Optimization of Speed Hump Profiles Based on

Vehicle Dynamic Performance Modeling

Amirarsalan Mehrara Molan 1 and Ali Abdi Kordani 2

Abstract: Speed humps are in widespread use around the world. Despite their effective performance in increasing safety, they cause con-
siderable damage to vehicles and discomfort to drivers and passengers. This paper investigates the dynamic response of vehicles on different
Downloaded from ascelibrary.org by Selcuk Universitesi on 02/06/15. Copyright ASCE. For personal use only; all rights reserved.

types of speed humps using a multibody simulation software package that is popular in the automotive industry. Following this evaluation,
two new profiles are recommended to optimize the dynamic performance of speed humps. A series of formulas are also presented to estimate
the dynamic performance of passenger cars on flat-topped and parabolic humps based on vehicle speed, hump dimensions, and driving
behavior while traversing the hump. The results show that, for flat-topped humps, the ramp length (or entrance slope) has the greatest effect
on vehicle performance. It is found that the dynamic performance of the recommended profiles is much better than that of conventional
humps. DOI: 10.1061/(ASCE)TE.1943-5436.0000686. © 2014 American Society of Civil Engineers.
Author keywords: Speed humps; Disadvantages of speed humps; Profile optimization; Driver discomfort; Vehicle damage.

Introduction According to Schlabbach (1997), traffic calming started in 1970

with the installation of the first speed hump in the city of Delft in
Almost 1.2 million traffic-related fatalities and 50 million traffic- the Netherlands. Since then, speed humps have achieved global
related injuries occur each year worldwide, more than half of which dominance as traffic-calming devices in comparison with other
involve pedestrians, especially in developing countries. Pedestrians approaches.
were involved in about 52% of traffic fatalities in New York be-
tween 2005 and 2009 (Ewing et al. 2013). Moreover, it is antici-
pated that these traffic fatalities and injuries will increase by 65% Statement of the Problem
by 2020 owing to the rapid increase in individual vehicle use in According to the U.K. Department for Transport’s Local Transport
developing countries (World Health Organization 2004). There- Note 1/07 on traffic calming, among all traffic-calming devices,
fore, governments and transportation agencies need to adopt pol- speed humps have the best performance in terms of reduction of
icies to reduce the threat posed to pedestrians by high-speed traffic speed, traffic flow, and injuries due to accidents, as well
vehicles in residential areas. Such policies include traffic calming. as having the greatest degree of public acceptability. However, they
According to Lockwood (1997), “Traffic calming is the combina- cause the longest delays to emergency services and also have neg-
tion of mainly physical measures that reduce the negative effects of ative impacts on vehicle emissions (carbon monoxide, nitrogen ox-
motor vehicle use, alter driver behavior, and improve conditions for ides, and particulate matter) and noise pollution. Moreover, a speed
non-motorized street users.” Previous investigations and accident hump transfers an upward force to vehicles and their occupants and
statistics show that traffic calming techniques have a considerable cargo as they traverse it (Weber 1998). Generally, these advantages
effect in decreasing the incidence of accidents. For example, traffic- and disadvantages of speed humps depend on their geometric de-
calming devices that cause a vehicle speed reduction of 14.4 km=h sign and profile shape. Speed hump geometry has a considerable
can reduce accidents by as much as 70% (Ahn and Rakha 2009). effect on the degree of discomfort experienced by passengers
These devices can be categorized into three groups according to the (Department for Transport 2007). Unfortunately, there have been
mechanism by which they reduce vehicle speed: volume control, few theoretical studies on the optimization of hump profiles to re-
horizontal speed control, and vertical speed control. duce passenger discomfort, vehicle damage, and adverse environ-
Vertical speed control devices are the most widely used, and mental effects (Baslamisli and Unlusoy 2009). Also, most (if not
include speed humps and road bumps (road bumps longer than all) current hump profiles were developed at least 20 years ago and
3 m are usually called speed humps) (Ansari Ardeh et al. 2008; are poorly compatible with modern vehicle design.
Alfredo et al. 2011; Abaza et al. 2012). Speed humps decrease traf-
fic flow about 25% (Department for Transport 2007).
Purpose of the Study
The primary objective of this study is to investigate the dynamic
1
Researcher, Young Researchers and Elites Club, Science and Research response of vehicles on different types of speed humps and to
Branch, Islamic Azad Univ., 1465836935 Tehran, Iran (corresponding present new recommendations for profiles to optimize the dynamic
author). E-mail: [email protected] performance with regard to forces acting on vehicles and the con-
2
Assistant Professor, Dept. of Civil Engineering, Imam Khomeini Inter-
sequent risk of damage. Many important factors are related to the
national Univ., 3414916818 Qazvin, Iran. E-mail: [email protected]
Note. This manuscript was submitted on August 6, 2013; approved on dynamic characteristics of vehicles on humps, including vehicle
March 12, 2014; published online on May 12, 2014. Discussion period speed, hump height, and hump length. For example, low humps
open until October 12, 2014; separate discussions must be submitted for are ineffective, and high humps can cause serious damage to ve-
individual papers. This paper is part of the Journal of Transportation En- hicles; therefore the choice of a suitable height is crucial. In this
gineering, © ASCE, ISSN 0733-947X/04014035(12)/$25.00. context, the dynamic response of the vehicle is the main factor

© ASCE 04014035-1 J. Transp. Eng.

