Applied Ergonomics 60 (2017) 348e355

Contents lists available at ScienceDirect

Applied Ergonomics
journal homepage: www.elsevier.com/locate/apergo

Effects of vibration on occupant driving performance under simulated

driving conditions
Amzar Azizan a, b, *, M. Fard a, Michael F. Azari c, Reza Jazar a
a
School of Engineering, RMIT University, Melbourne, Australia
b
University of Kuala Lumpur, Malaysia
c
School of Health and Biomedical Sciences, RMIT University, Melbourne, Australia

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

Article history: Although much research has been devoted to the characterization of the effects of whole-body vibration
Received 14 January 2016 on seated occupants' comfort, drowsiness induced by vibration has received less attention to date. There
Received in revised form are also little validated measurement methods available to quantify whole body vibration-induced
30 September 2016
drowsiness. Here, the effects of vibration on drowsiness were investigated. Twenty male volunteers
Accepted 27 December 2016
were recruited for this experiment. Drowsiness was measured in a driving simulator, before and after 30-
min exposure to vibration. Gaussian random vibration, with 1e15 Hz frequency bandwidth was used for
excitation. During the driving session, volunteers were required to obey the speed limit of 100 kph and
Keywords:
Human vibration
maintain a steady position on the left-hand lane. A deviation in lane position, steering angle variability,
Drowsiness and speed deviation were recorded and analysed. Alternatively, volunteers rated their subjective
Lane deviation drowsiness by Karolinska Sleepiness Scale (KSS) scores every 5-min. Following 30-min of exposure to
Speed deviation vibration, a signicant increase of lane deviation, steering angle variability, and KSS scores were observed
Steering angle in all volunteers suggesting the adverse effects of vibration on human alertness level.
2016 Elsevier Ltd. All rights reserved.

1. Introduction transmitted vibration to the seated human body has a signicant

Drowsiness is one of the leading causes of accidents on motor-
ways and major roadways, accounting for approximately 20% of
road accidents worldwide (Horne and Reyner, 1995). Drowsiness,
which refers to sleepiness, is a multifactorial state that may lead to
inappropriate driving behaviour such as lack of awareness, poor
judgement and slowed reaction times (Anderson and Horne, 2013).
In addition, drowsiness as a result of alcohol intake or monotonous
driving conditions or night driving is known to signicantly inu-
ence driving performance, compromising transportation safety
(Larue, 2010; Kamdar et al., 2004; Akerstedt et al., 2005; Fairclough
and Graham, 1999). Although the performance of vehicle drivers
has been well investigated under various conditions, vibration-
induced drowsiness is not well-characterised. Relationships be-
tween amplitude and frequency of vibration and drowsiness levels
have been assumed without sufcient quantitative data. According
to ISO 2631-1 (1997) International Standards (1997), the
* Corresponding author. School of Engineering, RMIT University, Melbourne,
Australia.
E-mail address: s3356158@student.rmit.edu.au (A. Azizan).
inuence on human perception and ride comfort (Grifn, 1990;
Baik, 2004; Factors et al., 1997). Exposure to vibration also has
been found to correlate with a range of physiological reactions of
the human body such as lower back pain and reduction in heart rate
variability (Vicente et al., 2011; Callaghan and McGill, 2001).
Although many studies have contributed much to the under-
standing and prediction of the subjective human body response to
vibration (Manseld and Maeda, 2011; Maeda et al., 2008; Tewari
and Prasad, 2000; Kjellberg and Wikstro m, 1985), few studies
have considered the effect of vibration specically on drowsiness
levels for seated occupants in the vehicle.
Therefore, there is considerable scope for dening the exact
effects of vehicle and particularly seat vibration on driver drowsi-
ness levels. According to several published reports on drowsiness
and vehicle control, there is a close relationship between drowsi-
ness and vehicle lateral control (standard deviation of lateral
position-SDLP, steering angle variability) as well as longitudinal
control (speed deviation) (Wierwille and Kirn, 1994; Thiffault and
Bergeron, 2003). SDLP is calculated as a standard deviation of the
average lateral position and corresponds to the amount of weaving
in the car and increases in SDLP may ultimately result in the lane

