21

A study of the eect of rear suspension auxiliary roll

damping on vehicle-handling dynamics
E Uzunsoy and O A Olatunbosun*
Department of Mechanical Engineering, The University of Birmingham, Birmingham, UK
The manuscript was received on 21 April 2004 and was accepted after revision for publication on 24 August 2004.
DOI: 10.1243/095440705X6523

Abstract: The handling dynamics of a University of Birmingham Racing Formula Student racing car,
incorporating a novel rear suspension load transfer system which provides auxiliary roll damping, is
modelled using a simulation tool, HANDSIM, developed for vehicle-handling studies, and the eect
of the system on the vehicles handling performance is discussed by considering the results obtained
from the simulation. The aim of implementing this suspension load transfer system is to improve the
handling characteristics of the vehicle in severe cornering manoeuvres.
Keywords: vehicle-handling dynamics, suspension system, simulation, fuzzy logic control, driver
model

1 INTRODUCTION

2 TYRE AND VEHICLE MODEL

This paper discusses the eects of a rear suspension

load transfer mechanism, featuring a pair of auxiliary
dampers which provide roll damping only, on the
general handling dynamics of a Formula Student
racing car. The handling responses of the vehicle with
the modied suspension system (including rear suspension auxiliary roll dampers) are compared with
those with the normal suspension system (i.e. without the auxiliary dampers). Both the front and the
rear suspension systems are modelled to determine
the complete eect of the suspension system on the
vehicle-handling dynamics. While the front suspension system is an application of the double-wishbone
design, the rear suspension is a combination of a
conventional double wishbone and two diagonally
mounted additional damper units, referred to as the
rear load transfer mechanism. The kinematics of
both systems are analysed using the velocity diagram
approach. Then, a four-degree-of-freedom vehicle
model, including the roll motion, is used to analyse
the eects of the suspension system in terms of the
handling dynamics.

A four-degree-of-freedom vehicle model is used in this

study. The model takes into account lateral, yaw, roll,
and longitudinal motions, enabling the inclusion of
traction and braking forces on handling manoeuvres.
The model considers the vehicle as travelling at
constant speed with small corrective braking and
tractive force inputs provided under the control of
a fuzzy logic controller [1]. In accordance with this
act, a fuzzy path planning method representing an
intelligent driver model [2] is used to create the
vehicle simulation path. The fuzzy controller consists
of three inputs related to the vehicle orientation from
the drivers point of view and six outputs representing the steering angle, braking pressure applied to
each wheel, and throttle position. A very general ow
diagram of the vehicle model and fuzzy pilot interaction is shown in Fig. 1. The outputs provided by
the vehicle model block (heading, yaw, roll, etc.) are
processed in the fuzzy inputs block and then sent
to the fuzzy logic controller block. The controller is
intended to react as a real driver; so it determines
the direction in which to turn the steering wheel,
and by how much, by considering the membership
functions of the relevant inputs. It also determines
any application of braking pressures or the throttle
as appropriate. The vehicle model block is replaceable by any vehicle model regardless of its complexity.
The path that the vehicle is supposed to follow is
pre-mapped on the Earth xed coordinates [13].

* Corresponding author: Department of Mechanical Engineering,

Vehicle Dynamics Research Group, Birmingham University,
Edgbaston, Birmingham, B15 2TT, UK. email: o.a.olatunbosun@
bham.ac.uk

D08004 IMechE 2005

Proc. IMechE. Vol. 219 Part D: J. Automobile Engineering

22

E Uzunsoy and O A Olatunbosun

Fig. 2 Forces acting on the vehicle: top view

Fig. 1 General ow diagram of the vehicle path
planning

The pitching eects of the small corrective braking

and tractive forces and the longitudinal load transfer
are ignored. The basic equations are
+rV)
F =m(U
(1)
x
+rU )
F =m(V
(2)
y
M =I r
(3)
z
zz
M =I p
(4)
x
xx
Equations (1), (2), (3) and (4) can be written in a
more detailed form from Fig. 1 as
rV)=F +F +(F +F ) cos dF
m(U
xrL
xrR
xfL
xfR
brL
F (F +F ) cos d,
brR
bfL
bfR
(F +F ) sin d
(5)
yfL
yfR
+rU )=F +F +(F +F ) cos d
m(V
yrL
yrR
yfL
yfR
+(F +F ) sin d,
xfL
xfR
(F +F ) sin d
(6)
bfL
bfR
+rU )(h h )+m g(h h )w
I p =m (V
xx
s
cg
rc
s cg
rc
C w K w
(7)
w
w
and

vehicle during roll motion is represented by Fig. 3.

