Abstract—Active vibration isolation systems are not commonly

used due to their associated cost and power requirements. In
principle, semi-semi active isolation systems can deliver adaptability,
good performance and less power consumption compared with fully
active control. This paper presents a semi-active concept of a 7-
degree of freedom (DOF) full vehicle model for heavy vehicle. To
develop suspension control system that can improve road holding and
improve ride comfort, dynamic modeling of passive and semi-active
suspension for heavy vehicle model was constructed. The simulation
model was developed using MATLAB Simulink software. Passive
heavy vehicle model was validated using vehicle dynamics
simulation software known as TruckSim. The validation was done by
comparing the simulation results. The primary focus of this paper is
on ride quality control and road holding. The paper investigates
analytically, the use of passive and semi-active suspension for ride
quality and road holding on heavy vehicle. The control scheme of the
semi-active suspension system is based on the groundhook control. A
ride test was conducted at constant speed, and the simulation results
of passive and semi-active suspension consist of roll, pitch, vehicle
body heave and tire forces are compared and analyzed. The results
show that the semi-active system controlled by groundhook strategy
provides better isolation than a conventional passive damped system.

Keywords—Heavy vehicle, semi active, groundhook control,
ride comfort, road holding





Syabillah Sulaiman, is with the Department of Automotive Engineering,
Faculty of Mechanical Engineering, Universiti Teknologi Malaysia (UTM),
81300, Skudai, J ohor, Malaysia (corresponding author, phone:
+60162405589; e-mail: syabillahsulaiman@gmail.com).
Pakharuddin Mohd Samin, is with the Department of Automotive
Engineering, Faculty of Mechanical Engineering, Universiti Teknologi
Malaysia (UTM), 81300, Skudai, J ohor, Malaysia (e-mail:
pakhar@fkm.utm.my).
Hishamuddin J amaluddin, is with the Department of SystemDynamics and
Control, Faculty of Mechanical Engineering, Universiti Teknologi Malaysia
(UTM), 81300, Skudai, J ohor, Malaysia (e-mail: hishamj@fkm.utm.my).
Roslan Abd Rahman, is with the Department of SystemDynamics and
Control, Faculty of Mechanical Engineering, Universiti Teknologi Malaysia
(UTM), 81300, Skudai, J ohor, Malaysia (e-mail: roslan@fkm.utm.my).
Mohammad Safwan Burhaumudin, is with the Department of Automotive
Engineering, Faculty of Mechanical Engineering, Universiti Teknologi
Malaysia (UTM), 81300, Skudai, J ohor, Malaysia (e-mail:
safwan_burhaumudin@yahoo.com).


I. INTRODUCTION
HEAVY military vehicle that transports troops needs high
vehicle stability, ride comfort and road friendliness [1],
[2]. This heavy vehicleis regularly driven on different terrains
[3], and thus the stability of the vehicle needs to be studied to
improve the vehicle ride performance [4], [5]. This simulation
model was validated with vehicle simulation software to
represent the vehicle’s ride behavior. This approach has
similarly been adopted by other researchers [6]-[10].
There are three types of model commonly used to represent
vehicle suspension behavior. These are quarter-vehicle model
[11], half vehicle model [12], [13] and full vehicle model
[14]-[16]. In this paper, a full car model was used to
investigate the ride performance behavior of passive and semi-
active suspension of heavy vehicle.
The design of a good suspension system is concerned with
isolation of the disturbances from the vehicle’s body. A
conventional passive suspension system has either low or high
damping characteristic. To insulate against load disturbance,
the damper needs to cover the full range of damping
characteristic. Therefore, suspension design is a compromise
between these two goals. There are two main categories of
suspension systems, namely passive and active suspension
system [17]. Passive suspension, which means there is no
external energy source in the system compared with active
suspensions, it is incorporated with external energy source to
achieve the required damping characteristic. Semi-active
suspension typically consists of controllable dampers and
passive spring. The performance of semi-active suspension
typically lies between active and passive suspension system
In suspension design, a full vehicle model is found to be
advantageous in studying a real vehicle performance. The full
vehicle model is a linear 7-degree of freedom consists of a
sprung mass that is connected to unsprung masses. Unsprung
mass consists of four suspensions and four tires located at
each corner of the body. The sprung mass has 3-degree of
freedom representing body bounce, roll and pitch movements,
while the unsprung masses has 4-degree of freedom in vertical
motions [18], [19].
The heavy vehicle simulation model was constructed using
MATLAB Simulink and was validated by using heavy vehicle
simulation software known as TruckSim.
Groundhook Control of Semi-Active suspension
For Heavy Vehicle
Syabillah Sulaiman, Pakharuddin Mohd Samin, Hishamuddin J amaluddin, Roslan Abd Rahman and
Mohammad Safwan Burhaumudin
A
International Journal of Research in Engineering and Technology (IJRET) Vol. 1, No. 3, 2012 ISSN 2277 – 4378
146
This paper presents the ride performance of a passive and
semi-active suspension systems of a heavy vehicle based on
the developed 7-degree of freedom full vehicle model.
II. MATHEMATICAL MODELING
A. Passive Suspension Model
The heavy vehicle ride model in this study is based on a
four wheels vehicle. The ride model has 7-degree of freedom
which involves bounce, pitch, roll of the vehicle body, and
four wheels vertical motions. Fig. 1 shows the vehicle ride
model.

