Accident Analysis and Prevention xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Accident Analysis and Prevention

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

A fundamental experimental approach for optimal design of speed bumps

⁎
A. Hakan Lava, , Ertugrul Bilginb, A. Hilmi Lavb
a
Koc University, Istanbul Turkey
b
Istanbul Technical University, Istanbul Turkey

A B S T R A C T

Speed bumps and humps are utilized as means of calming traffic and controlling vehicular speed. Needless to
say, bumps and humps of large dimensions in length and width force drivers to significantly reduce their driving
speeds so as to avoid significant vehicle vertical acceleration. It is thus that this experimental study was con-
ducted with the aim of determining a speed bump design that performs optimally when leading drivers to reduce
the speed of their vehicles to safe levels. The first step of the investigation starts off by considering the following
question: “What is the optimal design of a speed bump that will − at the same time − reduce the velocity of an
incoming vehicle significantly and to a speed that resulting vertical acceleration does not jeopardize road safety?
The experiment has been designed to study the dependent variables and collect data in order to propose an
optimal design for a speed bump. To achieve this, a scaled model of 1:6 to real life was created to simulate the
interaction between a car wheel and a speed bump. During the course of the experiment, a wheel was accelerated
down an inclined plane onto a horizontal plane of motion where it was allowed to collide with a speed bump.
The speed of the wheel and the vertical acceleration at the speed bump were captured by means of a Vernier
Motion Detector.

The optimal width was found by testing speed bumps with widths
ranging from 1 to 10 cm with 1 cm increments, while keeping height
constant at 1 cm. The optimal height was determined by using test
heights ranging from 0.7-3.4 cm with 0.3 increments, while the width
was kept constant at 5 cm. The simultaneous use of motion detectors
and computer software enabled the measurement of two important
mechanical factors affecting the performance of speed bumps: namely
velocity reduction, which is the measure of decrease in velocity of the
wheel, and the peak vertical acceleration, which is believed to be one of
the main factors affecting passenger comfort. The gathered raw data
were then refined for each speed bump and a Speed Bump Coefficient of
Performance (SBCP) was introduced to compare the speed bumps. The
study focusing on optimal width determination provided evidence of
the existence of a parabolic relationship between the width of the speed
bump and the reduction in velocity. Within these conditions, the op-
timal speed bump dimensions were found to be 5.0 cm in width and
2.8 cm in height, scaled up to 30.0 cm in width and 16.8 cm in height at
real conditions.
1. Introduction
Traffic engineers use a variety of methods in their efforts to control
vehicle speeds and to prevent accidents caused by speeding. One of the
most common methods used is the placement of traffic calming devices,
of which speed bumps and humps are popular types. These speed
bumps and humps are obstacles that are placed across various roads to
reduce the velocity of incoming vehicles.
The utilization of speed bumps allows both for a steady flow of
traffic on private roads and vehicular speed management in parking lots
(Kramberger 2010). Despite their popular use, however, little con-
sideration has been paid for the design of these devices and this hap-
hazard approach oftentimes leads to incorrect constructions. It is
especially important that speed bumps are built correctly, as even the
slightest deficiency may lead to such problems as damage to vehicles;
strong jolts caused by excessive vertical acceleration, which have the
potential of causing passenger discomfort, and, in severe cases, injuries;
and insufficient velocity reduction due to inefficient designs, etc.
(Hessling et al., 2008). Hence, it is critical and necessary to investigate
the optimal design for a speed bump that will both reduce the velocity
of an incoming vehicle while also keeping peak vertical acceleration at
an acceptable level, since severe vertical acceleration levels also en-
dangers traffic safety.
Because the primary interactions being observed are collisions and
accelerations caused by either gravity or macroscopic object interac-
tion, this study is also closely related to the field of classical mechanics.
More specifically, the Principles of Momentum and Newton’s Laws of

⁎
Corresponding author.
E-mail address: alav15@ku.edu.tr (A.H. Lav).

http://dx.doi.org/10.1016/j.aap.2017.05.022
Received 14 February 2017; Received in revised form 23 May 2017; Accepted 25 May 2017
0001-4575/ © 2017 Elsevier Ltd. All rights reserved.