The force distribution at each wheel is derived from
both Fig. 2 and Fig. 3 as shown by the equations
F

vof

=0.5m g
f
1
+rU )h +(K +K
)w+C w ]
+ [m (V
u
wf
wAroll
w
t uf
f
(9)

F =0.5m g
vif
f
1
+rU )h +(K +K
)w+C w ]
[m (V
u
wf
wAroll
w
t uf
f
(10)
F

vor

=0.5m g
r
1
+rU )h +K w+(C +C
+ [m (V
)w ]
u
wr
w
wAux
t ur
r
(11)

F =0.5m g
vir
r
1
+rU )h +K w+(C +C
)w ]
[m (V
u
wr
w
wAux
t ur
r
(12)

t
I r =a(F +F ) cos d+ f (F F ) sin d+,
zz
yfL
yfR
yfR
2 yfL
+a(F

xfL

+F

yrL

+F

xfR

) sin d

t
+ r (F F F +F ),
brR
xrL
xrR
2 brL
b(F

t
) f (F +F F F )
yrR
bfR
xfR
bfL
2 xfL

cos da(F

bfR

+F

bfL

) sin d

(8)

respectively.
Figure 2 shows the forces acting on a vehicle
during yaw motion while the force analysis of a
Proc. IMechE. Vol. 219 Part D: J. Automobile Engineering

Fig. 3 Force analysis of a vehicle during roll motion

D08004 IMechE 2005

Rear suspension auxiliary roll damping

23

Fig. 4 Tyre characteristics for a Goodyear 20.06.5-13 tyre [4] for Formula SAE racing cars

The tyre-cornering characteristics are provided

by interpolation of measured data. Tyre data were
obtained from Goodyear Tyres [4] (Fig. 4) for a
Goodyear 20.06.5-13 tyre which is specically
produced for Formula SAE racing cars.

3 VEHICLE SUSPENSION SYSTEM AND

ANALYSIS
The Formula Student car (Car 201) (Fig. 5) suspension system is of a double-wishbone design at both
the front and the rear. While the front suspension
system also includes an anti-roll bar, the rear design,
which is the main subject of this study, includes
two diagonally mounted dampers (Fig. 6) aimed at
improving the handling properties of the vehicle by
providing auxiliary damping in roll only.

Fig. 5 Car 201

D08004 IMechE 2005

Fig. 6 Rear suspension (diagonal damper) load transfer

mechanism: front view

The main objective for modifying the rear suspension system is instantaneously to produce high
roll stiness at the rear of the vehicle by the action
of the diagonal dampers at the beginning of a turn.
This instantaneously results in oversteer characteristics, which enables the car to turn into the corner
quickly. As the extra roll stiness due to the diagonal
dampers is removed, the handling characteristics
revert to understeer, giving the driver the condence
to accelerate out of the corner.
The velocity diagram approach is used to analyse
the kinematics of both front and rear suspension
systems. The velocity diagram to provide the ratio
between the wheel and springdamper motions was
derived from the drawings of the suspension systems
using the computer aided design software Solidworks
2003A. In the classical double-wishbone suspension
Proc. IMechE. Vol. 219 Part D: J. Automobile Engineering

