticipants were regular drivers with at least one year of licensed
(unsupervised) driving experience. Participants were compensated
$30 for their time. The study was approved by the University ethics
committee.
2.3. Equipment
The experiment used an EF-X by Eca-Faros driving simulator
with modified software for research purposes. This consists of a
driver’s seat with dashboard, steering wheel, accelerator and brake
pedals, handbrake and gear shift. The simulated environment is
displayed on three screens providing 120◦ field of view including
rear view mirror. Data is collected at 30 Hz.
2.4. Scenarios
The simulated road had two lanes in each direction with a double
white line (no median strip) in between. There were sidewalks with
pedestrians on both sides and occasional oncoming traffic. Lane and
parking bay dimensions were based on information supplied by the
local road authority. Fig. 1 shows examples of the simulated road
environment.
Scenarios were designed to replicate the four road environments
(parking conditions) of interest. The four conditions comprised: the
Arterial No Parking condition, in which buildings were set far back
from the road and there were no parking bays (see Fig. 1, top panel);
the Urban No Parking condition, a shopping-strip style road with
shopfronts on low multi-storey buildings set close to the road on
both sides; the Urban Empty Parking condition, identical to Urban
No Parking but with the addition of empty parking bays marked
in the kerbside lanes on each side of the road; and the Urban Full
Parking condition (see Fig. 1, lower panel), identical to Urban Empty
Parking but with 90% of the parking bays occupied by vehicles.
These four conditions were each presented (as a 1 km-long block)
six times over the course of three experimental drives, in order
to minimise fatigue and order effects. Each drive contained eight
blocks: two Arterial No Parking blocks, and one block with car fol-
lowing and one block without for each of Urban No Parking, Urban
Empty Parking, and Urban Full Parking. The blocks were in a differ-
ent (randomised) order for each drive. In addition to these blocks,
each drive started and ended with a section similar to the Arterial
No Parking condition that was not analysed.
Half of the urban blocks (i.e. Urban No Parking, Urban Empty
Parking and Urban Full Parking conditions) contained a car-
following task, in which a lead vehicle turned into the road ahead
of the participant’s vehicle, travelling initially at the same speed as
the participant. After 160 m, this vehicle decelerated to 70% of the
participant’s speed (at that point) and remained at that speed for
 J. Edquist et al. / Accident Analysis and Prevention 45 (2012) 759–765
Fig. 1. Screenshots of (a) Arterial No Parking and (b) Urban Full Parking conditions.
the rest of the block. Additionally, in one non-car-following block
for each of these three conditions, a pedestrian suddenly turned and
walked onto the road in front of the subject vehicle. The pedestrian
turned and stepped into the road when the driver was 3.64 s away
from the potential collision site (based on the drivers’ speed, not a
set distance), thus allowing plenty of time to avoid the collision if
the driver responded immediately.
2.5. Procedure
After completing a demographic and driving experience
questionnaire, participants performed a short practice drive to
familiarise them with the control dynamics of the simulated
vehicle. Participants were instructed to practise accelerating and
braking gently, and to practise driving at a consistent speed of
60 km/h. This drive was repeated with the addition of the PDT; this
task involved the participant activating the left or right turn indica-
tor in response to a pedestrian icon which appeared at the far side
of the left or right outer screens. The icon appeared at randomised
intervals of 200, 300, 400, 500 or 600 m (12–36 s at 60 km/h), and
disappeared after 100 m (6 s) if the driver failed to respond. Fol-
lowing this practice, participants completed the three main drives
(while also performing the PDT). The whole experiment took just
over an hour.
In the main drives, participants were instructed to drive as they
normally would, but remain in the centre lane, and try to main-
tain a consistent speed of around 60 km/h (the posted speed limit).