http://dx.doi.org/10.1016/j.apergo.2016.12.020
0003-6870/ 2016 Elsevier Ltd. All rights reserved.
 A. Azizan et al. / Applied Ergonomics 60 (2017) 348e355
crossing into the road shoulder and adjacent trafc lane. Steering
angle variability is calculated as a deviation from the center of
steering angle. Zero deviation means the center of the vehicle co-
incides or is parallel to the center of lane position. Speed adjust-
ment from the posted speed limit will result in speed deviation.
Therefore, the primary dependent variables for this investigation
were volunteers' SDLP measured from simulated driving vehicle,
steering angle variability and speed deviation.
Although many studies have attempted to demonstrate the links
between driving performance and drowsiness, drowsiness caused
by vehicle vibration has not been experimentally assessed by
simulated driving. Therefore, it was also important to investigate
the feasibility and utility of simulated driving in the detection of
drowsiness caused by vibration. Hence, it was the primary aim of
this study to investigate the effects of vibration on human drows-
iness level using both objective (Simulated Driving Test) and sub-
jective (Karolinska Sleepiness Scale) measurement methods.
2. Methods
2.1. Recruitment and screening
Twenty young male (n 20) participated in this investigation
with a mean age (SD) 23.0 1.3. They were randomly selected
from university students. They had no history of low back pain
(LBP) and normal or corrected-to-normal vision. Their de-
mographic were recorded at enrolment. They were (mean SD)
168.2 4.0 cm and weighed (mean SD) 64.2 12.2 kg. The
average BMI of participants was (mean SD) 22.6 2.54 kg/m2.
One week prior to a laboratory experiment, volunteers were
required to maintain their normal amount of sleep (between 7 and
9 h) and keep a normal sleep schedule. Therefore, they were asked
to keep a sleep diary of when they go to sleep at night and wake up
each day. All volunteers were screened before the day of experi-
ment using Pittsburgh Sleep Quality Index (PSQI) to measure the
sleep quality (Buysse et al., 1989). Volunteers who showed poor
sleep quality index (PSQI > 5) were excluded from the
investigation.
2.2. Ethical considerations
Before the investigation, volunteers were provided with verbal
and written explanations of the purpose and contents of the
experiment. They were also informed that they have right to refuse
participation in the experiment, and the results of the experiment
would remain condential. Following this, the informed written
consent form was obtained from all the volunteers after the pro-
cedure of the experiment was explained, and the laboratory facil-
ities were introduced to them. The experimental protocol was
reviewed and approved by the RMIT University Human Research
Ethics Committee (Approval Number: EC 00237).
2.3. Experiment setup
The experiment setup for drowsiness assessment is illustrated
in Fig. 1. The vehicle seat with adjustable headrest was mounted on
a vibration table. The vibration table was mounted on four air
cushions. The vehicle seat's inclination angle was set at 15  to the
vertical direction. Experiment set-up has been developed with a
single vertical hydraulic actuator to replicate the vibration
perceived by seated occupant in a moving vehicle. Although, the
input vibration was not independent on each axis, however, the
input vibration generated from the hydraulic vertical actuator is
located below the table away from the center of the table. The off-
center excitation provides the multi-axial (x,y, z-axis) input
349
vibration. This vibration setup also was built to be somewhat
similar to the vibration that is transferred from the vehicle oor to
the seat. The vibration table below the seat was designed to be
dynamically rigid in frequencies below 100 Hz. This is to ensure
that there is no interaction with vehicle seat structural dynamics.
Prior to drowsiness measurement, measurement of total trans-
mitted vibration for each volunteer has been done in accordance
with ISO 2631-1 (1997) International Standards (1997). The total
transmitted vibration was measured both from the vehicle's seat-
back and seat pan. The measurement was carried out to adjust the
required hydraulic input force for every volunteer to become
0.2 ms2 r.m.s.
Two tri-axial accelerometer pads (SVANTEK SV-38V model)
were used to measure the total transmitted vibration to the human
body located at the seat cushion and the seatback (Fard et al., 2014).
The SV 106 Human Vibration Exposure (HVE) meter (analyser),
which was connected to the accelerometer pads, was used to obtain
the total frequency weighted transmitted vibration to the seated
human body. The HVE analyser uses the weighting factors (Wk, Wd,
Wc) and multiplication factors (Table 1) to calculate the total
frequency-weighted transmitted vibration. The weighting curves
(Wk, Wd, Wc) given in Table 1 were from ISO 2631-1 (1997)
International Standards (1997). The frequency weighting curves
dene the values by which the vibration magnitude at each specic
frequency is to be multiplied in order to weight the measured vi-
bration in accordance with the human body (Grifn, 1990). The
multiplication factors (Table 1) were used to weight the effects of
seatback and seat cushion vibrations (Grifn, 1990; M. Amzar and
Fard, 2013; Fard et al., 2014).
2.4. Drowsiness measurement
2.4.1. Objective measure (Simulated Driving Test)
Volunteers were tested on the York driving simulator software
(York Computer Technologies, Kingston, Ontario, Canada) as shown
in Fig. 2. The simulator has been determined to be an ecologically
valid research tool to measure psychomotor performance related to
driving (Arnedt et al., 2005; Chung et al., 2005). The simulator as-
sembly consists of a personal computer, a 40-inch monitor and
peripheral steering wheel, accelerator and brake accessories. A
customized driving scenario was developed in which volunteers
were presented with a forward view from the driver's seat. The
driving simulation showed a cross-country highway, with two
lanes in each direction.
The two primary instruction for volunteers are:
1. Maintain a steady position within the left trafc lane during the
entire test.
2. Maintain a constant speed (100 kph).
Outcome variables measured by the simulator included (a)
standard deviation of lateral position (SDLP) (b) speed deviation
and (c) steering angle variability. The variation in these three
outcome measures shows how well the volunteers able to conduct
the test according to this instruction.
2.4.2. Subjective measure (Karolinska Sleepiness Scale-KSS)
Subjective drowsiness level was assessed using Karolinska
Sleepiness Scale (KSS). KSS is a self-reported and subjective
assessment of sleepiness that measures changes in sleepiness level
at the time (Gillberg et al., 1994). It is a 9-point Likert scale varied
from 1 extremely alert, 2 very alert, 3 alert, 4 rather alert,
5 neither alert nor sleepy, 6 some sign of sleepiness, 7 sleepy,
but no effort to stay awake, 8 sleepy, some effort to stay awake,
9 very sleepy, great effort to stay awake. KSS has been widely
350 A. Azizan et al. / Applied Ergonomics 60 (2017) 348e355