Fig. 1 Seven degree of freedom of vehicle ride model

There are some assumptions made in this study. The vehicle
aerodynamic effect is neglected and the road is assumed to be
level except for road disturbance. The vehicle is also assumed
to be rigid where the load transfer from one point to another is
one hundred percent effective. Parameters of the vehicle are
also assumed to be constant throughout the simulation process
such as tire stiffness, spring stiffness, and damper coefficient.
Based on the 7-degree of freedom model in Fig. 1, the
dynamic of the sprung mass is defined by:

DRR SRR
DRL SRL DFR SFR DFL SFL B B
F F
F F F F F F Z m
− −
− − − − − − =
 
(1)

Where, m
B
is the mass of the vehicle,
B
Z
 
is the body
acceleration and F is the forces acting on vehicle model(
S
for
spring,
D
for damper,
FL
for front left,
FR
for front right,
RL
for
rear left,
RR
for rear right). The spring forces,
Sij
F
(i for front
or rear and j for left or right) that act on the suspensions are
given by:

) (
Uij Bij Sij Sij
Z Z K F − =
(2)

Where,
Bij
Z
is the sprung vertical displacement,
Uij
Z
is the
unsprung mass vertical displacement and K
Sij
is the
suspension spring stiffness. Then the damper forces,
Dij
F
of
the
suspensions are given by:


) (
Uij Bij Dij Dij
Z Z C F
 
− =
(3)

Where,
Bij
Z

is the sprung vertical velocity,
Uij
Z

is the
unsprung mass vertical velocity and
Dij
C
is the suspension
damper coefficient. Acceleration at unsprung mass is given
by:

Tij Dij Sij Uij Uij
F F F Z m − + =
 
(4)

Where,
Uij
m
is the unsprung mass,
Uij
Z
 
is the vertical
acceleration at unsprung mass and
Tij
F
is the dynamic tire
forces. Dynamic tire forces,
Tij
F
is defined as:

) (
Rij Uij Tij Tij
Z Z K F − =
(5)

Where,
Tij
K
is the tire stiffness and
Rij
Z
is the road profile
acting as the disturbance. The pitch effect of the vehicle is
given by:

b F F F F
a F F F F J
DRR SRR DRL SRL
DFR SFR DFL SFL y
) (
) (
+ + + +
+ + + − = θ
 
(6)

Where,
Y
J is the moment of inertia about y-axis and θ
 
is
the pitch acceleration, while a is the length of vehicle from the
center of gravity to the front end and b is the length of vehicle
from the center of gravity to the rear end of the vehicle. The
roll effect of the vehicle is given by:

d F F F F
c F F F F J
DRR SRR DFR SFR
DRL SRL DFL SFL x
) (
) (
+ + + +
+ + + − = ϕ 
(7)

Where,
X
J is the moment of inertia about x-axis and ϕ  is
the roll acceleration, while c is the length of the vehicle from
the center of gravity to the right end and d is the length of
vehicle from the center of gravity to the left end of the
vehicle.