The order of presentation of the three drives was counterbalanced
761
Mean sp
peed Ma
ax speed
70.0
65.0
Speed (km/h)
60.0
55.0
50.0
45.0
Arterial
r No Urban No U
Urban Emptty Urban F
Full
Parrking Park
P ing Parking Parking
Fig. 2. Mean and maximum speed (for non-following blocks only) by road environ-
ment and parking condition (bars represent 95% CIs).
across participants. At eight points (once each after blocks of each
Urban condition with and without car following, and twice after
Arterial No Parking blocks), participants stopped at a red traffic
light and completed a Mental Workload Questionnaire for the block
they had just driven though. The Mental Workload Questionnaire
was based on the NASA-TLX (Hart and Staveland, 1988); however
the question about physical demand was replaced with a question
about safety. Participants rated mental demand, time pressure, suc-
cess, effort, frustration, and safety (reverse scored) out of ten, and
these scores were summed to give an overall measure of work-
load (results for the two Arterial No Parking blocks were averaged,
after confirming that they were not significantly different from one
another).
2.6. Data analyses
After checking that there were no significant differences
between blocks of the same condition due to order effects, data
from different blocks of the same condition were averaged. Each of
the dependent variables was analysed using a repeated-measures
analysis of variance (RMANOVA) with the within-subjects factor of
road environment (four levels: Arterial No Parking, Urban No Park-
ing, Urban Empty Parking, Urban Full Parking). For the dependent
variables of workload rating, lateral position and peripheral detec-
tion, an additional within-subjects factor was the presence/absence
of a lead vehicle (these analyses do not include data from the
Arterial No Parking condition). For blocks that included the car-
following task, data from the first 400 m and the last 200 m were
discarded to avoid the potentially confounding effects of the decel-
erating lead vehicle (at the start of the block) and red traffic signals
(at the start and end of the block). For blocks without this task, the
first and last 200 m were discarded.
3. Results
3.1. Speed
As participants’ speed was restricted by the lead vehicle in
the car-following blocks, the analysis of speed includes only data
from blocks without a lead vehicle present (4 × 1 RMANOVA for
each dependent variable). The main effect of road environment
was significant for mean speed, F(3,84) = 75.405, p < .001, 2 = .729;
maximum speed, F(3,84) = 24.404, p < .001, 2 = .466; and speed
variability, F(3,84) = 181.778, p < .001, 2 = .867. Mean and maxi-
mum speed decreased with the complexity of the environment
(Fig. 2): the Arterial No Parking and Urban Full conditions were
significantly different from the other two conditions, while the
Urban No Parking and Empty bay conditions were not significantly
different from each other.
762 J. Edquist et al. / Accident Analysis and Prevention 45 (2012) 759–765

No lead vehicle Slow lead vehicle

.200

right
.100

(metres)
.000

-.100

left
-.200

-.300
Arterial No Urban No Urban Empty Urban Full
Parking Parking Parking Parking
Fig. 3. Standard deviation of speed over several 1 km blocks, by road environ- Fig. 4. Mean lateral position by road environment, parking condition and presence
ment and parking condition (bars represent 95% CIs) *Significant difference between of lead vehicle; 0 represents the centre of the driver’s lane, with negative values
‘Urban No Parking’ and ‘Urban Full Parking’. representing positions left of the centre (bars represent 95% CIs).

Speed variability increased with environmental complexity

(Fig. 3): the Arterial No Parking condition was associated with
significantly less variable speeds than all other conditions, while
the Urban Full Parking condition was associated with significantly
more variable speeds than the Urban No Parking condition (the
Urban Empty Parking condition was not significantly different from
either the Urban Full Parking or Urban No Parking conditions).
3.2. Headway to lead vehicle
A 3 × 1 RMANOVA for each dependent variable revealed that
there were no significant effects of road environment on mean
time headway (F < 1), minimum time headway (F < 1), or time head-
way variability (F < 2). Mean overall headway was 2.7 s, minimum
(averaged across participants) was 2.1 s, and variability was 0.37 s.