Fig. 1. (a) Schematic drawing of the experiment setup. An actual vehicle's seat was mounted on a vibration table. A hydraulic actuator located at the corner of the table will provide
multi-axial input to the volunteer. A simulator also consists of a personal computer, a 40-inch monitor and peripheral steering wheel, accelerator, and brake accessories. (b) Forward
view of simulated driving software. Volunteers were required to obey the speed limit of 100 kph and maintain a steady position on the left-hand lane.

Table 1 2.5. Experimental procedures

The frequency weighting coefcients and the multiplication factors, from ISO 2631-1
(1997) International Standards corresponding to X (fore-aft), Y (lateral), and Z
(vertical) accelerations in the seat cushion and the seatback. Wk, Wd, Wc are shown
The experiment was carried out in a temperature and light
here. controlled (21  Ce23  C, <70 lux) laboratory and the noise level
was below 60 dB. Volunteers arrived at the laboratory at 8.00 h. The
Seat accelerometer pad
experiment began at 8.30 for all volunteers. Prior to the experi-
Seat cushion Seatback ment, all volunteers were screened using Epworth Sleepiness Scale
Axis X Y Z X Y Z (ESS) to detect any abnormalities in sleep (John, 1991). Those with
Weighting Wd Wd Wk Wc Wd Wd scores above ten, which indicated excessive daytime sleepiness,
Multiplication factor k 1 1 1 0.8 0.5 0.4 were excluded from the experiment.
Volunteers performed two separate test conditions [baseline
(no-vibration condition) and with-vibration condition] in a ran-
domized cross-over design, one week apart. The orders of two
conditions were randomly ordered to avoid order-related in-
uences. To standardise the learning effect, volunteers underwent a
10-min practice session before baseline and with-vibration condi-
tions to familiarize themselves with the simulator interface. During
with-vibration condition, volunteers were asked to drive for 10-
min with no vibration followed by 30-min sitting with exposure
to vibration with their eyes open. Volunteers were exposed to
Gaussian random vibration, with 1e15 Hz frequency bandwidth.
Total transmitted acceleration to the human body measured from
seat pan and backrest was kept constant at 0.2 ms2 r.m.s. Volun-
teers rated their subjective sleepiness scale using KSS before vi-
bration exposure, every 5-min of vibration and after vibration
exposure (Akerstedt and Gillberg, 1990). The rating was initiated by
the test leader saying KSS. Immediately after 30-min sitting,
Fig. 2. SDLP and steering angle variability were calculated as a deviation from the volunteers were required to drive for another 10-min. Similar
center of lane position. Zero deviation occurs when the center of the vehicle coincides procedures and sitting arrangement as in with-vibration condition
with the center of lane position. Speed deviation is measured as the deviation from the with the only difference being no vibration exposure were applied
posted speed limit (100 kph). for 30-min sitting. See Fig. 3 for an overview of the experiment
setup. The total duration of each condition (no-vibration and with-
vibration) is 50-min.
used as a prediction tool particularly to measure the changes over
time of sleepiness/drowsiness level. Lower KSS scores (<5) may
indicate a difference state of alertness level. Higher KSS scores (>7) 3. Results
may refer to the state of drowsiness where subjects may experience
involuntary dozing behaviour and frequent lapse episodes. At every 3.1. Objective measurement
5-min interval during the simulated driving task, volunteers were
prompted by the word KSS by the investigator to provide sub- Statistical analysis was performed using GraphPad software
jective rating score according to scales visible at all times next to (GraphPad Prism 6). All data were checked for normality prior
the monitor screen. analysis. Two-tailed (p < 0.05) was considered statistical signi-
cance. Two-way repeated measures - Analysis of Variance (ANOVA)
 A. Azizan et al. / Applied Ergonomics 60 (2017) 348e355 351