24

E Uzunsoy and O A Olatunbosun

system, the springdamper system is assumed to be

in the same vertical axis as the tyres for simplicity.
All suspensions are functionally equivalent to a pair
of springs and dampers, as shown in Fig. 3. The
lateral separation of the springs causes them to
develop a roll resisting moment proportional to the
dierence in roll angle between the body and axle.
The stiness is given by
(13)
K = 1 K s2
w 2 s
The velocity diagram approach reduces any type
of springdamper placement to the classical system
by providing the motion ratios. In this graphical
analysis method [57] of the suspension behaviour,
roll characteristics are determined about an arbitrary
point, which does not move as the body rotates.
In the classical velocity diagram approach, the diagram starts from the roll centre. However, for each
increment in the roll angle, it is necessary to update
the diagram. This need could be eliminated by
selecting an arbitrary point analysis instead. Therefore, the analysis presented here starts with the
classical approach rst (Fig. 7) to provide a reference
velocity to be used in the creation of the velocity
diagram as a result of the arbitrary point analysis
(Fig. 8). In Fig. 9, O is at the intersection of the vehicle
centre-line with the ground plane. XC is the link
which transfers the tyre forces to the main spring
damper system (YZ) and the auxiliary damper (KL ),
2
via the rocker. The rocker is expressed by its edge
points: TLYKX. It rotates about the pivot point T.
The inboard points of attachment of the suspension
links are A and B. The wheel contact point is O. The
position of the roll centre is determined by extending
an imaginary axis from the suspension links AD and
BC until a point of intersection is obtained.
A line from the intersection point through the
point O cuts the centre line of the body at O, which
is the roll centre. The formation of the velocity dia-

Fig. 8 Complete velocity diagram of the rear suspension system

Fig. 9 Rear suspension links and springdamper placement

Fig. 7 Velocity diagram (classical approach)

gram (Fig. 8) starts from O with vectors ob and oa

in Fig. 7 perpendicular to OB and OA respectively
in Figure 9, and also equal in length to these distances by the assumption of a rotational velocity of
1 rad/s. Similarly, ad is perpendicular to AD, and od
is perpendicular to OD. In Fig. 7, only the significant vectors are shown for the analysis. In fact, the
arbitrary point analysis is the rst step to obtain
the initially unknown vector measures [5]. By using

Proc. IMechE. Vol. 219 Part D: J. Automobile Engineering

D08004 IMechE 2005

Rear suspension auxiliary roll damping

the measures from Fig. 7, then Fig. 8 can be obtained

to give the ratio of the spring to the wheel motions.
Considering the suspension shown in Fig. 9, an
arbitrary angular velocity is applied to the link BC.
bc is perpendicular to BC, and ad is perpendicular
to AD (see Figs 8 and 9). Both a and b are points of
zero velocity and therefore coincident with o. Since
the length bc in Fig. 7 is known from the application
of the arbitrary angular velocity to BC in Fig. 9,
similarly d is located since cd is perpendicular to CD;
o is located since oc is perpendicular to OC, and
od is perpendicular to OD.
By considering Figs 8 and 9, let [5]
oy
=n
oo

(14)

Therefore, the vertical force F can be shown to be

v
proportional to the spring force F according to
s
F =F n
(15)
v
s
and
qn
qF
qF
v =F
n s
s
qv
qv
qv

(16)

where v is the vertical displacement at o.

Therefore,

qn
qF
+Kn2
K = v = F
s qv
v
qv

25

If it is assumed that positive and negative suspension

movements are symmetrical, then the expression
may be simplied [6] to
qM
=2Kn2
w
qw

(21)

Since the springs and dampers are mounted concentrically and have the same displacement, the same
coecients are valid for the eective damping stiness. Addition of the roll damping provided by the
auxiliary dampers results in
qM
=2Cn2 +2C n2
w
Aux Aux
qw

(22)

Using the same approach outlined above, a similar

analysis is carried out on the front suspension
geometry. If the numerical values obtained from the
analyses are applied to equations (21) and (22), the
resulting eective roll stiness K and roll damping
w
C for both rear and front suspensions are presented
w
in Table 1. n is a constructive constant that is
w
determined in the initial design studies (function of
suspension geometry). Spring and damper rates are
the variable design parameters that can be modied
after the initial design.