B. Semi-Active Suspension Model
The mechanical model used for simulation uses a method
known as ideal ‘groundhook’ control. It adopts a
hypothetical damper, C
grd
connected between the unsprung
mass and the fixed fictitious frame on the ground.
Theoretically, groundhook control will improve the
responses of the unsprung system [18]. Fig. 2 shows the
ideal groundhook control on a quarter vehicle model.

Figs. 4 and 5 show the Simulink and TruckSim
simulations have similar trend but slightly different in
magnitude. Fig. 8 also shows the simulation result of
Simulink body displacement has the same trend as
TruckSim simulation but a slight different in magnitude.
This error maybe due to simplified model used in Simulink,
while TruckSim model is based on an actual tested vehicle
International Journal of Research in Engineering and Technology (IJRET) Vol. 1, No. 3, 2012 ISSN 2277 – 4378
147
simulation process.










Fig. 2 Ideal groundhook control

The groundhook control is given by:

0 if
0 if
< −
≥ −
st t
st t
v v
v v

then
then
0 =
=
f
v c f
t gnd
(8)
Where, v
t
is unsprung mass velocity, v
st
is relative velocity
between sprung mass and unsprung mass, f is semi active
suspension force and c
gnd
is groundhook damping coefficient.
The forces generated by the semi active controller are added
to the passive vehicle model. The equation of the semi active
suspension control is given by:

f F F F Z m
Tij Dij Sij Uij Uij
+ − + =
(9)

III. PASSIVE MODEL VALIDATION
A ride test conducted for Simulink and TruckSim is
presented. The model was tested using two different speeds
which are 36 km/h and 43 km/h for both simulations and the
results of the simulations are compared and discussed.
A. Simulation: 36 km/h
The performances of the simulation models in terms of
pitch, roll, and vertical displacement responses are compared
between Simulink and TruckSim simulation models. The
speed of the vehicle model is kept constant throughout the
simulation that is 36 km/h.
Figs. 3 to 5 show the simulation results of both Simulink
and TruckSim performances in the presence of external
disturbance. The vehicle hits the first bump at 5 second on the
left side and hit the second bump at 5.6 second on the right
side.

Fig. 3 Roll angle response at 36 km/h



Fig. 4 Pitch angle response at 36 km/h


Fig. 5 Body displacement response at 36 km/h

B. Simulation: 43 km/h
The results of the simulation at 43 km/h show similar
trend of roll angle, pitch angle, and body displacement
between Simulink and TruckSim. These are shown in Figs.
6 to 8.
The different vehicle speed used in the simulations is to
show that the trend of the output of the simulink model is
consistent and independent of speed. Figs. 6 to 8 show
similar trends as Figs. 3 to 5 for the speed at 36 km/h.


Fig. 6 Roll angle response at 43km/h

C
grd

International Journal of Research in Engineering and Technology (IJRET) Vol. 1, No. 3, 2012 ISSN 2277 – 4378
148

Fig. 7 Pitch angle response at 43km/h


Fig. 8 Body displacement response at 43km/h
IV. MODEL SIMULATION
The simulation of passive and semi-active suspensions of
heavy vehicle model was performed using Simulink. A ride
test was conducted for both simulations. The road profile as
disturbance was applied on the left tires followed by the right
tires for both simulations. The height and length of the bumps
are 0.1 m (incremental elevation) and 5 m (station)
respectively for both sides, as shown in Fig. 9.

Fig. 9 Road disturbance profile
Fig. 10 shows graphically the arrangement of the bumps
and the vehicle movement when it hits the bumps. The left
and the right bumps were arranged such that the change in the
direction of disturbance occurs instantaneously.


Fig. 10 Ride test road profile

A. Passive Suspension Model
Fig. 11 shows the passive suspension Simulink block
diagram of the ride model. The road profile disturbance acts
on the unsprung mass system. The signal from the unsprung
mass block diagram namely suspension tire forces are
transmitted to the sprung mass, pitch, and roll block diagram
to compute the output variable. Then the output from sprung
mass, pitch, and roll are fed back to the unsprung system.