J. Transp. Eng. 2014.140.

 in the design of speed humps (Kassem and Al-Nassar 1982). In the Salau et al. (2004), Alfredo et al. (2011), Johansson et al. (2011),
present study, 12 humps of different types with different dimen- and Abaza et al. (2012).
sions are considered for a range of different vehicle speeds to de- The purpose of this study, and one of the most important aspects
termine the optimum values of the primary factors in the geometric of speed hump efficiency, is the effect of the geometric design of
design of humps, namely, height and length. Then, two recom- humps on optimizing the dynamic response of vehicles. As already
mended profiles are presented and their performance is compared mentioned, only a few studies have dealt with optimization of hump
with that of conventional speed humps. A series of formulas are profiles. One of the first studies of the dynamic effect of speed
also presented to estimate the dynamic performance of vehicles humps on vehicles was conducted by Kassem and Al-Nassar
against the forces acting on them based on hump dimension, ve- (1982). According to their results, ride comfort is directly propor-
hicle speed, and driving behavior. Two different driving behaviors tional to the maximum vertical acceleration of the driver and should
are considered in the analysis: (1) the driver negotiates the hump at not exceed 0.6g. Higher humps produce greater dynamic effects on
constant speed, and (2) the driver needs to use the brake while tra- vehicles. Kassem and Al-Nassar (1982) also recommended a maxi-
versing the hump. mum height of 8 cm for humps based on safety considerations and
Generally, since the dynamic characteristics of vehicles signifi- vehicle protection. Since 1999, highway regulations have limited the
Downloaded from ascelibrary.org by Selcuk Universitesi on 02/06/15. Copyright ASCE. For personal use only; all rights reserved.

cantly affect the performance of traffic-calming devices, a unified maximum height of speed humps to 10 cm (Department for Transport
approach involving both highway engineering and vehicle engi- 2007), although most guidelines and studies recommend a height of
neering is essential to solve many of the current problems in traffic 7.5 cm (close to that recommended by Kassem and Al-Nassar).
calming. This study adopts such an approach to obtain new profiles Baslamisli and Unlusoy (2009) proposed an optimization
for speed humps by multibody simulation modeling of a vehicle’s algorithm to model the characteristics (acceleration responses) of
dynamic response. various vehicle types in three case studies using a half-car model
and presented an optimal profile for each case. However, the rec-
ommended profiles are rather long, and the height is more than
Literature Review 10 cm, which is not in accord with current guidelines. Also, longer
humps do not necessarily reduce vehicle speed (Ansari Ardeh et al.
Watts (1973), one of the pioneers in the study of speed hump fea- 2008). Baslamisli and Unlusoy (2009) showed that conventional
tures, tested a series of speed humps at different crossing speeds sinusoidal, circular, and parabolic profiles do not perform well ac-
and showed that a speed hump 10 cm high and 3.7 m wide is cording to hump efficiency requirements. Garcia et al. (2012) pre-
the best option for residential areas. The results of Watts’ research sented a new traffic-calming device called a speed kidney that
led to the widespread adoption of speed humps from the 1970s on- consists of a raised and curved area at the center of each lane. Five
ward (Antic et al. 2013). main advantages have been mentioned for the speed kidney in com-
Most previous studies were conducted to evaluate the effect of parison with other raised traffic-calming devices: it has no effect on
traffic-calming devices on accident reduction. Zein et al. (1997) buses and emergency vehicles; it reduces the speed of passenger
considered four traffic calming-projects in the greater Vancouver cars and motorcycles with less discomfort to passengers; it is bicycle-
area to study the benefits of traffic calming, and they found an aver- friendly; it does not cause damage to vehicles; and it reduces the negative
age 40% reduction in collision frequency and a 38% reduction in environmental impact of noise, fuel consumption, and emissions.
annual claims costs. Following acceptance by the public that traffic-
calming devices (especially speed humps) are effective in reducing
accidents, recent work has been aimed at investigating the effi- Simulation Modeling Process
ciency of these devices from a number of aspects, including safety,
A total of 112 simulation tests were run in this study. The simu-
technical, and economic. Arien et al. (2013) used a driving simu-
lation process was conducted using CarSim simulation software,
lator to study the effect of traffic-calming devices on traffic safety,
which is one of the most widely used dynamic simulation packages
and they found that gate constructions and curves can reduce
in the automotive industry. This multibody simulation software was
vehicle speed and thereby improve traffic safety.
developed by Mechanical Simulation Corporation (MSC) to predict
Daham et al. (2005) and Ahn and Rakha (2009) examined
the performances of various two-axle vehicles and to analyze with
traffic-calming devices from the viewpoint of energy consumption high accuracy the dynamic behavior of vehicles in response to a
and environmental impact. Daham et al. (2005) concluded that given road geometry. Parameters, such as roadway geometric de-
speed humps can increase emissions of hydrocarbons, carbon sign and the inertial properties and characteristics of vehicles, are
monoxide, carbon dioxide, and nitrogen oxides by about 148, user-defined. Although vehicle dynamics simulations have long
117, 90, and 195%, respectively. Traffic calming also affects fuel been common in vehicle design, their use by highway engineers
consumption, possibly increasing it (Ahn and Rakha 2009). Behzad in roadway design is rare (Stine et al. 2010).
et al. (2007) and Kokowski and Makarewicz (2006) investigated the The simulation process and details are as follows.
effect of road bumps on noise pollution: they reduce the level of
traffic noise during the deceleration phase (usually before the ve-
hicle traverses the bump) and increase it during the acceleration Vehicle
phase after traversing the bump (Kokowski and Makarewicz CarSim supports two-axle passenger cars. An E-class sedan was
2006). According to Behzad et al. (2007), the noise level depends selected as a typical passenger car for simulation. Its structural
on the height of the bump, with the peak noise level for a bump of parameters are shown in Table 1. The suspension type was inde-
height 5.5 cm being measured as about 35% more than that for a pendent, with an unsprung mass (both sides) of 90 kg.
bump of height 4 cm.
Another important issue that arises in traffic calming is the dis-
Speed Hump Profiles
tance between devices. Salau et al. (2004) developed a mathemati-
cal model based on the concept of isolation factor versus frequency As already mentioned, the dynamic response of a vehicle to a speed
ratio to determine the effective distance between road bumps. For hump depends on the geometric parameters of the latter, such as its
more information about spacing of traffic calming devices, see height, length, and, especially, its profile shape, and so it is to these

© ASCE 04014035-2 J. Transp. Eng.