3.3. Lateral position
For mean lateral position data, the blocks without a lead vehicle
were analysed in a repeated measures ANOVA with road environ-
ment as a 4-level, within-subjects factor. The main effect of road
environment was significant, F(3,84) = 42.003, p < .001, 2 = .600.
Pairwise comparisons showed that participants drove closest to the
left (kerbside) edge in the Arterial No Parking condition and fur-
thest from the kerb in the Urban Full Parking condition, with the
Urban No Parking and Empty conditions being in between these
two extremes (Fig. 4). When the blocks with lead vehicles were
included (and the Arterial No Parking condition excluded) in a 3 × 2
RMANOVA, parking condition was significant, F(2,56) = 56.346,
p < .001, 2 = .668. Presence of a lead vehicle was also significant,
targets (less than 2% of the overall sample) were deleted from the
subsequent analyses of reaction time.
Reaction time to peripheral targets during the blocks without
any lead vehicles was analysed in a repeated measures ANOVA
with road environment as a 4-level, within-subjects factor. The
main effect of road environment was significant, F(3,84) = 45.917,
p < .001, 2 = .621. Pairwise comparisons showed that PDT reac-
tion time during the Urban Full Parking condition was significantly
slower than during all other conditions, which did not differ from
each other (Fig. 5).
A second RMANOVA without the data from the Arterial No
Parking condition included parking as a 3-level, within-subjects
factor, and lead vehicle as a 2-level, within-subjects factor. The
main effect of parking condition was significant (F(2,56) = 58.273,
p < .001, 2 = .675). The main effect of the presence of a lead vehicle
was borderline significant, F(1,28) = 4.031, p = .054, 2 = .126. There
was no interaction between parking condition and presence of a
lead vehicle in terms of PDT reaction time.
3.5. Response to unexpected pedestrians
In one block each of the Urban No Parking, Empty Parking and
Full Parking conditions (always with no lead vehicle present), a
pedestrian suddenly turned from the footpath and walked onto the
road in front of the subject vehicle.
2.00
No vehicle Slow lead vehicle
Response time (seconds)

F(1,28) = 7.057, p = .013, 2 = .201; participants drove further away

from the kerb in the presence of a lead vehicle. There was no sig-
1.50
nificant interaction between lead vehicle and parking.
There was no significant main effect of parking condition on
lane position variability. Lateral position variability was signifi-
cantly lower in the presence of a lead vehicle: F(1,28) = 35.758,
p < .001, 2 = .561; the standard deviation of lane position (SDLP) 1.00
was 0.144 m (SE = 0.008) without the lead vehicle present and
0.123 m (SE = .007) with the lead vehicle present.

0.50
3.4. Peripheral detection task
Arterial No Urban No Urban Empty Urban Full
Parking Parking Parking Parking
There were only 34 missed PDT targets of a total of 1740 target
presentations across all participants. Thirty-two of these missed Fig. 5. Response time to peripheral icon by road environment, parking condition,
targets (94%) occurred in the Urban Full Parking condition. Missed and presence of lead vehicle (bars represent 95% CIs).
 J. Edquist et al. / Accident Analysis and Prevention 45 (2012) 759–765
Table 1
Response to unexpected pedestrian event.
Measure
Time to accelerator release
Time to brake
Maximum brake pressure
Minimum distance
Minimum time-to-collision (TTC)
Number of collisions
Dependent measures of participants’ responses to the unex-
pected pedestrian event included: time to accelerator release, time
to brake, maximum brake pressure, minimum time to collision,
minimum distance to the pedestrian, and whether or not the par-
ticipant’s vehicle collided with the pedestrian. The first five of these
measures were included in a repeated measures multivariate anal-
ysis of variance (MANOVA). The main effect of parking condition
was significant, F(10,106) = 6.063, p < .001, 2 = 364 (using Pillai’s
trace); and univariate analyses showed that it was significant for
every measure. Table 1 presents the results of the analyses. In every
case, there was a significant difference between the Urban Full Park-
ing condition and the Urban No Parking/Empty conditions, which
do not differ from each other.