Please cite this article as: Lav, A.H., Accident Analysis and Prevention (2017), http://dx.doi.org/10.1016/j.aap.2017.05.022
A.H. Lav et al.
Motion are key to predicting an overview of the interaction. This issue
was investigated here with a study based on an experimental approach.
As a first step, a basic model for the interaction between a vehicle wheel
and a speed bump was created. This model supports experimental
procedures by providing a precise projection and stable contact be-
tween the projectile and the speed bump. A scaled environment was
constructed as a means of testing various speed bumps under con-
trollable conditions. The horizontal velocity and vertical acceleration of
the wheel during the interaction were monitored to measure both the
velocity reduction and peak vertical acceleration of the wheel. The
collected data were analyzed to assign an efficiency level to each speed
bump tested, and to determine the most optimal speed bump.
2. Literature review
These speed humps and bumps have been playing vital roles as
traffic calming devices for the past several decades and this utilization
has catalyzed the publishing of a wide number of studies on this subject,
a development that began to significantly increase in the mid 1980′s
and that continues to date. These investigations have been published as
scientific articles in many journals, as reports, and as MSc and PhD
theses, etc. While many of these scientific research studies on this
subject have been conducted globally, most have originated thus far
from the United States of America, Canada, Great Britain, and Australia.
Some of these studies have concluded that, for instance, as com-
pared to speed bumps, speed humps provide a much smoother and
more gradual flow of vehicle speeds (Namee and Witchayangkoon,
2011). Other studies have dealt with various aspects of speed humps
and bumps. Among these were the first attempts at determining the
kinds of designs needed for speed humps (Watts and Seminol Profiles).
This design, which has become the most common speed hump design,
was created by Watts in 1973 and has been termed the ‘Watts Profile’ or
circular hump (Weber, 1998). As a response to this design, concerns
over the abruptness of Watts Profile humps at higher speeds in Semi-
nole County, Florida resulted in the creation of a design called the
Seminole Profile Hump. These latter humps are similar in profile to
trapezoidal humps with the difference that they are circular rather than
straight ramps (Nicodemus, 1991).
Further concerns about the designs of these early humps led to ef-
forts that began in the mid 1990s towards the creation of a design of a
profile to serve as an optimum speed hump. To this end, the maximum
acceleration of the driver at a speed specified by the hump design was
selected as the objective function for the optimization of the humps. In
addition to this aim, the maximum acceleration of the driver at a speed
that exceeded that specified by the hump design served as the objective
function for the optimization of the suspension. These efforts led to a
method proposed to create an optimum hump design that would
minimize these two objective functions (Maemori, 1995).
The Transportation and Road Research Laboratory (TRRL) of Great
Britain conducted extensive laboratory research on test tracks in-
vestigating the outcomes of vehicles travelling at various speeds over
various hump size and shapes. In these studies, the British researchers
concluded that the ideal speed hump is 12 feet long and 3–4 inches
high. It was as a result of these experiments that the parabolic shaped
speed hump used throughout Europe, Australia and New Zealand were
developed by TRRL (Smith and Giese, 1997).
A Canadian MSc thesis entitled, Towards a Canadian Standard for
The Geometric Design of Speed Humps was developed with the aim of
determining a set of geometric design standards for speed humps in
Canada. In this study, accelerations were recorded for vehicles travel-
ling over several of on-road Watts and Seminole Profile wooden speed
humps, and those records were compared to discomfort criteria. This
study concluded that precisely designed speed humps could reduce
automobile and heavy vehicle speeds, produce acceptable levels of
discomfort for vehicle occupants, result in no vehicle damage, max-
imize overall road safety, and minimize vehicle noise (Weber, 1998).
2
Accident Analysis and Prevention xxx (xxxx) xxx–xxx
Another study found that a lower ramp of speed humps having spe-
cific other profiles lead the vehicle to gently cross the hump at a straddle,
which at the same time forces the driver to reduce the speed to a degree
at which a catastrophic jump over the road pavement will be avoided.
This study’s analysis of drivers’ responses to installation of such raisings
shows that users prefer larger humps since they are more comfortable in
terms of jolts perceived inside the car (Pau and Angius, 2001).
The year 2008 marked the introduction of a new approach to the
shape optimization of speed humps. Based on multi-objective genetic
optimization of the hump profile, this approach takes into consideration
the separation phenomenon that occurs when the front tires of the ve-
hicle momentarily lose contact with the road surface. The results of the
same research have shown that longer humps do not necessarily perform
better in terms of speed reduction, a conclusion that is somewhat dif-
ferent from those previously reported by others (Ardeh et al., 2008).
In 2009, a similar study of shape optimization of speed hump based
on a multi-objective optimization technique called “GAM,” which was
first proposed by Gembicki in 1974, was also carried out. The main
objective of the study was to minimize vehicle response functions below
the hump crossing speed limit and to maximize them above the speed
limit. This study concluded that the commonly used sinusoidal, cir-
cular, and parabolic profiles do not perform well in terms of hump ef-
ficiency requirements and that polynomial and trapezoidal profiles
tended to produce better results. (Başlamışlı and Ünlüsoy, 2009).
While frequent exposure to speed humps has been found to cause
back discomfort in both drivers and passengers, and despite the im-
portance of speed humps as a traffic device, a review of the literature
demonstrates that very little research has been carried out to in-
vestigate their design. This lack of suitable study shows that there is a
real need to identify an effective, suitable design that is less hazardous
to cars driving at an acceptable speed on residential streets (Zainuddin
et al., 2014).
In 2015, two different investigations towards the design of a
“Simple Harmonic Hump” and “Polynomial Hump” were published. In
these studies, a ride comfort diagram was created by using the
“MATLAB” simulation. This work led to the design of simple harmonic
hump and polynomial hump for any desired hump crossing speed in the
range of 5–30 km/h (Hassan, 2015a, 2015b).
2.1. Flat top trapezoidal humps and raised pedestrian crosswalks (Speed
tables)
Another study (Vejdirektoratet, 1991) concluded that flat top tra-
pezoidal humps are particularly useful when combined with pedestrian
crossings. In response to this, Australia prefers to use these trapezoidal
humps, and does so even on bus routes (Jarvis, 1992).
A study completed in 2009 investigated the optimization of cross-
section dimensions of raised pedestrian crosswalks (RPC) in Qazvin,
Iran. This research concluded that the second ramp length ranks as the