Fig. 3. The volunteers were required to complete a 10-min driving session both before and repeated after 30-min sitting with no-vibration and with vibration. 10-min practice
session was provided for all the volunteers at the beginning of every condition.

was used to assess the changes in objective measurement over the
course of vibration exposure. All post hoc testing was performed via
the Tukey post-hoc test. There was no signicant differences in
alertness level measured by Epworth Sleepiness Scale (ESS) at
baseline measurement between no-vibration condition and with-
vibration conditions. Comparison of driving performance indexes
(SDLP, speed deviation, and steering angle variability) and subjec-
tive sleepiness scale (KSS) between no-vibration and with-
vibration condition are presented here.
Results of Standard Deviation of Lateral Position (SDLP) are
shown in Fig. 4. Each panel of the gure shows the average and
standard deviation (SD) of SDLP before and after 30-min of expo-
sure to vibration and sitting (no-vibration). From Fig. 4, it can be
seen that in the rst 10-min driving, there was no signicant dif-
ference in SDLP was observed between before no-vibration con-
dition and before with-vibration condition (p > 0.05). As compared
to the baseline (no-vibration condition) SDLP measure between
before exposure and after exposure to vibration showed that 30-
min exposure to vibration had signicant inuences on volunteer
lateral keeping performance. We found that following 30-min
exposure to vibration, the deviation from a lateral position or
lateral variability was signicantly increased from (mean SD:
23.6 0.07 to 26.2 0.06; t(19) 2.65, p < 0.01). The analysis of
lateral variability showed difculties in maintaining the vehicle in
the middle of the left-hand lane when alertness was lowest
following exposure to vibration. No signicant difference was
observed in the baseline (no-vibration condition) where SDLP was
reduced from (mean SD: 23.5 0.05 to 22.8 0.04; t(19) 0.91,
p > 0.05). The repeated-measures analysis of both conditions (no-
vibration and with-vibration) revealed a signicant difference be-
tween group variations (p < 0.01) that suggest a considerable in-
uence of vibration on drowsiness level measured by SDLP.
The secondary outcome index measured in this study was speed
deviation. Speed deviation is calculated as the average sum of the
differences of the vehicle speed from the posted speed limit. The
results obtained from the speed deviation are presented in Fig. 5.
This gure shows the comparison of speed deviation measured for
all volunteers between the no-vibration condition and with-
vibration condition. Even though there seemed to be a slight in-
crease of speed deviation following exposure to 30-min vibration
from (mean SD: 5.30 5.96 to 6.28 5.97; t(19) 1.26, p > 0.05)
and 30-min sitting from (mean SD: 4.68 3.28 to 5.59 3.70;
t(19) 1.45, p > 0.05), this did not reach statistical signicance
(p > 0.05). Speed deviation did not show signicant variation as
compared to SDLP. These results suggest that speed deviation was
not sensitive to levels of alertness caused by vibration as used
under these experimental conditions.
The third outcome index measured in this study was steering

Fig. 4. This bar graph represents the mean (SD) of SDLP for twenty volunteers in no-vibration condition and with-vibration condition. The measurement was taken at 10-min
driving before and 10-min driving after 30-min sitting in no-vibration and 30 min sitting with-vibration The signicant increase of lateral position variability (SDLP) (p < 0.01)
was observed in with-vibration condition. These differences, corresponding to common legal limits for driving were 2.4 cm (BAC 0.05%) (Verster and Roth, 2011). The changes
indicate the inability of the volunteer to maintain straight position once subjected to 30-min vibration. **p < 0.01.
352 A. Azizan et al. / Applied Ergonomics 60 (2017) 348e355

Fig. 5. This bar graph represents the mean (SD) of speed deviation for twenty volunteers in no-vibration condition and with-vibration condition. The measurement was taken at
10-min driving before and 10-min driving after 30-min of sitting in no-vibration and 30-min sitting with-vibration. No signicant changes (p > 0.05) in speed deviation were
observed in both conditions (no-vibration and with vibration) following 30-min exposure to vibration and sitting.