There was only one collision during the Urban Empty Parking
block, after which the driver said that he had seen the pedestrian
(and braked) but expected that the pedestrian would stop, so had
not come to a full stop. There were no collisions during the Urban
No Parking block, while 11 of the 29 drivers collided with the pedes-
trian during the Urban Full Parking block.
Time to brake was almost twice as long in the Urban Full Park-
ing condition (2.00 s) compared to the Urban No Parking condition
(1.12 s). Assuming a friction coefficient of 0.7 during braking, ini-
tial travel speed in the Urban Full Parking condition would need to
be 47 km/h (instead of 60 km/h) to achieve the same stopping dis-
tance as in the Urban No Parking condition (based on the equation
of motion: velocity squared = initial velocity squared plus 2 times
acceleration times distance, solving for distance, where accelera-
tion is gravity times friction coefficient).
3.6. Mental workload
Blocks without any lead vehicle present were analysed in a 4 × 1
RMANOVA with road environment as a within-subjects factor. The
main effect of road environment was significant, F(3,84) = 11.948,
p < .001, 2 = .299. Pairwise comparisons showed that the Arterial
No Parking condition was associated with significantly lower work-
load ratings than all other conditions, and the Urban Full Parking
condition received significantly higher workload ratings than all
other conditions, while Urban No Parking and Urban Empty Parking
workload ratings were not significantly different from one another.
30
Workload sum (out of 60)
No vehicle Slow lead vehicle
25
20
15
10
5 appropriate countermeasure.
0
Arterial No Urban No Urban Empty
Parking Parking Parking
Fig. 6. Workload ratings by road environment and parking condition (bars represent
95% CIs).
763
F(2,56) P 2 Difference
39.491 <.001 .585 Full > No park/Empty
32.515 <.001 .537 Full > No park/Empty
11.691 <.001 .295 Full > No park/Empty
20.401 <.001 .421 Full < No park/Empty
20.056 <.001 .417 Full < No park/Empty
2 (2) = 21.46 <.001 Full > No park/Empty
A second RMANOVA without the Arterial No Parking condition
involved parking as a 3-level within-subjects factor, and the pres-
ence of a lead vehicle as a 2 level within-subjects factor. The main
effect of parking was significant, F(2,52) = 17.612, p < .001, 2 = .404.
The Urban No Parking and Empty conditions were not significantly
different from one another, but the Full condition was rated by par-
ticipants as significantly more demanding in terms of workload
(Fig. 6). Neither presence of a lead vehicle, nor the interaction of
lead vehicle and parking, was significant.
4. Discussion
As expected, vehicles parked on-street in a complex, urban
environment influenced driver behaviour in several ways: drivers
reported experiencing higher workload, reduced their vehicle
speed, positioned the vehicle further from the kerb towards
oncoming traffic, reacted more slowly to a peripheral target, and
responded more slowly to an unexpected pedestrian when driving
in the Urban Full Parking condition compared to the Urban Empty
condition. Although there was evidence of some compensation by
drivers in response to the increased workload in terms of reduced
speed and a change in lane position further from the parked vehi-
cles, these changes were not sufficient to protect drivers from an
increase in crash risk. The combination of behaviour changes seen
in the driving simulator is therefore likely to translate to higher
crash rates in real world, urban areas where vehicles are parked
on-street.