J. Transp. Eng. 2014.140.

 Table 1. E-Class Sedan Parameters (ITE 1997; Department for Transport 2007; Weber 1998). The lon-
Parameters Value gitudinal profile of a speed hump can be based on a number of
different shapes, such as a segment of a circle, a sine wave, or a
Wheelbase 3,048 mm
flat top with straight ramps (Department for Transport
Width 1,880 mm
Height 1,480 mm 2007). The conventional speed-hump profiles are flat-topped
Track 1,600 mm [Fig. 1(a)], parabolic [Fig. 1(b)], and sinusoidal [Fig. 1(c)]. The
Wheel center height 375 mm flat-topped configuration is also sometimes called a speed table.
Sprung mass 1,653 kg These conventional profiles are generally popular because of their
Height of mass center of sprung mass 590 mm simplicity and ease of construction (Baslamisli and Unlusoy 2009).
Length between mass center to front axle 1,402 mm As is clear from Fig. 1, some hump profiles are much longer than
Length between mass center to rear axle 1,646 mm others. To reduce dynamic effects at the entrance and exit points, a
hump should be at least as long as the wheelbase of any vehicles
traversing it (Jarvis 1992). Therefore, longer humps are appropriate
where trucks and articulated vehicles, with their longer wheelbases,
Downloaded from ascelibrary.org by Selcuk Universitesi on 02/06/15. Copyright ASCE. For personal use only; all rights reserved.

are likely to be using the road.

Fig. 2 shows two new profiles that have been recommended on
the basis of the geometry of overpasses (although the scale of the
overpass here is undersized). They can be considered as sinusoidal
profiles and their detailed geometry is shown in Fig. 3.

Driver Behavior
Two types of driving behavior are considered: (1) the driver drives
over the speed hump at constant speed, and (2) the driver needs to
use the brake while traversing the hump. In most previous speed
hump simulations, owing to software limitations, braking was
not considered. In CarSim, braking inputs are defined by the pres-
sure applied to the brake system. Thus, another driving situation
was defined by using 5 MPa braking. This pressure is more relevant
when traversing speed humps. For example, 5 MPa braking can
Fig. 1. Conventional speed humps: longitudinal profile: (a) flat topped
reduce a speed of 50 to 34.8 km=h in a 1.5 s braking reaction time.
(in this study: L1 ¼ 1.8, 2, 2.5, 2.7 m; L2 ¼ 3, 4.3, 5.2, 5.5, 7 m;
Speed humps usually reduce a vehicle’s speed by about
and h ¼ 10, 7.5 cm); (b) parabolic (in this study: L ¼ 3.65, 4.26 m
10-15 km=h (Abdi Kordani and Mashhadi Zadeh 2008). The brake
and h ¼ 10, 7.6 cm); (c) sinusoidal (in this study: L ¼ 3.7 m and
is applied at the start of the speed hump and released when the
h ¼ 10 cm)
second axle passes the hump, so braking reaction times were se-
lected based on the vehicle speed and the hump length; they were
usually between 1 and 1.5 s. Note that the E-class sedan uses an
parameters that most attention should be paid during speed-hump antilock braking system (ABS).
design. In the simulations in this study, 12 different speed- The simulation considered five different vehicle speeds (20, 30,
hump configurations were considered. Fig. 1 shows their profiles, 40, 50, and 60 km=h) for the first type of driving behavior (no brak-
all of which were selected from reliable guidelines and studies ing) and three speeds (40, 50, and 60 km=h) for the second type

Fig. 2. New recommended profiles for speed humps: (a) recommended profile 4 × 0.08 (length ¼ 4 m and height ¼ 8 cm); (b) recommended profile
8 × 0.08 (length ¼ 8 m and height ¼ 8 cm)

© ASCE 04014035-3 J. Transp. Eng.

J. Transp. Eng. 2014.140.

Downloaded from ascelibrary.org by Selcuk Universitesi on 02/06/15. Copyright ASCE. For personal use only; all rights reserved.

Profile X1 X2 X3 X4 Y G1 G2 LSag LCrest

8*0.08 1.5 m 3m 5m 6.5 m 10 cm 4% 4% 3m 2m

4*0.08 0.5 m 1m 3m 3.5 m 12 cm 8% 8% 1m 2m

Fig. 3. Geometrical details of recommended profiles

(with braking). At speeds of 20 and 30 km=h, it is unusual to brake more useful for hump design. A series of dynamic models are
on speed humps, and so these speeds were not considered in the presented based on vehicle speed, type of speed hump, hump
braking tests. dimensions (height and length), and driver behavior (no braking
or braking) to estimate the acceleration and forces acting on ve-
hicles. An IBM SPSS statistics 19 was used to analyze the results.
Plots and Data Analysis
After inserting data on vehicle, hump geometry, and driving behav-
ior, the final step in the simulation process is plot definition. The Simulation Outputs
defined plots in this study are the vertical acceleration of the center
of gravity, the vertical force, the spring force, and the damper force. Simulation results are presented below to show the impact of vari-
Speed humps can cause large pitching motion and large suspension ous designs of speed humps on vehicles under no-braking and brak-
travel and can even result in wheel loss at high traversal speeds ing conditions. The terminology describing the speed humps is as
(Baslamisli and Unlusoy 2009). To find the effect of speed humps follows:
on vehicles, three force components were studied: spring forces, • flat-topped humps: ramp length × flat length × height
damper forces, and vertical forces on tires. Also, the vertical accel- • parabolic, sinusoidal, and recommended models: length ×
eration was taken as a comfort criterion for drivers and passengers height
while traversing the hump. Watts (1973) indicated that the main For example, parabolic profile 3.65 × 0.1 means a parabolic
cause of a driver’s discomfort is his or her peak vertical accelera- hump with a length of 3.65 m and a height of 0.1 m.
tion. In addition, vertical, spring, and damper forces can damage
the vehicle. Most conventional suspensions use passive springs
to absorb impact and dampers (or shock absorbers) to control Vertical Acceleration
spring motion. The shock absorbers dampen the vertical motion Vertical acceleration is known to be one of the main factors deter-
of a vehicle on its springs. The suspension system of a typical pas- mining driver discomfort. The maximum vertical acceleration of
senger car is illustrated in Fig. 4. One of the advantages of CarSim the center of gravity is shown in Fig. 5 for different types of humps
is its ability to provide an animation of the simulation results to give and two new recommended models. According to Fig. 5 and the
users a clear perception of the vehicle’s performance. results of previous studies (Jarvis 1992; Weber 1998; Weber and
Following the simulation, a multiplex regression analysis was Braaksma 2000; Ansari Ardeh et al. 2008), the maximum vertical
performed on the simulation outcomes to make them clearer and acceleration occurs at higher speeds. The key point is that the maxi-
mum vertical acceleration is determined primarily by the hump
height and the ramp length (or the entrance slope) in flat-topped
humps, with the effect of the length of the straight section being
negligible. The height also has the primary role for parabolic
and sinusoidal humps. The performance of flat-topped humps
was found to be better than that of parabolic humps or sinusoidal
humps, but it is also notable that there is a considerable difference
between the results for the recommended profile 8 × 0.08 and for
conventional humps. Profile 8 × 0.08 performs much better in
terms of the vertical acceleration acting on vehicles, and conse-
quently drivers and passengers experience less discomfort on this
Fig. 4. Spring and damper in an independent suspension system of
profile. Profile 4 × 0.08 also exhibits a better dynamic response
typical passenger cars
than parabolic and even sinusoidal humps.