highest impact variable. This is followed by the first ramp length, top
flat crown length, before-RPC spot speed, height and street width, in
order of magnitude (Mohammadipour and Alavi, 2009).
2.2. Optimization of speed hump spacing
A further study considered the optimization of speed hump spacing
(Abaza et al., 2012).
2.3. Effects of Speed Humps in Reducing Various Type of Crashes,
Pedestrian Injuries and Fatalities in Residential Areas
A study carried out in Oakland, California with the purpose of de-
termining the effectiveness of speed humps in reducing child pedestrian
injuries in residential areas concluded that the presence of block of
speed humps within the boundaries of a neighborhood significantly
reduces (53% to 60% reduction) the odds of being struck and injured by
A.H. Lav et al.
an automobile (Tester et al., 2004).
Another study demonstrated that since motor vehicle speeds were
low on pedestrian crossings irrespective of distance between cushion
and crossing, distances of two-car lengths (app. 10 m) between such
bumps had shown positive effects concerning mobility for vulnerable
road users. The study thus suggested the use of this distance when in-
stalling speed cushions. These researchers also emphasized that a
longer distance would also make it easier − especially for children and
the elderly − to distinguish if a driver was going to stop or not, a factor
that increases both safety and mobility (Johansson et al., 2011).
Another investigation recommended that median separation, speed
humps, and road lighting be incorporated into road design and con-
struction, as such inclusion significantly reduces pedestrian accidents
and the likelihood of hit-and-run incidents (Aidoo et al., 2013).
The performance of 13 different type of countermeasures, including
the speed humps that had been installed in New York City during the
past 20 years by New York City Department of Transportation
(NYCDOT), were also evaluated. This study concluded that speed
humps have only a limited impact on crashes, including injurious and
fatal crashes, which does not meet with their expected performance
(Chen et al., 2013).
2.4. Studies of design of speed humps utilized in rural areas
Although speed humps and speed tables have been shown to be
effective in reducing vehicle speeds, the majority of these studies have
been conducted in large urbanized areas. Little information as to how
these devices function in smaller or more rural cities is available in the
literature. Smith et al’s study was one of the rare and useful examples
for their use in such kinds of residential areas. The results of this study
led the authors to conclude that speed humps − especially speed tables
− effectively reduce the number of the vehicles exceeding the speed
limits in those particular areas (Smith et al., 2002).
Practical applications have also demonstrated that on highways
located in mountainous areas, the use of appropriate speed control
humps situated on the downhill slopes can lead drivers to actively
control their vehicular speed, improve the operational safety of large
vehicles, and effectively reduce traffic accidents (Jianbo et al., 2010).
2.5. Studies on the design of bicycle and motorcycle friendly humps
Early uncertainty about the effects of humps on vehicle stability
initially restrained the use of humps. Despite this, time has shown that
the use of humps has relatively little negative effect on vehicles, in-
cluding even motorcycles, and this is offset by their effectiveness in
controlling vehicle speed (Kjemtrup and Herrstedt, 1992).
Speed hump can also be especially designed to be bicycle friendly.
Local authorities can adjust their use for cyclists by changing their
hyperbolic design, tapering the edges, and providing a one meter gap at
the curb lane side (Jacobs, 2004).
2.6. Studies on the design of emergency vehicle and public transportation
friendly speed humps
One study included the search for ways to solve those problems
arising from the passing of public transportation and emergency ve-
hicles over speed humps. This study of speed bumps located in several
countries over the face of the globe demonstrated that if speed bumps
are designed and executed correctly such problems as vehicle damage
or driver’s loss of control can be effectively eliminated (Zaidel et al.,
1992).
2.7. Predicting the environmental effects of speed humps in residential areas
According to the preliminary findings of a research study conducted
in 2010, a significant increase in tail pipe emissions was noted after the
3
Accident Analysis and Prevention xxx (xxxx) xxx–xxx
speed hump was crossed. But while this is obviously a negative im-
plication and a setback for their use, it should not be considered as an
isolated variable. Any assessment of speed bump efficacy should con-
sider total relevant implications, such as the speed reduction achieved
and the resulting increase in safety levels (Silva et al., 2010).
2.8. Dynamic simulation of speed Humps/Bumps to investigate the dynamic
response of vehicles
In 2006 and 2007 two similar MSc theses prepared by Gürer and
Sözen respectively considered the dynamic simulations of speed bumps
and speed humps under traffic loading, and their vibration effects on
vehicles. These studies measured the vertical accelerations of vehicles
while passing over a speed hump, derived the frequency analysis of the
vibration on the vehicle, and obtained the frequencies of vibration at
different locations (Gürer, 2006). In these studies, the researchers used
the Matlab/Simulink method to analyse and compare/contrast wheel
acceleration, vehicle acceleration, the change of the location of the
wheel, and change of the location of the (Sözen, 2007).
Other researchers investigated the dynamic responses of vehicles on
different types of speed humps using the kind of multibody simulation
package which is popular in the automotive industry and found that the
dynamic performance of the recommended two sinusoidal type profiles
had significantly better performance than that of conventional humps;
i.e. parabolic, sinusoidal and four types of flat-topped humps (speed
table). The authors expected the results of their study to be useful in the
future geometric design of speed humps (Molan and Kordani, 2014).
2.9. Effects of speed humps on pavement condition
Yet another study looked at the effects of speed bumps on the pa-
vement itself and concluded that the presence of improper speed humps
significantly and negatively affects pavement condition by reducing the
PCI (Pavement Condition Index) value (Bekheet, 2014).
3. The model
In this study, a model of the interaction between a vehicle and a
speed bump was created by replicating some properties involved in
collisions. This model only looks at one wheel, for such an investigation
is sufficient in providing data about the speed bump performance.
The interaction can be described according to three main phases:
pre-collision, during collision, and post-collision, as follows:
3.1. Pre-Collision
In this phase, the wheel is moving towards a speed bump as shown
in Fig. 1. Several factors impact the motion of the wheel in this part of
the investigation.
Firstly, the wheel displaces over time parallel to the plane of mo-
tion. This means that it has a certain mass (m), has a velocity (v1),
kinetic energy, and momentum (P). The momentum is equal to the
product of the mass of the object and its velocity;
Momentum (P ) = mv (1)
According to Newton’s 2nd Law of Motion, the net force acting on
an object is equal to the rate of change in momentum within the di-