Roll Rate
7
Pitch Rate
6
Roll Angle
5
Pitch Angle
4
Body Acceleration
3
Body Velocity
2
Body Displacement
1
unsprung
Zb
teta (pitch)
phi (roll)
Zrfl
Zrfr
Zrrl
Zrrr
Fzfl
Fzfr
Fzrl
Fzrr
Sprung Mass (Ms)
Suspension Forces
Zb
Zb dot
Zb dotdot
Roll (phi )
Suspension Forces
Roll Angle (phi)
Roll Rate (phi dot)
Road profile
Zrfl
Zrfr
Zrrl
Zrrr Pitch (teta)
Suspension Forces
Pitch Angle
Pitch Rate

Fig. 11 Passive Simulink block diagram

The output variables namely roll, pitch, body heave and tire
forces are recorded and compared with semi-active simulation
results. All parameters of the vehicle are assumed constant
throughout simulation.
B. Semi-Active Suspension Model
The same source of road disturbance applied in passive
simulation is used by the semi-active vehicle. The semi-active
suspension model used the same model as the passive system
except the extra forces are added to each suspension, these
forces are generated by the semi-active suspension control.
Fig. 12 shows the semi-active suspension Simulink block
diagram. The diagram shows the addition of four forces from
groundhook control.
Fig. 13 shows the groundhook control constructed in
Simulink. The velocities of sprung mass and unsprung masses
are link as an input of the control algorithm. The outputs of
the groundhook control are the forces applied to the
suspension.
Vehicle movement
International Journal of Research in Engineering and Technology (IJRET) Vol. 1, No. 3, 2012 ISSN 2277 – 4378
149
15
Vus RR
14
Vus RL
13
Vus FR
12
Vus FL
11
Fzrr
10
Fzrl
9
Fzfr
8
Fzfl
7
Roll Rate
6
Pitch Rate
5
Roll Angle
4
Pitch Angle
3
Body Acceleration
2
Body Velocity
1
Body Displacement
Zb
teta (pitch)
phi (roll)
Zrfl
Zrfr
Zrrl
Zrrr
Fsa
Fsa1
Fsa2
Fsa3
Fsfl
Fsfr
Fsrl
Fsrr
Fzfl
Fzfr
Fzrl
Fzrr
VusFL
VusFR
VusRL
VusRR
unsprung
Sus Forces
Zb
Zb dot
Zb dot dot
Sprung Mass (Ms)
Sus Forces
Roll Angle (phi)
Roll Rate (phi dot)
Roll (phi)
Sus Forces
Pitch Angle
Pitch Rate
Pitch (teta)
8
Fsa3
7
Fsa2
6
Fsa1
5
Fsa
4
Zrrr
3
Zrrl
2
Zrfr
1
Zrfl

Fig. 12 Semi-active Simulink block diagram
F
-Vu
Vs-Vu
1
Fsa
Merge
Merge1
u1
if(u1 >0)
else
If1
else {}
In1 Out1
If Action
Subsystem3
if {}
Vu Out1
If Action
Subsystem2
-1
0
2
V2
1
V1

Fig. 13 Groundhook control
V. RESULTS OF SIMULATION AND DISCUSSION
The performances of the simulation models are roll, pitch,
body heave and tire forces responses where comparison
between passive and semi-active suspension simulation
models are made. The road profile as shown in Fig. 9 was
used as the road disturbance. The speed of the vehicle model
is kept constant throughout the simulation that is 36 km/h.
Fig. 14 shows the tire force produced by passive and semi-
active suspensions. Both graphs show the same trend but in
semi-active graph it produces less force compared with
passive, with improvement of about 5.9 percent. The
improvement occurs by controlling the velocity of the tire as
appeared in the control algorithm equation. The fictitious
damper slows down the tire velocity by absorbing the force,
so the tire force is dependent on the groundhook damping
coefficient of the semi active suspension.

Fig. 14 Tire force

Fig. 15 and 16 show the body vertical displacement and
performance of body velocity respectively. It clearly shows
the improvement of the semi-active suspension system
compare with passive system. Semi-active suspension system
has lower body displacement and body velocity and 0.5
second faster settling time compared to passive suspension.
This situation occurs because semi-active suspensions slow
down the vehicle body as it slows down tire displacement.
Semi-active suspension improves about 6.9 percent in body
vertical displacement and 26 percent in body velocity.