The pattern of compensation to on-street parking is interesting:
of those variables which were significantly affected by parking, sub-
jective workload and speed variability increased the least between
the Urban No Parking and Full Parking conditions, and by about the
same amount as speed itself reduced. These results support the con-
cept of mental workload as an important mediator of speed control
(e.g. Fuller, 2005; Hoyos, 1988). Reaction time on the PDT increased
more than did workload, with response time to the unexpected
pedestrian almost doubling in the Urban Full Parking condition
compared to the Urban No Parking condition. There appears to be
a separation between the effects of on-street parking on perceived
workload and speed control, and its effects on drivers’ perception-
response time to hazards. This has negative implications for road
safety. Perceived workload could be more dependent on drivers’
ability to maintain lateral control, rather than on their ability to
maintain a safe stopping distance. This is supported by findings
such as Brooks et al. (2011) simulator study of curve navigation
in fog. Although they did not collect workload data, they found
that participants did not slow down until visibility was massively
reduced, well below the level at which they could stop in time
to avoid hitting a pedestrian hidden in the fog. Further research
is required to confirm this suggestion, and to determine the most
It should be noted that the effects of parked vehicles on reac-
tion time to the PDT and to the unexpected pedestrian were found
Urban Full
even though drivers had slowed down. If response times had been
Parking
adjusted to account for the increased time available when partic-
ipants drove more slowly (as was done by Trick et al., 2010), the
results would have been even more marked.
764 J. Edquist et al. / Accident Analysis and Prevention 45 (2012) 759–765
The slower free speeds chosen by drivers in the Urban Full Park-
ing condition were not reflected in longer time headways when
car-following, perhaps because the time headways drivers chose
in this study were very conservative (an average of 2.7 s across all
conditions; the standard guideline is 2 s). Lateral position shifted
towards the centre of the lane, which probably reflects the use of
the lead vehicle’s position (centred in the lane) as a guide. This may
also account for the finding that lane position variability decreased
when a lead vehicle was present. Alternatively, this finding may
reflect the amount of attention drivers had available to concentrate
on the lateral control portion of the driving task. In the presence of
a lead vehicle, longitudinal control could be maintained by moni-
toring the closeness of the lead vehicle (while looking ahead) rather
than the speedometer (requiring a downward glance); in addition,
the lower speed imposed by the lead vehicle would have made the
task of maintaining a consistent lateral position easier.
Attentional allocation can also explain the variability of lateral
and longitudinal control in the Urban Full Parking condition. In the
presence of parked cars, which essentially narrowed the available
road width, participants may have paid more attention to main-
taining their lateral position; thus SDLP remained low compared to
the Urban No Parking condition, while speed variability increased.
Road environment and complexity appear to influence driving
performance. A number of performance measures showed differ-
ences between the Arterial No Parking condition (where buildings
were few and set far back from the road, and the roadside envi-
ronment was simple) and the Urban No Parking/Empty Parking
condition (where the roadside environment was visually complex,
comprising continuous shopfronts at a small distance from the road
edge). In the latter conditions, subjective workloads were higher,
speed was lower and more variable, and lateral position was fur-
ther away from the road edge. These are the same effects as those
found when comparing the Urban No Parking/Empty Parking to the
Urban Full Parking condition, and imply that, in addition to char-
acteristics of the road itself, the visual complexity of the roadside
environment is an important contributor to driver workload and
performance. Interestingly, PDT performance remained constant
between the Arterial No Parking and the Urban No Parking/Empty
Parking conditions, but was impaired in the Urban Full Parking con-
dition. Previous research has shown that PDT response times are
affected by environmental complexity in terms of traffic (Nunes
and Recarte, 2005) and level of visual information (Stinchcombe
and Gagnon, 2010); further research is required to clarify the effect
of complex backgrounds on peripheral hazard detection.
It is noteworthy that there were no significant differences in
any of the dependent measures between the Urban No Parking and
Empty Parking conditions. These conditions were identical except
for the presence of road markings implying that vehicles were per-
mitted to park on the street. Drivers in the present study were never
directed to park, so it appears that they interpreted these markings
as being irrelevant in the absence of any vehicles actually parked
there. This is consistent with previous research, which has found
little effect on speed of similar markings such as bus lanes and cycle
lanes unless they are occupied (Elliot et al., 2003).
4.1. Limitations and future research
While the present study included both male and female drivers,
of a range of different ages and experience levels, all were car
drivers who primarily drove as commuters. It is possible that differ-
ent risk factors will affect certain driver groups to different extents.