© ASCE 04014035-4 J. Transp. Eng.

J. Transp. Eng. 2014.140.

 0.45

Maximum Vertical Acceleration of CG (g's)

Flat 1.8*3*0.1
Flat 2*3*0.1
0.35 Flat 2.5*3*0.1
Flat 2.7*3*0.1
Flat 1.8*5.5*0.075
0.25
Flat 1.8*7*0.075
Flat 1.8*4.3*0.1
Flat 1.8*5.2*0.1
0.15
Model 4*0.08
Model 8*0.08

0.05
20 30 40 50 60
(a) Speed (km/h)
Downloaded from ascelibrary.org by Selcuk Universitesi on 02/06/15. Copyright ASCE. For personal use only; all rights reserved.

0.65
Maximum Vertical Acceleration of CG (g's)

Parabolic 3.65*0.1
0.45 Parabolic 3.65*0.076

Parabolic 4.26*0.076

Sinusodial 3.7*0.1
0.25 Model 4*0.08

Model 8*0.08

0.05
20 30 40 50 60
(b) Speed (km/h)

Fig. 5. Performance of speed humps with regard to vertical acceleration acting on vehicle: (a) flat-topped and recommendation speed humps;
(b) parabolic, sinusoidal, and recommendation speed humps

0.45
Maximum Vertical Acceleration of CG (g's)

Flat 1.8*3*0.1
0.4
Flat 2*3*0.1
0.35 Flat 2.5*3*0.1
Flat 2.7*3*0.1
0.3
Flat 1.8*5.5*0.075

0.25 Flat 1.8*7*0.075

Flat 1.8*4.3*0.1
0.2 Flat 1.8*5.2*0.1
Model 4*0.08
0.15
Model 8*0.08
0.1
40 50 60
(a) Speed (km/h)

0.64
Maximum Vertical Acceleration of CG (g's)

0.55
Parabolic 3.65*0.1
0.46 Parabolic 3.65*0.076

Parabolic 4.26*0.076
0.37
Sinusoidal 3.7*0.1

0.28 Model 4*0.08

Model 8*0.08
0.19

0.1
40 50 60
(b) Speed (km/h)

Fig. 6. Performance of speed humps with regard to vertical acceleration acting on vehicle under braking conditions: (a) flat-topped and
recommendation speed humps; (b) parabolic, sinusoidal, and recommendation speed humps

© ASCE 04014035-5 J. Transp. Eng.

J. Transp. Eng. 2014.140.

 34

Flat 1.8*3*0.1
Flat 2*3*0.1
30
Flat 2.5*3*0.1

Tire Load (kN)

Flat 2.7*3*0.1
Flat 1.8*5.5*0.075
26
Flat 1.8*7*0.075
Flat 1.8*4.3*0.1
Flat 1.8*5.2*0.1
22
Model 4*0.08
Model 8*0.08

18
20 30 40 50 60
(a) Speed (km/h)
Downloaded from ascelibrary.org by Selcuk Universitesi on 02/06/15. Copyright ASCE. For personal use only; all rights reserved.

42

37 Parabolic 3.65*0.1

Parabolic 3.65*0.076
Tires Load (kN)

32
Parabolic 4.26*0.076

Sinusoidal 3.7*0.1
27
Model 4*0.08

Model 8*0.08
22

17
20 30 40 50 60
(b) Speed (km/h)

Fig. 7. Tire load of vehicle for different types of speed humps: (a) flat-topped and recommendation speed humps; (b) parabolic, sinusoidal, and
recommendation speed humps

33

Flat 1.8*3*0.1
Flat 2*3*0.1
Flat 2.5*3*0.1
30
Flat 2.7*3*0.1
Tire Load (kN)

Flat 1.8*5.5*0.075
Flat 1.8*7*0.075
Flat 1.8*4.3*0.1
27
Flat 1.8*5.2*0.1
Model 4*0.08
Model 8*0.08

24
40 50 60
(a) Speed (km/h)

39

Parabolic 3.65*0.1

34 Parabolic 3.65*0.076
Tire Load (kN)

Parabolic 4.26*0.076

Sinusoidal 3.7*0.1

29 Model 4*0.08

Model 8*0.08

24
40 50 60
(b) Speed (km/h)

Fig. 8. Tire load of vehicle for different types of speed humps under braking conditions: (a) flat-topped and recommendation speed humps;
(b) parabolic, sinusoidal, and recommendation speed humps

© ASCE 04014035-6 J. Transp. Eng.

J. Transp. Eng. 2014.140.

 30

Flat 1.8*3*0.1
Flat 2*3*0.1

Suspension Spring Force (kN)

Flat 2.5*3*0.1
26
Flat 2.7*3*0.1
Flat 1.8*5.5*0.075
Flat 1.8*7*0.075
Flat 1.8*4.3*0.1
22
Flat 1.8*5.2*0.1
Model 4*0.08
Model 8*0.08

18
20 30 40 50 60
(a) Speed (km/h)
Downloaded from ascelibrary.org by Selcuk Universitesi on 02/06/15. Copyright ASCE. For personal use only; all rights reserved.