rection of the force. The law can be represented as a derivation equa-
tion;
dP dmv
F= =
dt dt (2)
Given that the mass of the wheel (m) remains constant throughout
the motion, it can be excluded from the derivation by the constant
factor rule in differentiation.
A.H. Lav et al.
Fig. 1. System Prior to Collision.
dP ΔP dv Δv
F= = =m =m
dt Δt dt Δt
Lastly, by refining Eqn. (3), a final equation can be deduced for the
mechanical variables;
F Δt = ΔP = mΔv
Eqn. (4) states that if a force was to act on the wheel during a time
interval, the wheel’s momentum and hence velocity would change. In
this study, Eqn. (4) is vital as it explains the relationship between the
speed bump design and change in the mechanical properties of the
wheel (further explained in System During Collision).
Additionally, wheel slippage has been controlled, and rolling re-
sistance has been neglected in this study.
3.2. System during collision
• Although the wheel and the speed bump are elastic objects, in rea-
lity they can be assumed to be rigid as their elasticity is negligible.
This assumption enables us to provide a model of the collision as
given in Fig. 2.
• As the two objects come into contact, the wheel starts to exert a
force on the speed bump. According to Newton’s 3rd Law of Motion,
the speed bump exerts an equal and opposite force on the wheel.
The direction and the magnitude of the forces exerted vary with
time over the duration of the collision. Nevertheless, a representa-
tion of the averages of the forces exerted is made by FC in Fig. 2.
• The speed bump is stationary and attached to the ground, and there-
fore, the forces acting on it have no effect and will not cause the speed
Accident Analysis and Prevention xxx (xxxx) xxx–xxx
bump to move. On the other hand, because the wheel is not stationary,
(3) the forces acting on the wheel will cause a change in its horizontal
velocity and vertical acceleration. The force exerted by the speed bump
can be investigated in two component force vectors: FC┴ and FC║.
(4) • FC║is the component vector which is parallel and opposite to the
momentum of the wheel. As already deduced in Eqn. (4), the overall
force of FC║ during the time of contact decreases the momentum of
the wheel as the direction of the force opposes the wheel’s mo-
mentum. The reduction in the net momentum acts to decrease the
velocity of the wheel. Hence, FC║ is important as it is the force
governing the velocity reduction.
• FC┴ is the component vector which is normal to the motion of the
wheel. Although FC┴ does not have any effect on the horizontal
momentum, it is the force that causes change in the vertical axis of
the wheel. Gravitational force is exerted on the center of the wheel
throughout its motion. During the time of collision, FC┴ is the force
acting on the wheel opposite to W. However, while W can be as-
sumed constant, FC┴ changes significantly over a short period of
time. Therefore, during collision, the drastic change in the net ver-
tical force will affect the vertical acceleration of the wheel. The
vertical acceleration of the wheel has been observed in this study
because the peak vertical acceleration is one of the important factors
affecting passenger comfort. Since the duration of passing the speed
bump is roughly constant, the peak acceleration indicates the degree
of jolt that occurs in the interaction between the wheel and the
speed bump.
• Additionally, the friction of the speed bump and back slope accel-
eration by the wheel are also accounted for in this study (Fig 3).
4
A.H. Lav et al.
Fig. 2. System During Collision.
3.3. Post-Collision system
This phase is similar to the pre-collision system. However, following
the collision the wheel moves through the plane of motion slower than
it had prior to the collision. Eqn. (5) can then be deduced by con-
sidering the final velocity (v2) measured in this phase:
FΔt = ΔP = m(v_2-v_1) (5)
The final velocity will always be smaller than the initial velocity as
FC║ acts opposite to the direction of motion. For this reason, the change
in overall velocity will always be a reduction in velocity.
From the theory, two dependent variables are deduced, as follows:
3.3.1. Velocity reduction
This variable is the change in velocity caused by the collision. As
seen in Eqn. (4), this factor is proportional to the average parallel force
exerted during the time interval of collision. Therefore, velocity re-
duction depends on the magnitude and duration of the force exerted
while the wheel collides with the speed bump.
3.3.2. Peak vertical acceleration
Humans are not disturbed by high accelerations while travelling in a
vehicle (e.g. sports cars accelerating at great rates); however, a sudden
change in acceleration causes human discomfort as a result of the jolting
effect (e.g. bumpy car rides). When a wheel passes over a speed bump, the
peak vertical acceleration should indicate an accurate estimate of the jolt
that was experienced by the wheel. This variable is a means of optimizing
speed bump designs to be more passenger-friendly and comfortable.
The dependent variables are affected by the following independent
variables:
3.3.3. Shape of the speed bump
The shape of a traffic calming device is determined by three distinct
elements: curvature, height, and width parallel to the traffic flow. The
term speed bump is used for vertical traffic calming devices that are
5
Accident Analysis and Prevention xxx (xxxx) xxx–xxx
parabolic. This means that this investigation into the design ignores the
curvature of the speed bump, for its parabolic shape is the determining
factor of a what defines a speed bump (Fig 4).
In order to achieve the greatest reduction in velocity reduction and
the least vertical acceleration, the height and width of the speed bump
must be optimized for the best ratio of force components. The experi-
ment conducted within this investigation is an attempt to find the op-
timal values.
3.3.4. Initial velocity
The initial collision velocity of the wheel affects the force exerted.
However, this variable is controlled in this experiment to focus on the
shape of the speed bump.
4. Experiment setup
The experiment conducted in this study was designed to study the
dependent variables and collect data in order to propose the optimal de-
sign for a speed bump. Here, a wheel is accelerated down an inclined plane
onto a horizontal plane of motion where it collides with a speed bump.
The wheel in the experiment is a 1/6th scale replica of an actual
wheel. It has a diameter of 1.05 × 10–1 ± 0.05 × 10–1 m and a mass
of 1.550 × 10-1 ± 0.001 × 10–1 kg. To carry out the experiment a
wheel of such size and mass was available and it satisfied acceptable
accuracy and precision. The curvature of each speed bump was para-
bolic, which means their highest point was located midway between the
width. Finally, Vernier Motion Detectors were used to measure the
horizontal velocity and vertical acceleration of the wheel. Fig. 5 re-
presents the setup of the experiment.
As seen in Fig. 5, the Vernier Motion Detector was placed on the
same level of motion as the wheel, or above the speed bump. The device
acts by sending ultrasonic waves and analyzing the reflections bouncing
back from the objects in the path of the waves to determine the distance
of a particular object. To meet the purposes of this experiment, the data
need to be collected very quickly; to achieve this, the data collection
speed was set at 20 Hz or 20 samples per second.
A.H. Lav et al. Accident Analysis and Prevention xxx (xxxx) xxx–xxx