angle variability. Steering angle variability is calculated as the de-
viation from the center of steering angle. Fig. 6 shows the mean of
steering angle variability measured for all twenty volunteers be-
tween the no-vibration condition and with-vibration condition. As
Fig. 6 shows, there is a signicant increase of steering angle vari-
ability (mean SD: 0.67 0.20 to 0.82 0.20; t(19) 3.34, p < 0.01)
between before 30-min vibration and after 30 min exposure to
vibration. Data for this table can be compared with the data in no-
vibration condition. No signicant differences were detected be-
tween before 30-min sitting and after 30-min sitting (mean s.d:
0.69 0.22 to 0.73 0.28; t(19) 0.58, p > 0.05).
3.2. Subjective measurement (Karolinska Sleepiness Scale)
Comparison between the means of subjective sleepiness scale
(KSS) for all volunteers in both conditions (no-vibration and with-
vibration) are plotted against time and presented in Fig. 7. To
investigate statistical signicance, two-way repeated measures-
ANOVA corrected for Turkey's multiple comparisons was carried
out. As illustrated here, initial KSS values did not differ between
both conditions (p > 0.05). The average KSS value was (mean SD:
2.11 0.57 in with-vibration condition and 2.21 0.85 in no-
vibration condition) at the beginning of the experiment. During
the rst 10-min of driving, no signicant difference was observed in
both conditions with KSS score of (mean SD: 3.58 1.26 in with-
vibration condition and 3.16 0.83 in no-vibration condition). It is

Fig. 6. This bar graph represents the mean (SD) of steering angle variability for twenty volunteers in no-vibration condition and with-vibration condition. The measurement was
taken at 10-min driving before and 10-min driving after 30-min of sitting in no-vibration and 30-min sitting with-vibration. As can be seen in this gure, with-vibration condition
reported signicantly increase in steering angle variability than in no-vibration condition (p < 0.01). **p < 0.01.
 A. Azizan et al. / Applied Ergonomics 60 (2017) 348e355
Fig. 7. The average score (mean SD) of subjective sleepiness scale (KSS) plotted against time for twenty volunteers in no-vibration condition and with-vibration condition. At the
beginning of the experiment, no signicant differences were observed between the two conditions (p > 0.05). The onset of drowsiness can be observed following 10-min of vi-
bration exposure in with-vibration condition, and the difference was signicant compared to no-vibration condition (p < 0.001).
apparent from Fig. 7 that there is a denite decline in alertness
levels indicated by a progressive increase in subjective sleepiness
score throughout the course of exposure to vibration. The onset of
drowsiness was more rapid following 15-min exposure to vibration
with KSS value of (mean SD: 6.00 1.41) whereas KSS value of
only (mean SD: 4.05 1.13) was observed in no-vibration con-
dition. After 30-min, KSS value increased signicantly with expo-
sure to vibration than without vibration (mean SD: 7.89 0.81
and 4.53 0.70 respectively; p < 0.0001). The second driving test
was conducted after the vibration stopped. The mean sleepiness
level reduced indicating a slight decrease in overall sleepiness level
(mean SD: 7.26 1.00 in with-vibration condition and 4.42 0.61
in no-vibration condition).
3.3. Correlation
Furthermore, we interrogated the correlation between subjec-
tive measure (KSS) and objective measure (SDLP, speed deviation
and steering angle variability of this study by the use of Pearson's
(r) correlation). It is apparent from Fig. 8(a) that, there was a sig-
nicant positive correlation between SDLP and KSS (r 0.972;
p < 0.05) and between steering angle variability and KSS (r 0.983;
p < 0.05). A positive correlation indicated that impairments in SDLP
and steering angle variability were related to subjective self-
reported sleepiness, and the relationship was statistically signi-
cant (p < 0.05). A positive correlation between speed deviation and
KSS did not reach statistical signicance (r 0.619; p > 0.05).
Further analysis between objective measures showed that there
were no signicant correlation between speed deviation and SDLP
(r 0.418; p > 0.05) and between speed deviation and steering
angle variability (r 0.678; p > 0.05). As expected, the correlation
between steering angle variability and SDLP was high, and the re-
sults were signicant (r 0.928; p < 0.05).
4. Discussion
Driver drowsiness has been one of the primary causes of road
accidents (Hosking, 2006). However, drowsiness that is caused by
vehicle vibration is not well understood or investigated. The pre-
sent study was designed to determine the relationship between
human drowsiness levels and exposure to whole-body vibration as
353
can be experienced in a vehicle. We demonstrated that the increase
of human drowsiness level, measured by simulated driving soft-
ware and Karolinska Sleepiness Scale (KSS) signicantly correlated
with exposure to vibration. These data support the hypothesis that
exposure to vibration (random vibration with 1e15 Hz frequency
band), even for as little as 30-min under these experimental con-
ditions, causes drowsiness and adversely affects psychomotor
performance as measured by lane keeping performance and
steering angle variability, found in all twenty volunteers. The def-
icits in performance observed in this study are comparable to
performance on the York Driving Simulator (YDS) under alcohol
intoxication (blood alcohol levels above 0.05% which corresponds
to severe performance decits) (Arnedt et al., 2001; Verster et al.,
2014) (Table 2).
SDLP is related to the levels of drowsiness in the driver. SDLP
values become larger as the driver becomes drowsier. As drowsi-
ness increases, situational awareness decreases and the drivers'
ability to predict an upcoming event is lowered. This lead to over-
compensation by the driver and the wheel tends to be moved
frequently. The continuous variations of the standard deviation of
lateral position (SDLP) suggest low levels of alertness when vol-
unteers were exposed to vibration and hence cause difculties in
maintaining the vehicle in the middle of the left-hand lane. The
ndings observed in this study mirror previous studies that have
reported SDLP changes as the most common and persistent nding
in sleep deprivation and drowsiness (Verster et al., 2014;
Ramaekers et al., 2000). Our results demonstrate, for the rst
time that exposure to whole-body vibration may cause psycho-
motor decits under simulated driving conditions. The current
study also found that there is an increase of steering angle vari-
ability following exposure to vibration. The observed increase in
steering angle variability could be attributed to drowsiness, and the
ndings are consistent with previous studies (He et al., 2011;
Forsman et al., 2013; Nakayama et al., 1999). Another important
nding to emerge from this study is that speed deviation failed to
exhibit any statistically signicant variation between no-vibration
condition and with-vibration condition. The present ndings
seem to be consistent with other research that found that a lon-
gitudinal measure such as speed deviation is not capable of
detecting drowsiness (Arnedt et al., 2001). As expected, results
from the subjective measurement (KSS) also showed a signicant
354 A. Azizan et al. / Applied Ergonomics 60 (2017) 348e355