29

27
Suspension Spring Force (kN)

Parabolic 3.65*0.1

Parabolic 3.65*0.076
25
Parabolic 4.26*0.076

23 Sinusoidal 3.7*0.1

Model 4*0.08
21
Model 8*0.08

19
20 30 40 50 60
(b) Speed (km/h)

Fig. 9. Suspension spring force of vehicle for different types of speed humps: (a) flat-topped and recommendation speed humps; (b) parabolic,
sinusoidal, and recommendation speed humps

30

Flat 1.8*3*0.1
Flat 2*3*0.1
Suspension Spring Force (kN)

Flat 2.5*3*0.1
28
Flat 2.7*3*0.1
Flat 1.8*5.5*0.075
Flat 1.8*7*0.075
Flat 1.8*4.3*0.1
26
Flat 1.8*5.2*0.1
Model 4*0.08
Model 8*0.08

24
40 50 60
(a) Speed (km/h)

29

Parabolic 3.65*0.1
Suspension Spring Force (kN)

Parabolic 3.65*0.076

Parabolic 4.26*0.076
26
Sinusoidal 3.7*0.1

Model 4*0.08

Model 8*0.08

23
40 50 60
(b) Speed (km/h)

Fig. 10. Suspension spring force of vehicle for different types of speed humps under braking conditions: (a) flat-topped and recommendation speed
humps; (b) parabolic, sinusoidal, and recommendation speed humps

© ASCE 04014035-7 J. Transp. Eng.

J. Transp. Eng. 2014.140.

 Based on Watts’ (1973) experiments, the maximum vertical Braking tests also showed interesting results, as can be seen in
accelerations acting on drivers should be between 0.7 and 0.9g. Fig. 8. Again, braking has the greatest effect on the sinusoidal
Jarvis (1992) also took a maximum vertical acceleration of 0.7g hump. An interesting point is the different performance of the
to represent a tolerable level of discomfort. However, Weber sinusoidal profile and profile 4 × 0.08 in comparison with the
(1998) suggested a vertical acceleration level of 0.57g for passenger others at higher speeds. Also, vehicles experience greater vertical
cars in North America. Fig. 5 also shows a maximum vertical forces on profile 8 × 0.08 in comparison with some flat-topped
acceleration of 0.6046g ≈ 0.6g. humps at higher speeds, with the differences being particularly
Fig. 6 illustrates the maximum vertical acceleration under large in the no-braking case.
braking conditions. Based on this figure, braking can increase
the maximum vertical acceleration, especially on sinusoidal humps
Spring Force
at higher speeds. The vertical acceleration of a sedan on a sinus-
oidal hump at 60 km=h under braking conditions was estimated to Vehicle suspension systems are especially vulnerable to damage
be 38% more than in the absence of braking. This causes sinusoidal from speed humps, and this can be costly. Fig. 9 shows the spring
humps in general to produce more vertical acceleration than most forces on vehicles on speed humps under the no-braking condition.
Downloaded from ascelibrary.org by Selcuk Universitesi on 02/06/15. Copyright ASCE. For personal use only; all rights reserved.

other speed humps. The most important finding from Fig. 9 is the unexpected response
of the sinusoidal profile and of profile 4 × 0.08, while parabolic
profile 4.26 × 0.076 shows a better response than these humps.
Vertical Force
Also, the spring forces are seen to be similar on parabolic humps
The vertical forces on vehicle tires are shown in Fig. 7. A linear re- at speeds of 40 and 50 km=h.
lationship is seen for most of the tests. As with the vertical acceler- The results of braking tests are shown in Fig. 10, which shows
ation results, hump length does not have a significant effect on the that profile 4 × 0.08 has one of the poorest performances with
vertical forces acting on a vehicle. Note that the maximum vertical regard to spring forces. In contrast to the no-braking tests, the sinus-
force on a tire for a straight path is about 17.95 kN, while it is be- oidal hump has the best performance of all the humps, even better
tween 19.8 and 34.9 kN on conventional humps. It is clear that con- than that of profile 8 × 0.08. Also, on a flat road (with no speed
ventional humps can increase the vertical forces by as much as 94%. humps), the spring force acting on the suspension system was
Also, profile 8 × 0.08 has the best dynamic response of all humps, calculated to be about 19 kN, while for the various humps the
and profile 4 × 0.08 is better in this respect than the parabolic and spring force was estimated to be between 19.7 and 29.5 kN in the
sinusoidal humps. Profile 4 × 0.08 also produces lower vertical no-braking case and between 23.3 and 28.7 kN with braking, de-
forces on vehicles than the four types of flat-topped humps. pending on the hump profile and the vehicle speed.

14

Flat 1.8*3*0.1
Suspension Damper Forces (kN)

Flat 2*3*0.1
Flat 2.5*3*0.1
10
Flat 2.7*3*0.1
Flat 1.8*5.5*0.075
Flat 1.8*7*0.075
Flat 1.8*4.3*0.1
6 Flat 1.8*5.2*0.1
Model 4*0.08
Model 8*0.08

2
20 30 40 50 60
(a) Speed (km/h)

12

10
Suspension Damper Forces (kN)

Parabolic 3.65*0.1

Parabolic 3.65*0.076
8
Parabolic 4.26*0.076

6 Sinusoidal 3.7*0.1

Model 4*0.08

4 Model 8*0.08

2
20 30 40 50 60
(b) Speed (km/h)

Fig. 11. Suspension damper force of vehicle for different types of speed humps: (a) flat-topped and recommendation speed humps; (b) parabolic,
sinusoidal, and recommendation speed humps

© ASCE 04014035-8 J. Transp. Eng.

J. Transp. Eng. 2014.140.

 10

Flat 1.8*3*0.1
Flat 2*3*0.1

Suspension Damper Forces (kN)

8
Flat 2.5*3*0.1
Flat 2.7*3*0.1
Flat 1.8*5.5*0.075
6
Flat 1.8*7*0.075
Flat 1.8*4.3*0.1
Flat 1.8*5.2*0.1
4
Model 4*0.08
Model 8*0.08

2
40 50 60
(a) Speed (km/h)
Downloaded from ascelibrary.org by Selcuk Universitesi on 02/06/15. Copyright ASCE. For personal use only; all rights reserved.