Fig. 3. System After Collision.

The motion detector was connected to the computer via Vernier Lab
Pro. The Lab Pro converts the measurements from the motion detector
into appropriate data which can be processed into Logger Pro 3.6.1
computer software. The simultaneous use of these devices enables the
collection of displacement, horizontal velocity, and vertical accelera-
tion of the wheel.
provides reliable data for the purposes of this experiment. Additionally,
the uncertainties of the results of these tests are used as a measure of
uncertainty for the raw data of the main experiment.
6. Experiment procedure

The analytical approach used in this study focused on acquiring sets

5. Calibration and accuracy tests of data points that could lead to a determination of the possible func-
tions, which can then be used to predict the optimal dimensions for
The calibration tests have shown that the Vernier Motion Detector speed bumps. The second purpose of the study was to discover the most

Fig. 4. Theoretical Change in Force of Collision (FC) By The Change of Speed Bump Design.

6
A.H. Lav et al.
Fig. 5. Experiment Setup.
appropriate functions that could describe the relationships between
independent and dependent variables. The width and height of speed
bumps were investigated separately and their results were combined to
find the optimal speed bump dimensions.
A scale of 1:6 scale was used to reflect both the weight and the
velocity of the wheel through a dimensional analysis and similarity
(Geometric and Kinematic Similarity).
Individual speed bumps were constructed for each trial and the
wheel was placed at the release point as shown in Fig. 5. As the wheel
approached the speed bump, the data collected from Logger Pro 3.6.1
software was processed to obtain the change in velocity and maximum
vertical acceleration of the wheel. Examples of the positions and ver-
tical acceleration of the wheel travelling over the speed bump are given
in Fig. 6 and Fig. 7, respectively.
7. Test results
The raw logger pro data was interpreted and then presented in a
table format. The reduction in velocity (-Δv) and the peak acceleration
data were extracted from the Logger Pro software and these are given in
Tables 1 and 2. It is important to note that the investigations on width
(Table 1) and height (Table 2) are shown separately. Also, the value of
the dimensions of the speed bumps were recorded in centimeter nota-
tion (cm, 10–2 m) for easier comprehension; the uncertainties for the
dimensions can be assumed negligible since they are relatively ex-
tremely small compared to the values themselves (10–5 m uncertainty
for a value of around 10–2 m, nearly a difference of 103 m). Repeated
trials were conducted to reduce any random uncertainties that might
have occurred during testing. The average of 3 trials is accepted as the
value of the measurement. The constant height dimension for the tested
widths was 1 cm, while the constant width dimension for the tested
heights was 5 cm. This selection was based on the fact that these scaled
height and width fall in the ranges published in such manuals as
Pennsylvania’s Traffic Calming Handbook and DDOT (2010) Speed
Hump Request Procedures and Engineering Guidelines.
7
Accident Analysis and Prevention xxx (xxxx) xxx–xxx
8. Analysis of the test results
The data collected from the study are presented in figures in order
to analyse the relationships between the dependent and independent
variables, and are given as follows:
As indicated in Fig. 8, there is a parabolic relationship between
velocity reduction and width of the speed bump. This shows that there
is an optimal width dimension that provides the highest degree of ve-
locity reduction.
During testing, it was observed that the back-slope acceleration
played a major role in forming a parabolic relationship between these
two variables. That velocity reduction increases as the width of the
speed bump increases is a logical outcome since the retarding forces
acting on the wheel would be exerted for a longer duration. However,
as the width of the speed bump increased after a certain point, the
intensity of the collisions decreased and the wheel was able to attain a
greater amount of velocity from rolling down the back slope of the
speed bump. This outcome explains the lesser values of velocity re-
duction for longer speed bumps. This argument can be supported by
comparing the maximum velocity reduction and back slope velocity
increase caused by the speed bump as indicated in(Table 3) Fig. 9.
As seen in Fig. 10, although the maximum velocity reduction re-
mained roughly the same for longer speed bumps, the velocity gained
from the back slope increased quite significantly. This would explain
the parabolic relationship in Fig. 8 and why longer speed bumps failed
to provide the most velocity reduction.
Fig. 11 shows a correlation coefficient of 0.05, thus appearing to be
little or no correlation between width and peak vertical acceleration. It
is possible that since the height of the speed bump remained constant,
the wheel could only be elevated to a certain level which limited the
possibility of a more abrupt or controlled freefall. Hence, the peak ac-
celeration remained roughly the same for all widths. This assumption
can be supported by analyzing the height and peak vertical acceleration
data to check if there is any correlation between those two variables.
As shown in Fig. 12, a correlation coefficient of 0.99 indicates a
A.H. Lav et al. Accident Analysis and Prevention xxx (xxxx) xxx–xxx

Fig. 6. Interpretation of Raw Vertical Position Graph of 5 cm Width, 1.9 cm Height Speed Bump.

Fig. 7. Interpretation of Raw Vertical Acceleration Graph of 5 cm Width, 1.9 cm Height Speed Bump.

Table 1
Width Data with respect to Reduction in Horizontal Velocity and Peak Vertical Acceleration.

Width of Speed Bump Reduction in Horizontal Velocity ms−1/ ± 0.01 Peak Vertical Acceleration ms−2/ ± 0.1
cm/(negligible)
Trial 1 Trial 2 Trial 3 Avg. Trial 1 Trial 2 Trial 3 Avg.

1 0.10 0.10 0.11 0.10 6.5 6.2 6.2 6.3

2 0.12 0.15 0.13 0.13 6.3 6.3 6.2 6.3
3 0.17 0.17 0.17 0.17 6.6 6.7 6.6 6.6
4 0.19 0.17 0.18 0.18 6.5 6.4 6.0 6.3
5 0.20 0.19 0.19 0.19 6.2 6.2 6.3 6.2
6 0.19 0.19 0.18 0.19 6.1 6.5 6.2 6.3
7 0.18 0.18 0.18 0.18 6.4 6.3 6.3 6.3
8 0.15 0.15 0.18 0.16 6.8 7.1 7.0 7.0
9 0.12 0.14 0.15 0.14 6.0 6.1 5.9 6.0
10 0.12 0.13 0.12 0.12 6.5 6.5 6.3 6.4

8
A.H. Lav et al. Accident Analysis and Prevention xxx (xxxx) xxx–xxx

Table 2
Height Data with respect to Reduction in Horizontal Velocity and Peak Vertical Acceleration.