Fig. 8. (a) The correlation between objective measures (SDLP, steering angle variability, speed deviation) and subjective sleepiness scale (KSS) in with-vibration condition. A high
correlation (r > 0.9) between objective measures (SDLP and steering angle variability) and KSS were observed, and the results were signicant (p < 0.05). (b) Correlation between
objective measures (SDLP, steering angle variability, speed deviation). A highly signicant correlation was observed between SDLP and steering angle variability (r 0.928; p < 0.05).
However, no signicant correlation was observed between Speed deviation and SDLP (r 0.418; p > 0.05) and between speed deviation and steering angle variability (r 0.678;
p > 0.05).
*p < 0.05.

Table 2 replicate actual driving conditions. Therefore, the vibration

Comparison of the SDLP (cm) in this study with those reported by Verster (Verster
et al., 2014).
This study J. C Verster
Baseline With-Vibration Baseline BAC 0.05%
23.6 26.2 24.2 26.1
*BAC Blood Alcohol Concentration.
decline in alertness level for all the volunteers following 30-min
exposure to vibration. The increase in subjective sleepiness scores
provides important corroborating evidence that exposure to 30-
min of vibration level can steadily reduce human alertness levels
that are linked to drowsiness (Hallvig et al., 2013). We also found
that sleepiness levels immediately decreased after the vibration
stopped, showing an acute improvement, albeit not down to alert
levels.
A high correlation was observed (r > 0.9) between the objective
measures (SDLP and steering angle variability) and the subjective
measure (KSS) adding weight to the results. This study also con-
rms the association between SDLP and steering angle variability
(r 0.928; p < 0.05). The results demonstrated that, as drowsiness
increases, there is a reduction in vehicle lateral control measured by
SDLP and steering angle variability. However, longitudinal control
measured by speed deviation is found to be much less affected by
drowsiness.
Our experiment design and procedures were developed to
perceived by volunteers in this study is similar to the actual vi-
bration felt in a typical vehicle. An actual vehicle seat was selected
to ensure proper vibration transmission to the human body. The
assessment and guidelines of human body ride comfort caused by
vibration are were founded in ISO 2631-1 (1997) International
Standards (1997). Since then, the relevant weighting factor has
been established to represent the human perception of vibration.
Although the guidelines for health effect due to exposure to vi-
bration are well documented, there is little quantitative research
data available on the inuences of vibration on seated human
drowsiness. This study demonstrates a link between exposure to
vibration and drowsiness, at least under these experimental con-
ditions. Therefore, it is imperative to further characterize this as-
sociation under combined noise and vibration conditions that more
closely resemble driving and to identify the component of noise
and vibration that are most responsible for the decline in driver
alertness.
5. Conclusion
The novel contribution of this study is the characterization the
role of vibration on occupant drowsiness. Our ndings have iden-
tied vibration as an important source of driver drowsiness. The
evidence from this study clearly demonstrates that exposure to
low-frequency random vibration between 1 and 15 Hz has
considerable inuence on subjective sleepiness levels, and more
 A. Azizan et al. / Applied Ergonomics 60 (2017) 348e355
importantly, human psychomotor performance and lapse of
attention. Exposure to vibration signicantly impacts lane keeping
performance of the driver (SDLP). Our ndings show a low trans-