12
Suspension Damper Forces (kN)

10
Parabolic 3.65*0.1

Parabolic 3.65*0.076
8
Parabolic 4.26*0.076

6 Sinusoidal 3.7*0.1

Model 4*0.08
4
Model 8*0.08

2
40 50 60
(b) Speed (km/h)

Fig. 12. Suspension damper force of vehicle for different types of speed humps under braking conditions: (a) flat-topped and recommendation speed
humps; (b) parabolic, sinusoidal, and recommendation speed humps

Damper Force • The length of flat-topped humps does not have a significant
effect on the dynamic response of passenger cars, while it is
Finally, damper forces are another main aspect of the study and are
found to be one of most important parameters for parabolic
shown in Figs. 11 and 12. With regard to these forces, both of the
humps.
recommended profiles have one of the best performances (only two
of the flat-topped profiles are slightly better than profile 4 × 0.08).
There is a considerable difference in damping forces at speeds of 20
and 30 km=h for some flat-topped humps. As with the previous re- Data Analysis and Modeling
sults, the hump length has no effect on the damping forces, in either
Following the simulation process and the classification of the re-
the braking or no-braking cases. The results for the damping forces
sults, a multiplex regression analysis was conducted to allow the
for the flat-topped humps in the braking tests are all very close.
development of dynamic models of passenger cars on different
types of speed humps based on vehicle speed and hump dimen-
Summary sions. Tables 3–6 show the coefficients and significances of vari-
ables in the statistical analysis for flat-topped and parabolic humps.
Table 2 shows a summary of the simulation results to provide a com- The modeling was not conducted for sinusoidal and recommended
parison of the dynamic performance of speed humps. From the view- humps owing to insufficient data being available.
point of dynamic performance, the following results were obtained: The variables are as follows:
• The recommended profile 8 × 0.08 shows the best dynamic re- AZ = vertical acceleration (g)
sponse, and its performance is considerably different from that FZ = vertical forces on tires (kN)
of the other humps. The dynamic performance of the recom- S = spring forces (kN)
mended profile 4 × 0.08 is also estimated to be much better D = damper forces (kN)
than that of the parabolic, sinusoidal, and four types of flat- V = vehicle speed (km=h)
topped humps. h = hump height (m)
• For flat-topped humps, the ramp length (or entrance slope) has L1 = ramp length (m), for flat-topped profiles
the greatest effect on performance. It is clear that the flat-topped L2 = length of flat section (m), for flat-topped profiles
profile 1.8 × 3 × 0.1 has one of the worst performances and that L = hump length (m), for parabolic profiles
the flat-topped profile 2.7 × 3 × 0.1 has the best performance of
the conventional humps.
Flat-Topped Humps
• Because of their shapes, sinusoidal and parabolic humps are
generally more expensive than flat-topped humps, and so the The dynamic response of a vehicle on flat-topped humps in the no-
latter might be a better option than other conventional humps. braking case is described by the following equations:

© ASCE 04014035-9 J. Transp. Eng.

J. Transp. Eng. 2014.140.

 Table 2. Dynamic Response of Vehicles on Road Humps Based on CarSim Simulation
Vertical acceleration Vertical forces Spring forces Damper forces
Road hump No braking Braking No braking Braking No braking Braking No braking Braking Total
Profile 8 × 0.08 a a a a a b a a
55
Flat-topped 2.7 × 3 × 0.1 b b b a b c e b
45
Flat-topped 1.8 × 7 × 0.075 b c b b b d b b
45
Flat-topped 1.8 × 5.5 × 0.075 b c b b b d b c
44
Flat-topped 2.5 × 3 × 0.1 b c b a b d e b
43
Profile 4 × 0.08 e d c e d f b d
34
Flat-topped 2 × 3 × 0.1 c d c c e e f d
31
Flat-topped 1.8 × 4.3 × 0.1 d e e d f e d d
27
Flat-topped 1.8 × 5.2 × 0.1 d e e d f e d d
27
Parabolic 4.26 × 0.076 f f f e c c c e
27
Sinusoidal 3.7 × 0.1 f f e f e a c f
26
Downloaded from ascelibrary.org by Selcuk Universitesi on 02/06/15. Copyright ASCE. For personal use only; all rights reserved.

Parabolic 3.65 × 0.076 f f f f d d d e

23
Flat-topped 1.8 × 3 × 0.1 d e e f f f g d
21
Parabolic 3.65 × 0.1 g g g g g g e g
10
a
The best.
b
Excellent.
c
Very good.
d
Good.
e
Not bad.
f
Bad.
g
The worst.

Table 3. Results of Statistical Analysis for Flat-Topped Humps Table 4. Results of Statistical Analysis for Flat-Topped Humps (Braking
(No-Braking Case) Using SPSS Software Conditions) Using SPSS Software
Variable B t Significance Variable B t Significance
Vertical acceleration Vertical acceleration
Constant −0.095 −1.849 0.073 Constant 0.064 1.374 0.185
V 0.006 37.431 0.000 V 0.005 21.462 0.000
h 2.488 7.020 0.000 h 2.614 8.385 0.000
L1 −0.068 −8.407 0.000 L1 −0.113 −15.893 0.000
L2 0.003 1.124 0.268 L2 −0.002 −0.591 0.561
R2 ¼ 0.978 R2 ¼ 0.978
Vertical force Vertical force
Constant 16.399 10.478 0.000 Constant 25.110 10.581 0.000
V 0.191 40.218 0.000 V 0.087 7.188 0.000
h 86.861 8.077 0.000 h 62.761 3.952 0.000
L1 −2.914 −11.907 0.000 L1 −3.296 −9.118 0.000
L2 0.15 0.159 0.875 L2 −0.110 −0.808 0.429
R2 ¼ 0.982 R2 ¼ 0.898
Spring force Spring force
Constant 16.525 11.365 0.000 Constant 25.530 23.834 0.000
V 0.148 33.574 0.000 V 0.064 11.625 0.000
h 71.282 7.134 0.000 h 11.576 1.615 0.123
L1 −2.190 −9.630 0.000 L1 −1.189 −7.288 0.000
L2 −0.060 −0.705 0.486 L2 −0.118 −1.919 0.070
R2 ¼ 0.974 R2 ¼ 0.912
Damper force Damper force
Constant 10.054 1.892 0.067 Constant 1.230 1.201 0.245
V 0.165 10.198 0.000 V 0.113 21.558 0.000
h −2.642 −0.072 0.943 h 39.634 5.780 0.000
L1 −1.431 −1.722 0.094 L1 −1.982 −12.701 0.116
L2 −1.357 −4.325 0.000 L2 −0.152 −2.575 0.019
R2 ¼ 0.819 R2 ¼ 0.974

MaxAZ ¼ −0.095 þ 0.006V þ 2.488 h − 0.068L1 þ 0.003L2 MaxD ¼ 10.054 þ 0.165V − 2.642 h − 1.431L1 − 1.357L2
ð1Þ ð4Þ
MaxFZ ¼ 16.399 þ 0.191V þ 86.861 h − 2.914L1 þ 0.15L2 For example, from Eq. (2), the vertical force for a design speed
ð2Þ of 50 km=h on the flat-topped hump profile 1.8 × 3 × 0.1 is esti-
mated as 29.84 kN.
MaxS ¼ 16.525 þ 0.148V þ 71.282 h − 2.190L1 − 0.060L2
The dynamic response under braking conditions is given by the
ð3Þ following equations:

© ASCE 04014035-10 J. Transp. Eng.