Height of Speed ump Reduction in Horizontal Velocity ms−1/ ± 0.01 Peak Vertical Acceleration ms−2/ ± 0.1
cm/(negligible)
Trial 1 Trial 2 Trial 3 Avg. Trial 1 Trial 2 Trial 3 Avg.

0.7 0.13 0.13 0.14 0.13 5.8 5.5 6.0 5.8

1.0 0.19 0.18 0.18 0.18 5.2 5.5 5.5 5.4
1.3 0.20 0.20 0.19 0.20 5.1 5.1 5.2 5.1
1.6 0.23 0.23 0.23 0.23 5.0 5.0 5.0 5.0
1.9 0.26 0.27 0.27 0.27 4.8 4.5 4.1 4.5
2.2 0.32 0.32 0.31 0.32 4.4 4.0 4.2 4.2
2.5 0.38 0.38 0.38 0.38 4.3 4.3 4.3 4.3
2.8 0.39 0.47 0.43 0.43 4.0 3.9 3.6 3.8
3.1 N/Aa
3.4

a
The wheel could not pass over the speed bump.

Fig. 8. Width Against Reduction in Horizontal Velocity.

Table 3
First Trial Data of Width with respect to Maximum Reduction in Horizontal Velocity and Velocity Gain from Back Slope.

Width of Speed Bump Maximum Reduction in Horizontal Velocity Velocity Gain from Back Slope
cm/(negligible) ms−1/ ± 0.01 ms−1/ ± 0.01

1 0.10 0.00
2 0.14 0.02
3 0.18 0.01
4 0.22 0.03
5 0.23 0.03
6 0.24 0.05
7 0.22 0.04
8 0.24 0.09
9 0.22 0.10
10 0.23 0.11

9
A.H. Lav et al. Accident Analysis and Prevention xxx (xxxx) xxx–xxx

Fig. 9. Raw Horizontal Velocity Graph of 5 cm Width, 1.6 cm Height Speed Bump.

Fig. 10. Comparison of Reduction in Horizontal Velocity, Maximum Reduction in Horizontal Velocity and Velocity Gain from Back Slope.

strong positive correlation between height and velocity reduction.
In Fig. 13, a correlation coefficient of −0.98 suggests that there is a
strong negative correlation between height and peak vertical accel-
eration. It seems that the correlation of velocity reduction and peak
vertical acceleration with regard to the same independent variable
(height) is of an opposite nature (positive and negative), and therefore
it can be assumed that these two variables are also inversely correlated
in the context of the study on height. This assumption has been more
thoroughly investigated in “Full-scope Analysis” with respect to the
hypothesis made about the effect of speed bump dimensions on the
components of the force exerted on the wheel in Fig. 4.

10
A.H. Lav et al. Accident Analysis and Prevention xxx (xxxx) xxx–xxx

Fig. 11. Width Against Peak Vertical Acceleration.

Fig. 12. Height Against Reduction in Horizontal Velocity.

11
A.H. Lav et al. Accident Analysis and Prevention xxx (xxxx) xxx–xxx

Fig. 13. Height Against Peak Vertical Acceleration.

Fig. 14. Width of The Speed Bump and Component Forces.

9. Full-scope analysis
The previous section hypothesized that there might be a relation-
ship between the shape of the speed bump and the force components of
the force exerted on the wheel. Based on this assumption, then the
dependent variables (velocity reduction and peak vertical acceleration)
would be inversely correlated because the magnitudes of the compo-
nent forces governing each of the two dependent variables are nega-
tively correlated.
The test results on width did not, however, provide the kind of
correlation described in the hypothesis. It may be that the independent
variable width has either slight or no effect on the force components of
the force exerted by the speed bump. Fig. 14 displays this outcome.
The investigation on height provided evidence of an inverse corre-
lation between the dependent variables with respect to height. This
relationship might support the hypothesis as presented in Fig. 15.
Furthermore, as shown in Table 6, it was observed that the wheel
could not pass over the speed bump after a certain height. This can also

12
A.H. Lav et al. Accident Analysis and Prevention xxx (xxxx) xxx–xxx

Fig. 15. Height of The Speed Bump and Component Forces.

Fig. 16. Height at which The Wheel Fails to Pass the Speed Bump.

be shown as an evidence that supports the hypothesis on force com- Ideally, the optimal speed bump would provide the most velocity
ponents, since the hypothesis can be used to explain the existence of reduction and the least peak vertical acceleration. Therefore, the per-
such an observations as seen in Fig. 16. formance of a speed bump is positively proportional to the velocity
reduction and inversely proportional to the peak vertical acceleration;
9.1. Speed bump coefficient of performance (SBCP)
SpeedBumpPerformance ∝ VelocityReduction
The aim of this study is to determine the design of an optimal speed 1
bump with respect to both velocity reduction and peak vertical accel- SpeedBumpPerformance ∝
PeakVerticalAcceleration (6)
eration. The Speed Bump Coefficient of Performance (SBCP) is a value
introduced to compare the performance of both of these variables as Combining these two relationships;
VelocityReduction
they affect speed bumps. SpeedBumpPerformance ∝ PeakVerticalAcceleration (7)

13
A.H. Lav et al. Accident Analysis and Prevention xxx (xxxx) xxx–xxx

Table 4 10. Evaluation of testing

Width of Speed Bump with respect to SBCP and Uncertainty in SBCP.
Width of Speed Bump SBCP/unit value
cm/(negligible)
1 1.5 × 10−2
2 2.1 × 10−2
3 2.6 × 10−2
4 2.9 × 10−2
5 3.1 × 10−2
6 3.0 × 10−2
7 2.8 × 10−2
8 2.3 × 10−2
9 2.3 × 10−2
10 1.9 × 10−2
Table 5
Height of Speed Bump with respect to SBCP and Uncertainty in SBCP.
Height of Speed Bump SBCP/unit value
cm/(negligible)
The speed bump with the dimensions of 5 cm in width and 1.0 cm in
Uncertainty in SBCP/unit value
height was tested twice, in both width and height investigations. The
values of the two trials can be compared to evaluate the accuracy of
± 0.2 × 10−2 testing as represented in Table 6.
± 0.2 × 10−2
It appears that the results for velocity reduction and SBCP lie within
± 0.2 × 10−2
± 0.2 × 10−2 the uncertainty range of the corresponding investigations results;
± 0.2 × 10−2 therefore, it can be assumed that the testing and results of these two
± 0.2 × 10−2 dependent variables were quite accurate.
± 0.2 × 10−2 However, the comparison of peak vertical acceleration data shows
± 0.2 × 10−2
that there is a significant discrepancy between the two results. This
± 0.2 × 10−2
± 0.2 × 10−2 discrepancy is likely an indication of a systematic error because the
repeated trials should have been able to reduce random errors to a point
where the discrepancy would be lower. One possible cause of this error
may be that the curvature of the speed bump in the width investigation
was incorrectly reconstructed in the height investigation. This error can
be eliminated by building a more sophisticated construction system.
Uncertainty in SBCP/unit value