mitted vibration at 0.2 ms2 r.m.s. Increased SDLP by 11% which is
comparable to SDLP under the inuence of alcohol (BAC 0.05%).
These data suggest that SDLP and steering angle variability are
stable measures of driving performance with high reliability. KSS
also showed utility as a sensitive measure of vibration-induced
drowsiness. These data provide tantalising evidence that should
prompt further research in this area. Such research would ulti-
mately lead to the development of relevant and practical guidelines
to limit vibration exposure in the automotive industry, in an effort
to reduce the burden of disease of road accidents. These guidelines
will complement the existing ISO 2631-1 (effects of vibration on
comfort) to extend these guidelines in assessment and establish-
ment of thresholds and safe limits for drowsiness-inducing
vibration.
References
Akerstedt, T., Gillberg, M., 1990. Subjective and objective sleepiness in the active
individual. Int. J. Neurosci. 52, 29e37.
Akerstedt, T., Peters, B., Anuand, A., Kecklund, G., 2005. Impaired alertness and
performance driving home from the night shift: a driving simulator study.
J. Sleep Res. 14 (1), 17e20. Available at: http://www.ncbi.nlm.nih.gov/pubmed/
15743329.
Amzar, A., Fard, M., 2013. Effects of Vehicle Seat Dynamics on Ride Comfort
Assessment. International Congress on Noise and Vibration.
Anderson, C., Horne, J.A., 2013. Driving drowsy also worsens driver distraction.
Sleep Med. 14 (5), 466e468. Available at: http://www.ncbi.nlm.nih.gov/
pubmed/23523431 [Accessed October 7, 2014].
Arnedt, J.T., Wilde, G.J., Munt, P.W., Maclean, A.W., 2001. How do prolonged
wakefulness and alcohol compare in the decrements they produce on a simu-
lated driving task? Accid. Anal. Prev. 33 (3), 337e344. Available at: http://www.
ncbi.nlm.nih.gov/pubmed/11235795.
Arnedt, J.T., Owens, J., Crouch, J., Stahl, Carskadon, M.A., 2005. Neurobehavioral
performance of residents after heavy night call vs after alcohol ingestion. JAMA:
J. Am. Med. Assoc. 294 (9), 1025e1033.
Baik, S., 2004. A study on the characteristics of vibration in seat system. KSCE J. Civ.
Eng. 8 (1), 135e139.
Buysse, D.J., Reynolds, C.F., Monk, T.H., Berman, S.R., Kupfer, D.J., 1989. The Pitts-
burgh Sleep Quality Index: a new instrument for psychiatric practice and
research. Psychiatry Res. 28 (2), 193e213. Available at: http://www.ncbi.nlm.
nih.gov/pubmed/2748771.
Callaghan, J.P., McGill, S.M., 2001. Low back joint loading and kinematics during
standing and unsupported sitting. Ergonomics 44 (3), 280e294. Available at:
http://www.ncbi.nlm.nih.gov/pubmed/11219760 [Accessed May 15, 2014].
Chung, F., Kayumov, L., Sinclair, D.R., Edward, R., 2005. What is the driving per-
formance of ambulatory surgical patients after general anesthesia? Anasthesi-
ology 103 (5), 951e956.
Factors, H., Kitazaki, S., Grifn, M.J., 1997. A modal analysis of whole-body vertical
vibration, using a nite element model of the human body. J. Sound Vib. 200,
83e103. Available at: http://www.sciencedirect.com/science/article/pii/
S0022460X96906744.
Fairclough, S.H., Graham, R., 1999. Impairment of driving performance caused by
sleep deprivation or alcohol: a comparative study. Hum. Factors 41 (1),
118e128. Available at: http://hfs.sagepub.com/cgi/doi/10.1518/
001872099779577336 [Accessed October 8, 2014].
355
Fard, M., Lo, L., Subic, A., Jazar, R., 2014. Effects of seat structural dynamics on
current ride comfort criteria. Ergonomics 57 (10), 1549e1561. Available at:
http://www.ncbi.nlm.nih.gov/pubmed/25017144 [Accessed December 11,
2014].
Forsman, P.M., Vila, B.J., Short, R.A., Mott, C.G., Dongen, H.P.A., 2013. Efcient driver
drowsiness detection at moderate levels of drowsiness. Accid. Anal. Prev. 50,
341e350. Available at: http://www.ncbi.nlm.nih.gov/pubmed/22647383
[Accessed December 8, 2014].