J. Transp. Eng. 2014.140.

 Table 5. Results of Statistical Analysis for Parabolic Humps (No-Braking Table 6. Results of Statistical Analysis for Parabolic Humps (Braking
Case) Using SPSS Software Conditions) Using SPSS Software
Variable B t Significance Variable B t Significance
Vertical acceleration Vertical acceleration
Constant −0.005 −0.046 0.965 Constant −0.009 −0.072 0.945
V 0.007 19.992 0.000 V 0.005 7.261 0.001
h 3.953 8.236 0.000 h 5.005 9.182 0.000
L −0.065 −3.229 0.008 L −0.058 −2.554 0.051
R2 ¼ 0.980 R2 ¼ 0.976
Vertical force Vertical force
Constant 13.910 3.374 0.006 Constant 18.200 3.224 0.023
V 0.284 20.487 0.000 V 0.159 4.984 0.004
H 136.483 7.213 0.000 h 186.814 7.448 0.001
L −1.704 −2.163 0.053 L −2.488 −2.382 0.063
R2 ¼ 0.979 R2 ¼ 0.963
Downloaded from ascelibrary.org by Selcuk Universitesi on 02/06/15. Copyright ASCE. For personal use only; all rights reserved.

Spring force Spring force

Constant 18.714 8.297 0.000 Constant 22.422 14.439 0.000
V 0.085 11.180 0.000 V 0.089 10.123 0.000
h 78.513 7.583 0.000 H 33.777 4.895 0.004
L −0.957 −2.220 0.048 L −0.654 −2.276 0.072
R2 ¼ 0.954 R2 ¼ 0.969
Damper force Damper force
Constant −0.141 −0.049 0.961 Constant −1.627 −0.888 0.415
V 0.109 11.425 0.000 V 0.108 10.411 0.000
h 79.102 −6.052 0.000 h 85.234 10.471 0.000
L −0.867 −1.592 0.140 L −0.695 −2.049 0.096
R2 ¼ 0.956 R2 ¼ 0.956

MaxAZ ¼ 0.064 þ 0.005V þ 2.614 h − 0.113L1 − 0.002L2 MaxD ¼ −1.627 þ 0.108V þ 85.234 h − 0.695L ð16Þ
ð5Þ According to the statistical analysis, there is a strong relation-
ship between a vehicle’s dynamic response and hump height.
MaxFZ ¼ 25.110 þ 0.087V þ 62.761 h − 3.296L1 − 0.110L2 Drivers need to use braking forces on downgrades or on hori-
ð6Þ zontal curves. Bonneson (1999) pointed out that 87% of drivers
reduce their speeds on curves. Therefore, the results and models
for the braking case are relevant for the design of speed humps
MaxS ¼ 25.530 þ 0.064V þ 11.576 h − 1.189L1 − 0.118L2 on downgrades and curves. However, the installation of humps
ð7Þ on horizontal curves is not recommended.

MaxD ¼ 1.230 þ 0.113V þ 39.634 h − 1.982L1 − 1.152L2

Conclusions
ð8Þ
Both of the recommended profiles show a good dynamic response.
From the viewpoint of vehicles’ dynamic performance, profile 8 ×
Parabolic Humps 0.08 (length 8 m and height 8 cm) is the best speed hump. The
The dynamic models for parabolic humps in the no-braking case performance of profile 4 × 0.08 is also estimated to be much better
are as follows: than that of parabolic, sinusoidal, and four types of flat-topped
humps. Since the length of a hump is determined by the wheelbase
MaxAZ ¼ −0.005 þ 0.007V þ 3.953 h − 0.065L ð9Þ of the types of vehicles for which it is principally intended, profile
8 × 0.08 is more appropriate for heavy vehicle routes (commercial
MaxFZ ¼ 13.910 þ 0.284V þ 136.483 h − 1.704L ð10Þ roads) and profile 4 × 0.08 for urban streets.
The maximum vertical force on a passenger sedan on a flat path
MaxS ¼ 18.714 þ 0.085V þ 78.513 h − 0.957L ð11Þ (no humps) is about 17.95 kN, while it is between 19.8 and 34.9 kN
on speed humps. Thus, speed humps can increase vertical forces by
as much as 94%.
MaxD ¼ −0.141 þ 0.109V þ 79.102 h − 0.867L ð12Þ The ramp length (or entrance slope) has the greatest effect on the
performance of flat-topped humps. According to Table 2, the flat-
Under the braking condition, the following equations are
topped profile 1.8 × 3 × 0.1 has one of the worst performances,
applicable:
while the flat-topped profile 2.7 × 3 × 0.1 is the best among the
MaxAZ ¼ −0.009 þ 0.005V þ 5.005 h − 0.058L ð13Þ conventional humps. On the other hand, the length of flat-topped
humps has a negligible effect on the dynamic response of passenger
cars, whereas it is found to be one of the most important parameters
MaxFZ ¼ 18.200 þ 0.159V þ 186.418 h − 2.488L ð14Þ
for parabolic humps.
Because of the greater cost of sinusoidal and parabolic humps,
MaxS ¼ 22.422 þ 0.089V þ 33.777 h − 0.654L ð15Þ flat-topped profile 2 × 3 × 0.1 might be a better option than

© ASCE 04014035-11 J. Transp. Eng.