0.7 2.2 × 10−2 ± 0.2 × 10−2 11. Conclusion and further recommendations
1.0 3.3 × 10−2 ± 0.2 × 10−2
1.3 3.9 × 10−2 ± 0.3 × 10−2
1.6 4.6 × 10−2 ± 0.3 × 10−2 The purpose of this study was to determine the optimal design for a
1.9 6.0 × 10−2 ± 0.4 × 10−2 speed bump that can best reduce the velocity of an incoming vehicle the
2.2 7.6 × 10−2 ± 0.4 × 10−2 most, while not causing a vertical acceleration that jeopardizes road
2.5 8.8 × 10−2 ± 0.4 × 10−2 safety. This study found that the optimal dimensions of a speed bump
2.8 11.3 × 10−2 ± 0.6 × 10−2
3.1 N/A*
are 5.0 cm in width and 2.8 cm in height. Keeping in mind the scaled
3.4 environment is of 1:6 ratio, these values are scaled up to:

30 cm width 16.8 cm height

Hence, to compare the tested speed bumps in this investigation, The investigation that focused on optimal width provided evidence
SBCP is defined as; of a parabolic relationship that exists between the width of the speed
VelocityReduction
SpeedBumpCoefficientofPerformance = PeakVerticalAcceleration (8) bump and velocity reduction. However, it failed to support the hy-
Example pothesis on force components. On the other hand, the investigation that
For Trial 1 of 1 cm width, 1 cm height speed bump; focused on optimal height supported the hypothesis by presenting both
VelocityReduction 0.10
SBCP = PeakVerticalAcceleration = 6.5 = 0.015 a negative correlation between dependent variables and an explanation
Uncertainty (Percentage uncertainties are added); as to why the wheel could not pass over some of the speed bumps.
The main limitation in this experiment was related to the ability to
0.01 0.1 control the initial velocity of the wheel. Because the apparatus used to
± SBCP = ⎛ + ⎞ x 0.015 = ± 0.002
⎝ 0.10 6.5 ⎠ project the wheel had an uncontrollable friction, a consistent collision
velocity could not be established for every trial. This limitation can be
The full SBCP; improved by using a smoother apparatus. Another limitation was the
use of only one projection speed for the wheel. A more reliable in-
1.5 × 10−2 ± 0.2 × 10−2 unit value vestigation could be achieved if the speed bumps are tested for a variety
of initial velocities.
Although significant manuals like MUTCD (2009) and SCDOT
(2006) focus on the placement and shape of speed hump and bump
9.2. Optimal dimensions signs, other manuals such as DDOT (2010) and Pennsylvania’s Traffic
Calming Handbook (2012) only briefly mention recommended ranges
The optimal dimensions for a speed bump found by modelling the of height and width of speed humps. Therefore, the scaled height and
SBCP values for width and height are given in Tables 4 and 5 as width selected in this basic experimental study complies with the
functions and calculating the values that yield the greatest SBCP within ranges given in those above-mentioned manuals.
the range of the test data (Figs. 17 and 18). Lastly, possible further investigations relevant to the topic may in-
Calculated Value: 2.8000000 cm clude investigations related to the optimal spacing for speed bumps and
Accepted Value: 2.8 cm the effect of the shape and curvature of the speed bump on their per-
Optimal Speed Bump; formance.
5.0 cm width, 2.8 cm height

Table 6
Average Reduction in Horizontal Velocity, Peak Vertical Acceleration and SBCP comparison for The Two Trials of 5 cm Width and 1.0 cm Height Speed Bump.

The Investigation Reduction in Horizontal Velocity Peak Vertical Acceleration SBCP

ms−1/ ± 0.01 ms−2/ ± 0.1 unit value/ ± 0.2 × 10−2

Width 0.19 6.2 3.1 × 10−2

Height 0.18 5.4 3.3 × 10−2

14
A.H. Lav et al. Accident Analysis and Prevention xxx (xxxx) xxx–xxx

Fig. 17. Width Against SBCP.

Fig. 18. Height Against SBCP.

References
Abaza, O.A., Hussein, Z.S., Malto, I.S., 2012. Optimization of speed control hump spaces.
ITE J. 82 (August (8)), 44.
Aidoo, E.N., Gyimah, R.A., Ackaah, W., 2013. The effect and environmental character-
istics on pedestrian hit-and-Run accidents in Ghana. Accid. Anal. Prev. 53, 23–27.
Ardeh, A.H., Shariatpanahi, M., Bahrami, N.M., 2008. Multiobjective shape optimization
of speed humps. Struct. Multidise. Optim. 2008 (37), 203–214. http://dx.doi.org/10.
1007/s00158-008-0226-7.
Başlamışlı, Ç.S., Ünlüsoy, S.Y., 2009. Optimization of speed control hump profiles. J.
Transp. Eng. 135, 260–269.
Bekheet, W., 2014. Short term performance and effect of speed humps on pavement
condition of alexandria governorate roads. Alexandria Eng. J. 53 (4), 855–861.
Chen, L., Chen, C., Ewing, R., McKnight, C.E., Srinivasan, R., Roe, M., 2013. Safety
countermeasures and crash reduction in new York city-Experience and lessons
learned. Accid. Anal. Prev. 50, 312–322.
DDOT, 2010. DDOT Speed Hump Request Procedures and Engineering Guidelines. pp. 19.
https://comp.ddot.dc.gov/Documents/Speed%20Hump%20Request%20Procedures
%20and%20Engineering%20Guidelines.pdf.
Gürer, M., 2006. Hız Tümseklerinin Taşıt Titreşim Tekniği Açısından İncelenmesi.
https://hdl.handle.net/11527/3810.
handbook, 2012. Pennsylvania’s Traffic Calming Handbook. Pennsylvania Dept. Of
Transportation (109 p).