Gillberg, M., Kecklund, G., Akerstedt, T., 1994. Relations between performance and
subjective ratings of sleepiness during a night awake. Sleep 17 (3), 236e241.
Available at: http://www.ncbi.nlm.nih.gov/pubmed/7939123.
Grifn, M.J., 1990. Handbook of Human Vibration. Academic Press Limited, London.
Hallvig, D., Anund, A., Fors, C., Kecklund, G., Karlson, J.G., Wahde, M., Akerstedt, T.,
2013. Sleepy driving on the real road and in the simulator - a comparison. Accid.
Anal. Prev. 50, 44e50.
He, Q., Li, W., Fan, X., 2011. Estimation of driver's fatigue based on steering wheel.
Eng. Psychol. Cognitive Ergon. 145e155.
Horne, J.A., Reyner, L.A., 1995. Sleep related vehicle accidents. BMJ Clin. Res. Ed. 310
(6979), 565e567.
Hosking, S., 2006. The Relationship between Driving Performance and the Johns
Drowsiness Scale as Measured by the Optalert System. Monash University Ac-
cident Research Centre (MUARC).
International Standards, 1997. ISO 2631-1 Mechanical Vibration and Shock - Eval-
uation of Human Exposure to Whole-body Vibration.
John, M.W., 1991. A new method for measure daytime sleepiness the Epworth
Sleepiness Scale. Sleep 14 (6), 540e545.
Kamdar, B.B., Kaplan, K.A., Kezirian, E.J., Dement, W.C., 2004. The impact of
extended sleep on daytime alertness, vigilance, and mood. Sleep. Med. 5 (5),
441e448. Available at: http://www.ncbi.nlm.nih.gov/pubmed/15341888
[Accessed September 28, 2014].
Kjellberg, A., Wikstrom, B.-O., 1985. Subjective reactions to whole-body vibration of
short duration. J. Sound Vib. 99, 415e424.
Larue, G.S., 2010. Predicting Effects of Monotony on Driver's Vigilance. Thesis-
Queensland University of Technology, p. 309.
Maeda, S., Manseld, N.J., Shibata, N., 2008. Evaluation of subjective responses to
whole-body vibration exposure: effect of frequency content. Int. J. Ind. Ergon.
38 (5e6), 509e515. Available at: http://linkinghub.elsevier.com/retrieve/pii/
S0169814107001667 [Accessed December 9, 2014].
Manseld, N.J., Maeda, S., 2011. Subjective ratings of whole-body vibration for
single- and multi-axis motion. J. Acoust. Soc. Am. 130 (6), 3723e3728. Available
at: http://www.ncbi.nlm.nih.gov/pubmed/22225028 [Accessed May 7, 2013].
Nakayama, O., Futami, T., Nakamura, T., Boer, E.R., 1999. SAE Technical Paper.
Development of a Steering Entropy Method for Evaluating Driver Workload, vol.
724, pp. 5e8. Available at: http://www.sae.org/technical/papers/1999-01-0892.
Ramaekers, J.G., Robbe, H.W.J., O'Hanlon, J.F., 2000. Marijuana, alcohol and actual
driving performance. Hum. Psychopharmacol. 15 (7), 551e558.
Tewari, V.K., Prasad, N., 2000. Optimum seat pan and back-rest parameters for a
comfortable tractor seat. Ergonomics 43 (2), 167e186. Available at: http://www.
ncbi.nlm.nih.gov/pubmed/10675057 [Accessed May 15, 2014].
Thiffault, P., Bergeron, J., 2003. Monotony of road environment and driver fatigue: a
simulator study. Accid. Anal. Prev. 35 (3), 381e391. Available at: http://www.
ncbi.nlm.nih.gov/pubmed/12643955.
Verster, J.C., Roth, T., 2011. Standard operation procedures for conducting the on-
the-road driving test, and measurement of the standard deviation of lateral
position (SDLP). Int. J. Gen. Med. 4, 359e371.
Verster, J.C., Bervoets, A.C., De Klerk, S., Vreman, R.A., Oliver, B., Roth, T.,
Brookhuis, K.A., 2014. Effects of alcohol hangover on simulated highway driving
performance. Psychopharmacology 231 (15), 2999e3008.
Vicente, J., Laguna, P., Bartra, A., Bailon, R., 2011. Detection of driver s drowsiness by
means of HRV analysis. Comput. Cardiol. 38, 89e92.
Wierwille, W., Kirn, C.L., 1994. Research on Vehicle-based Driver Status/perfor-
mance Monitoring; Development, Validation, and Renement of Algorithms for
Detection of Driver Drowsiness. Report no. DOT HS 808 247. NTIS, Springeld,
VA, p. 22161.