J. Transp. Eng. 2014.140.

 conventional parabolic and sinusoidal humps. The dynamic perfor- Behzad, M., Hodaei, M., and Alimohammadi, I. (2007). “Experimental and
mances of all of them are found to be almost the same in the numerical investigation of the effect of a speed bump on car noise emis-
simulations. sion level.” Appl. Acoust., 68(11–12), 1346–1356.
Braking increases the maximum vertical acceleration of Bonneson, J. A. (1999). “Side friction and speed as controls for horizontal
curve design.” J. Transp. Eng., 10.1061/(ASCE)0733-947X(1999)
vehicles, especially on sinusoidal humps at higher speeds. This
125:6(473), 473–480.
can have a considerable effect on driver discomfort. Daham, B., Andrews, G. E., Li, H., and Partridge, M. (2005). “Quantifying
The results and dynamic models from this study can be useful in the effects of traffic calming on emissions using on-road measure-
the geometric design phase to allow highway engineers to develop ments.” 2005 SAE World Congress, Detroit, MI.
efficient speed humps with a consequent increase in driver comfort Department for Transport, Department for Regional Development, Scottish
and a reduction in damage to vehicles, as well as an increase in Executive, and Welsh Assembly Government. (2007). Local transport
safety. note (LTN) 1/07: Traffic calming, The Stationary Office (TSO),
Norwich, U.K.
Ewing, R., Chen, L., and Chen, C. (2013). “Quasi-experimental study of
traffic calming measures in New York City.” No. 2364, Transportation
Further Studies Research Board of the National Academies, Washington, DC.
Downloaded from ascelibrary.org by Selcuk Universitesi on 02/06/15. Copyright ASCE. For personal use only; all rights reserved.

Garcia, A., Moreno, A. T., and Romero, M. A. (2012). “A new traffic-

This study has focused on the dynamic performance of passenger calming device: Speed kidney.” ITE J., 82(12), 28–33.
cars and has not taken account of environment and economic ITE. (1997). “Guidelines for the design and application for speed humps.”
parameters such as noise pollution, fuel consumption, and con- Traffic engineering council speed humps task force, Institute of Trans-
struction costs. An economic evaluation is essential to determine portation Engineers, Washington, DC.
the effect of the recommended profiles in reducing the costs of Jarvis, J. R. (1992). “An investigation of road humps for use on bus routes.”
damage to vehicles over a two-year to three-year period. Research Rep. ARR no. 222, Australian Road Research Board, Victoria,
Australia.
Johansson, C., Rosander, P., and Leden, L. (2011). “Distance between
speed humps and pedestrian crossings: Does it matter?” Accid. Anal.
References Prev., 43(5), 1846–1851.
Kassem, E., and Al-Nassar, Y. (1982). “Dynamic considerations of speed
Abaza, O. A., Hussein, Z. S., and Malto, I. S. (2012). “Optimization of
control humps.” J. Transp. Res. Part B, 16(4), 291–302.
speed control hump spacing.” ITE J., 82(8), 44–47.
Kokowski, P., and Makarewicz, R. (2006). “Predicted effects of a speed
Abdi Kordani, A., and Mashhadi Zadeh, A. R. (2008). “Evaluation of speed bump on light vehicle noise.” Appl. Acoust., 67(6), 570–579.
humps efficiency on urban streets.” J. Traffic Eng., 9(3), 37–41 (in Lockwood, I. M. (1997). “ITE traffic calming definition.” ITE J., 67(7),
Persian). 22–24.
Ahn, K., and Rakha, H. (2009). “A field evaluation case study of the envi- Salau, T. A. O., Adeyefa, A. O., and Oke, S. A. (2004). “Vehicle speed
ronmental and energy impacts of traffic calming.” Transport. Res. control using road bumps.” J. Transp., 19(3), 130–136.
Transport. Environ., 14(6), 411–424. Schlabbach, K. (1997). “Traffic calming in Europe.” ITE J., 67(7), 38–40.
Alfred, G., Antonio Jose, T., Mario Alfonso, R., and Ana Tsui, M. (2011). Stine, J. S., Hamblin, B. C., Brennan, S. N., and Donnell, E. D. (2010).
“Traffic micro-simulation study to evaluate the effect of type and “Analyzing the influence of median cross-section design on highway
spacing of traffic calming devices on capacity.” J. Procedia Soc. Behav. safety using vehicle dynamics simulations.” Accid. Anal. Prev.,
Sci., 16, 270–281. 42(6), 1769–1777.
Ansari Ardeh, H., Shariatpanahi, M., and Nikkhah Bahrami, M. (2008). Watts, G. R. (1973). “Road humps for the control of vehicle speeds.”
“Multiobjective shape optimization of speed humps.” J. Struct. Multi. Transport and Road Research Laboratory Rep. no. 597, Dept. of Envi-
Optim., 37(2), 203–214. ronment and Dept. of Transport, Crowthorne, U.K.
Antic, B., Pesic, D., Vujanic, M., and Lipovac, K. (2013). “The influence Weber, P. A. (1998). “Towards a Canadian standard for the geometric
of speed bumps heights to the decrease of the vehicle speed–Belgrade design of speed humps.” M.E. thesis, Dept. of Civil and Environmental
experience.” J. Saf. Sci., 57, 303–312. Engineering, Carleton Univ. Ottawa, ON, Canada.
Arien, C., Jongen, E. M. M., Brijs, K., Brijs, T., Daniels, S., and Wets, G. Weber, P. A., and Braaksma, J. P. (2000). “Towards a North American geo-
(2013). “A simulator study on the impact of traffic calming measures in metric design standard for speed humps.” ITE J., 70(1), 30–34.
urban areas on driving behavior and workload.” Accid. Anal. Prev., 61, World Health Organization. (2004). World report on road traffic injury
43–53. prevention, Geneva, 4.
Baslamisli, S. C., and Ünlüsoy, Y. S. (2009). “Optimization of speed con- Zein, S. R., Geddes, E., Hemsing, S., and Johnson, M. (1997). “Safety
trol hump profiles.” J. Transp. Eng., 10.1061/(ASCE)TE.1943-5436 benefits of traffic calming.” Transportation Research Record 1578,
.0000002, 260–269. Transportation Research Board, Washington, DC.

© ASCE 04014035-12 J. Transp. Eng.

J. Transp. Eng. 2014.140.