For example, Tarko (2009) found that pavement width had more
effect on speed for truck drivers than car drivers. Speed selec-
tion may be different for drivers such as a taxicab driver, a young
driver who drives for excitement, or a cautious older driver who
self-restricts, to give a few examples.
Another limitation of the study is that with a moderate fidelity
simulator there are questions around the extent to which the speed
perception cues available in real world are present. Such cues may
help drivers on real roads to slow down more in conditions of
high roadside object density and low pavement width, such as
the Urban Full Parking condition. Ideally, future research would
include a wider field of view and more on-road measures to deter-
mine if the effects are of the same magnitude. However, the fact
that the present study showed changes in driving performance and
workload in the absence of these cues is a testament that useful
results can be obtained even without the highest degree of simula-
tor fidelity.
The present study examined driver behaviour under the instruc-
tion to drive at 60 km/h (the speed limit for most connector roads
in Australian cities). It would be useful to understand how drivers
adapt to on-street parking under a range of speed restrictions (i.e.
speed limits). Previous research has shown that drivers under load
from a secondary task actually drive faster than when not under
load, if the speed they are asked to drive at is lower than the speed
they would habitually drive at (Lewis-Evans et al., 2011). It is there-
fore important to understand how workload induced by the road
environment interacts with speed restrictions, particularly where
the speed limit is lower than drivers would usually choose in simi-
lar situations (for example, variable speed limits). Further research
could also examine whether traffic calming devices and/or visual
treatments can successfully change behaviour to match risk levels
posed by factors such as on-street parking.
Role of funding source
This project was funded through MUARC’s Baseline Research
Program for which grants have been received from: the Victorian
Department of Justice, VicRoads, and the Transport Accident Com-
mission (TAC). The sponsors gave input into the overall direction
of the research (investigating crashes in urban environments); the
design, analysis, interpretation and writing were all performed by
the researchers.
Acknowledgements
The authors would like to thank Mr Ashley Verdoorn for pro-
gramming the driving simulator scenarios and Ms Karen Stephan
for her contribution towards study design.
References
Allsop, R., 1995. Reducing Traffic Injuries Resulting From Excess and Inappropriate
Speed. European Transport Safety Council, Brussels.
Brooks, J.O., Crisler, M.C., Klein, N., Goodenough, R., Beeco, R.W., Guirl, C., et al., 2011.
Speed choice and driving performance in simulated foggy conditions. Accident
Analysis & Prevention 43 (3), 698–705.
Chinn, L., Elliot, M.A., 2002. The effect of road appearance on perceived safe travel
speed: Final report (No. PA3828/02): Transport Research Laboratory.
Daisa, J.M., Peers, J.B., 1997. Narrow residential streets: do they really slow down
speeds? In: ITE Annual Meeting Compendium, Washington DC: Institute of
Transportation Engineers, pp. 546–551.
De Waard, D., 1996. The measurement of drivers’ mental workload. The Uni-
versity of Groningen, Groningen, Netherlands. From http://dissertations.ub.
rug.nl/faculties/ppsw/1996/d.de.waard/.
Elliot, M.A., Mccoll, V.A., Kennedy, J.V., 2003. Road design measures to reduce drivers’
speed via ‘psychological’ processes: A literature review (No. TRL564): Transport
Research Laboratory.
Engstrom, J., Johansson, E., Ostlund, J., 2005. Effects of visual and cognitive load in
real and simulated motorway driving. Transportation Research Part F: Traffic
Psychology and Behaviour 8 (2), 97–120.
Fuller, R., 2005. Towards a general theory of task behaviour. Accident Analysis &
Prevention 37, 461–472.
Fuller, R., McHugh, C., Pender, S., 2008. Task difficulty and risk in the determina-
tion of driver behaviour. Revue Européenne de Psychologie Appliquée/European
Review of Applied Psychology 58 (1), 13–21.