Fig. 15 Body vertical displacement


Fig. 16 Body velocity
Figs. 17 and 18 show the comparison between passive and semi-
active suspension of roll angle and roll rate respectively. Both semi-
active suspension results show an improvement compared to passive
suspension. Similar as before, the semi-active suspension control the
velocity between the sprung mass and unsprung mass. Roll angle of
the semi-active suspension improve about 4 percent and roll rate
improved higher than roll angle that is about 13.6 percent. This
happen by controlling the velocity of the vehicle movement.
Figs. 19 and 20 show the pitch angle and pitch rate of the heavy
vehicle body respectively. Pitch response also show an improvement
compared to passive suspension. But the improvement is less than the
roll response. The settling time of the pitch angle and pitch rate are
faster than passive suspension. The improvement of semi-active
suspension is about 3.4 percent in pitch angle and about 6.1 percent
in pitch rate.
International Journal of Research in Engineering and Technology (IJRET) Vol. 1, No. 3, 2012 ISSN 2277 – 4378
150

Fig. 17 Roll angle


Fig. 18 Roll rate




Fig. 19 Pitch angle


Fig. 20 Pitch rate
Table 1 shows the summary of the improvement of the semi-active
suspension system compared to passive suspension. The entire
variable is based on groundhook control. It shows that groundhook
control reduce the heavy vehicle tire force and improve ride comfort.