15
A.H. Lav et al.
Hassan, G.A., 2015a. Car dynamics using quarter model and passive suspension, part III: a
novel polynominal hump. IOSR J. Mech. Civil Eng. (IOSR-JMCE) 12 (1), 51–57.
Hassan, G.A., 2015b. Car dynamics using quarter model and passive suspension, part II: a
novel simple harmonic hump. IOSR J. Mech. Civil Eng. (IOSR-JMCE) 12 (1), 93–100.
Hessling, J.P., Zhu, P.Y., Gupta, M., 2008. Analysis and Synthesis of Speed Limiting.
https://www.npl.co.uk/upload/pdf/hessling_1. pdf.
Jacobs, M.A., 2004. Using traffic calming measures on local residential streets to faciliate
an alternative designated bicycle route. Master Urban Rural Planning. . https://0-
search.proquest.com. divit.library.itu.edu.tr/docview/3050988380?pq-origsite=
summon.
Jarvis, J.R., 1992. An investigation of road humps for use on bus routes. Aust. Road Res.
Board (ARRB) 222.
Jianbo, W., Rummou, X., Bin, C., Cao, H., 2010. The application research on mountain
highway speed-control hump. Int. Conf. Optoelectron. Image Process. 518–521.
http://dx.doi.org/10.1109/ICOIP.2010.69.
Johansson, C., Rosander, P., Leden, L., 2011. Distance between speed humps and pe-
destrian crossings: does it matter? Accid. Anal. Prev. 43 (5), 1846–1851.
Kjemtrup, K., Herrstedt, L., 1992. Speed management and traffic calming in urban areas
in europe: a historical view. Accid. Anal. Prev. 24 (1), 57–65.
Kramberger, A., 2010. Placement of Speed Bump to Be Reviewed. https://
offislandgazette.com/news/story/placement-of-speed-bump-to-be-reviewed.
MUTCD, 2009. Manual on Uniform Traffic Control Devices for Streets and Highways. pp.
862. https://mutcd.fhwa.dot.gov/pdfs/2009/mutcd2009edition.pdf.
Maemori, K., 1995. Alternate optimization of speed control hump for automobiles and
automobile suspension. Jsme Int. J. C-38 (3), 552–557.
Mohammadipour, H.A., Alavi, H.S., 2009. The optimization of the geometric cross-
Section dimensions of raised pedestrian crosswalks: a case study in qazvin. Accid.
Anal. Prev. 41 (2), 314–326.
Molan, A.M., Kordani, A.B., 2014. Optimization of speed hump profiles based on vehicle
dynamic performance modelling. J. Transport. Eng. 140 (8), 1–12.
Namee, S., Witchayangkoon, B., 2011. Crossroads vertical speed control devices:
16
Accident Analysis and Prevention xxx (xxxx) xxx–xxx
suggestion from observation. Int. Trans. J. Eng. Manage. Appl. Sci. Technol. 2 (2),
161–171.
Nicodemus, D.A., 1991. Safe and effective roadway humps —The seminole county profile.
Compendium of Technical Papers. Institute of Transportation Engineerspp. 102–105.
Pau, M., Angius, S., 2001. Do speed bumps really decrease traffic speed: an italian ex-
perience. Accid. Anal. Prev. 33 (5), 585–597.
Sözen, A., 2007. Hız Kesicilerin Trafik Yüklemesi Altındaki Dinamik Simülayonu. https://
tez.sdu.edu.tr/Tezler/TF01052. pdf.
SCDOT, 2006. Traffic Calming Guidelines. South Carolina Dept Of Transportation
(SCDOT) (p.35).
Silva, J.P.C., Seco, A.J.M., Bastos, S.A.M., 2010. Evaluation of the Effects of Road Humps
on Pollution Emissions, 12th WCTR. pp. 1–16. https://www.wctrs.leeds.ac.uk/wp/
wp-content/uploads/abstracts/lisbon/general/02794. pdf.
Smith, D.E., Giese, K.L., 1997. A Study on Speed Humps. https://www.ctre.iastate. edu/
research/roadhump/.
Smith, D.J., Knapp, K., Hallmark, 2002. Speed Impacts of Temporary Speed Humps in
Small Iowa Cities. https://www.researchgate.net/publication/238077510.
Tester, J.M., Rutherford, G.W., Wald, Z., Rutherford, M.W., 2004. A matched case control
study evaluating the effectiveness of speed humps in reducing child pedestrian in-
juries. Am. J. Health 94 (4), 646–650.
Vejdirektoratet, 1991. Road Standarts Part 7?Speed Reducers. Denmark Ministry of
Transport Urban Traffic Areas Report, Road Data Laboratory, Road Standarts
Division, Denmark (ISBN: 8774913484 9788774913481).
Weber, P.A., 1998. Towards a canadian standard for the geometric design of speed
humps. ITE Journal 70 (1), 30–34. https://0-search. proquest. com. divit. library. itu.
edu.tr/.
Zaidel, D.A., Hakkert, A.S., Pistiner, A.H., 1992. The use of road humps for moderating
speeds on urban streets. Accid. Anal. Prev. 24 (1), 45–56.
Zainuddin, N.I., Adnan, M.A., Diah, J., 2014. Optimization of speed hump geometric
design: case study on residential streets in Malaysia. J. Transport. Eng. 140 (3), 1–9.
http://dx.doi.org/10.1061/(ASCE)TE.1943-5436.0000611.