Greibe, P., 2003. Accident prediction models for urban roads. Accident Analysis &
Prevention 35, 273–285.
 J. Edquist et al. / Accident Analysis and Prevention 45 (2012) 759–765
Gugerty, L.J., 1997. Situation awareness during driving: explicit and implicit knowl-
edge in dynamic spatial memory. Journal of Experimental Psychology: Applied
3 (1), 42–66.
Harms, L., 1986. Drivers’ attentional response to environmental variations: a dual-
task real traffic study. In: Gale, A.G., Freeman, M.H., Haslegrave, C.M., Smith, P.,
Taylor, S.P. (Eds.), Vision in Vehicles, I. North-Holland, pp. 131–138.
Hart, S.G., Staveland, L.E., 1988. Development of NASA-TLX (Task Load Index):
results of empirical and theoretical research. In: Hancock, P.A., Meshkati,
N. (Eds.), Human Mental Workload. Elsevier Science, Amsterdam, pp.
139–183.
Horberry, T., Anderson, J., Regan, M.A., Triggs, T.J., Brown, J., 2006. Driver distraction:
the effects of concurrent in-vehicle tasks, road environment complexity and age
on driving performance. Accident Analysis & Prevention 38 (1), 185–191.
Hoyos, C.G., 1988. Mental load and risk in traffic behaviour. Ergonomics 31 (4),
571–584.
Lewis-Evans, B., Charlton, S., 2006. Explicit & implicit processes in behavioural adap-
tation to road width. Accident Analysis & Prevention 38, 610–617.
Lewis-Evans, B., de Waard, D., Brookhuis, K.A., 2011. Speed maintenance under cog-
nitive load – implications for theories of driver behaviour. Accident Analysis &
Prevention 43 (4), 1497–1507.
Martens, M.H., Comte, S., Kaptein, N.A., 1997. The effects of road design on speed
behaviour: a literature review: TNO Human Factors Research Institute.
Nunes, L.M., Recarte, M.A., 2005. Speed, traffic complexity and visual performance:
a study on open road. In: Underwood, G., Rothengatter, T., Huguenin, R.D. (Eds.),
765
Traffic and Transport Psychology Theory and Application: Proceedings of the
ICTTP 2004. Elsevier, pp. 339–354.
Pande, A., Abdel-Aty, M., 2009. A novel approach for analyzing severe crash patterns
on multilane highways. Accident Analysis & Prevention 41 (5), 985–994.
Parkes, A.M., Luke, T., Burns, P.C., Lansdown, T., 2007. Conversations in cars: the
relative hazards of mobile phones (No. 664): Transport Research Laboratory.
Roberts, I., Norton, R., Jackson, R., Dunn, R., Hassall, I., 1995. Effect of environmental
factors on risk of injury of child pedestrians by motor vehicles: a case-control
study. British Medical Journal 310, 91–94.
Stinchcombe, A., Gagnon, S., 2010. Driving in dangerous territory: complexity
and road-characteristics influence attentional demand. Transportation Research
Part F: Traffic Psychology and Behaviour 13 (6), 388–396.
Summala, H., 1988. Risk control is not risk adjustment: the zero-risk theory of driver
behaviour and its implications. Ergonomics 31 (4), 491–506.
Tarko, A.P., 2009. Modeling drivers’ speed selection as a trade-off behavior. Accident
Analysis & Prevention 41 (3), 608–616.
Trick, L.M., Toxopeus, R., Wilson, D., 2010. The effects of visibility conditions, traffic
density, and navigational challenge on speed compensation and driving perfor-
mance in older adults. Accident Analysis & Prevention 42 (6), 1661–1671.
van der Horst, R., de Ridder, S., 2007. Influence of roadside infrastructure on driving
behavior: driving simulator study. Transportation Research Record 2018 (2018),
36–44.
Wilde, G.J.S., 1982. The theory of risk homeostasis: implications for safety and health.
Risk Analysis 2, 209–225.