TABLE 1
RMS VALUE OF PASSIVE AND SEMI ACTIVE SUSPENSION SYSTEM
Variable Improvement (%)
Tire force 5.9
Body heave 6.9
Body velocity 26.0
Roll angle 4.0
Roll rate 13.6
Pitch angle 3.4
Pitch rate 6.1
VI. CONCLUSION
A heavy vehicle ride model with passive suspension system
consists of 7-degree of freedom has been developed and
validated. This passive vehicle simulation model was
constructed using Simulink and compared with multi body
dynamics software known as TruckSim. These comparisons
were made to validate the Simulink model.
The passive model is then compared with a semi-active
suspension system. Both passive and semi-active suspensions
were simulated for ride test with the vehicle speed of 36 km/h.
The semi-active suspension was controlled using groundhook
control algorithm.
Graph of roll, pitch, body and tire responses of the heavy
vehicle were recorded from the simulations. The simulation
results module semi-active suspension responses for all
variables. The significant observation is, the trend of the
vehicle movement when it hits the bumps, the semi-active
suspension resulted with better isolation compared to passive
suspension.
The results indicate that semi-active suspension can reduce
road damage by reducing the vehicle tire forces and also
improved ride comfort.
ACKNOWLEDGMENT
The authors wish to thank the Ministry of Higher Education
(MOHE) and the Universiti Teknologi Malaysia (UTM) for
providing the research facilities and support especially all
staff’s of Faculty of Mechanical Engineering, Universiti
Teknologi Malaysia. This research is supported using a
research grant, Vote No.78608.
REFERENCES
[1] Y. Chen, J . He, W. Zhang, "System Optimization to Improve Ride
Comfort and Road Friendliness," International Conference on
Mechanical and Electrical (ICMET 2010), Singapore, 10-12 Sept. 2010.
[2] T.R.M. Rao, G.V. Rao, K.S. Rao, A. Purushottam, "Analysis of Passive
and Semi Active Controlled Suspension Systemfor Ride Comfort in an
Omnibus Passing Over a Speed Bump," Analysis of Passive & Semi
Active Controlled Suspension Systems, Vol. 5, pp. 7-17, 2010.
International Journal of Research in Engineering and Technology (IJRET) Vol. 1, No. 3, 2012 ISSN 2277 – 4378
151
[3] S.A. Pazooki, D. Cao, "Modeling and Validation of Off-Road Vehicle
Ride Dynamics," Mechanical Systems and Signal Processing, vol. 28,
pp. 679-695, 2012.
[4] K. Hudha, H. J amaluddin, P.M. Samin, "Disturbance Rejection Control
of a Light Armored Vehicle Using Stability Augmentation Based Active
Suspension System," International Journal Heavy Vehicle Systems, vol.
15, pp. 152-169, 2008.
[5] D.J .M Sampson, G. McKevitt, D. Cebon, "The Development of an
Active Roll Control Systemfor Heavy Vehicles," Proc. 16th IAVSD
Symposium on the Dynamics of Vehicles on Roads and Tracks, Pretoria,
South Africa, 30 Aug. – 3 Sept. 1999, pp. 704-715.
[6] S.A.A Bakar, R. Masuda, H. Hashimoto, T. Inaba, H. J amaluddin, R.A.
Rahman, P.M. Samin, “Ride Comfort Evaluations on Electric Vehicle
Conversion and Improvement Using Magnetorheological Semi Active
Suspension System,” SICE Annual Conference, Waseda University,
Tokyo, 13-18 Sept. 2011.
[7] P.M. Samin, H. J amaluddin, R.A. Rahman, S.A.A. Bakar, K. Hudha,
"Modeling and Validation of a 7-DOF Full Car for Ride Quality,"
CADME07, Putra Brasmana Hotel, Kuala Perlis, Malaysia, 25-26 Oct.
2007.
[8] J .D. Setiawan, M. Safarudin, A. Singh, "Modeling, Simulation and
Validation of 14 DOF Full Vehicle Model," Instrumentation,
Communications, Information Technology, and Biomedical Engineering
(ICICI-BME), 2009 International Conference, Bandung, 23-25 Nov.
2009.
[9] G. Schade, "Vehicle Ride Analysis of a Tractor-Trailer," 2000
International ADAMS User Conference, pp. 1 – 13, Orlando, Florida,
J une 2000.
[10] A. Forsén, "Heavy Vehicle Ride and Endurance– Modelling and Model
Validation," Department of Vehicle Engineering, Royal Institute of
Technology, Doctor of Philosophy Thesis, Stockholm, 1999.
[11] A. W. Burton, A. J . Truscott and P. E. Wellstead, "Analysis, Modeling
and Control of an Advanced Automotive Self-Leveling Suspension
System," IEEE Proc. on Control Theory Appl., vol. 142, No. 2, pp. 129-
139, 1995.
[12] W. Can, W. Weirui, "Chaotic Behaviors of Half Car Model Excited by
the Road Surface Profile," Information Science and Engineering
(ICISE), 2009 1st International Conference, pp.3752-3755, Nanjing,
China, 18-20 Dec. 2009.
[13] C.Y. Tang, G.Y. Zhao, H. Li, S.W. Zhou, "Research on Suspension
System Based on Genetic Algorithm and Neural Network Control,"
Intelligent Computation Technology and Automation, 2009. ICICTA '09.
Second International Conference, pp.468-471, vol.1, Zhangjiajie, China,
10-11 Oct. 2009.
[14] C.P. Cheng, C.H. Chao, T.H Li, "Design of Observer-Based Fuzzy
Sliding-Mode Control for An Active Suspension SystemWith Full-Car
Model," Systems Man and Cybernetics (SMC), 2010 IEEE International
Conference, pp.1939-1944, 10-13 Oct. 2010.
[15] P.S.A. Singh, I.Z.M. Darus, "Enhancement of SUV Roll Dynamics
Using Fuzzy Logic Control," Informatics and Computational
Intelligence (ICI), 2011 First International Conference, pp.106-111,
Bandung, Indonesia, 12-14 Dec. 2011.
[16] R. Darus, Y.M. Sam, "Modeling and Control of Active Suspension
Systemfor A Full Car Model," Signal Processing & Its Applications,
2009. CSPA 2009. 5th International Colloquium, pp.13-18, Kuala
Lumpur, Malaysia, 6-8 March 2009.
[17] T.D. Gillespie, “Active Suspension,” In: Fundamental of Vehicle
Dynamics, Warrendale, PA: Society of Automotive Engineers Inc.,
1993, pp. 269-271.
[18] E. Guglielmino, T. Sireteanu, C.W. Stammers, G. Ghita, M. Giuclea,
"Dampers and Vehicle Modelling," Semi-active Suspension Control:
Improved Vehicle Ride and Road Friendliness, London: Springer, 2008,
pp. 27-36.
[19] F.F. Ling, “Design and Analysis of Passive Automotive Suspension," In:
R. Rajamani, Vehicle Dynamics and Control, London: Springer, 2006,
pp. 287-323.









International Journal of Research in Engineering and Technology (IJRET) Vol. 1, No. 3, 2012 ISSN 2277 – 4378
152