(17)

Since both suspension springs are involved, the

subscripts L and R are used and M is the moment
about the centre O and considering Fig. 9, let [5]
n OO=n
w
From equation (15),

(18)

M=F (n )+F (+n )

(19)
sL
w
sR
w
Dierentiating equation (19) with respect to the roll
angle w gives
q(n )
q(+n )
qF
qM
w +F
w +(n ) sL
=F
sL qw
sR qw
w qw
qw
qF
+(+n ) sR
w qw

(20)

4 THE EFFECT OF ROCKER SHAPE ON LOAD

TRANSFER
The ratio of the velocity vector of the auxiliary
damper to that of the tyre is the key element for the
amount of additional roll damping obtainable from
the vehicles suspension [see equation (22) and the
velocity diagram in Fig. 8]. The higher the ratio
obtained, the higher is the roll damping that can be
provided by the load transfer mechanism. Higher roll
damping leads to a higher transient load transfer at
the rear and this can result in an increased oversteer
characteristic and, as a result, an improved cornering
performance. A very simple design change to the
rockers is suggested to obtain higher roll damping
from the system, which did not seem to be eective
with the original design parameters. The rocker is

Table 1 Stiness and damping values obtained by the velocity diagram approach

Front suspension
Rear suspension

D08004 IMechE 2005

Roll stiness
(N m/rad)

Auxiliary stiness
(N m/rad)

Roll damping
(N m s/rad)

Auxiliary damping
(N m s/rad)

0.991K
f
1.362K
r

3000

0.991C
f
1.362C
r

0.188C
Aux

Proc. IMechE. Vol. 219 Part D: J. Automobile Engineering

26

E Uzunsoy and O A Olatunbosun

provided to transfer the vertical motion of the tyre

to the struts, which are expected to absorb the disturbing force caused by the tyreroad interaction and
to keep the tyre in contact with the road. In doing
this, a load transfer is also developed from the inner
tyre to the outer tyre of a vehicle in a cornering
manoeuvre. Therefore, from the given geometry of
the vehicles suspension system, rockers are important
in terms of providing the motion ratio between the
tyre and struts.
The modied rear axle and rocker design are shown
in Figs 10 and 11. Figure 11a shows the original and
proposed rocker geometries while Fig. 11b presents
the modied form of the velocity diagram which was
originally presented in Fig. 8. K and X show the new

extended points of the rocker instead of K and X

respectively (Fig. 11a). The new shape of the rocker
is proposed as TLYKX instead of TLYKX.
From equation (18) and the velocity diagram given
in Fig. 11b, the ratio n =0.716. This ratio provides
w
4.096 times the auxiliary damping stiness of the
original design. The new coecients used in the calculation of roll stiness and damping are presented
in Table 2.
From the new resulting coecients, the new
damper rates required in the modied suspension
system to provide the same damper stiness value
as the original design are shown in Table 3. This table
provides much lower damper rates to be used instead
of those provided in the original design.

Fig. 10 Modied rear axle design for Car 201

Fig. 11 (a) Final modication for the rocker; (b) the resulting velocity diagram of the new rear
suspension system
Proc. IMechE. Vol. 219 Part D: J. Automobile Engineering

D08004 IMechE 2005

Rear suspension auxiliary roll damping

27

Table 2 Stiness and damping values obtained by the velocity diagram approach

Front suspension
Rear suspension

Roll stiness
(N m/rad)

Auxiliary stiness
(N m/rad)

Roll damping
(N m s/rad)

Auxiliary damping
(N m s/rad)

0.991K
f
1.704K
r

3000

0.991C
f
1.704C
r

1.0253C
Aux

Table 3 Equivalent damper rates provided by the

design modications
Auxiliary damper rates
(original design)
(N s/m)

Equivalent damper rates

(modied rear axle and rocker)
(N s/m)

750
1500
3000
6000
10 000

183
366
732
1464
2441

5 SIMULATION RESULTS AND DISCUSSION

Simulation of vehicle response was carried out
using a vehicle-handling dynamics simulation tool,
namely HANDSIM, which was developed by the
Vehicle Technology Research Centre, University
of Birmingham, using the MATLABASIMULINKA
environment [8]. The simulation was carried out at
a vehicle speed close to the maximum speed at which
the vehicle can traverse the given path without loss
of control. Vehicle speed is kept constant during the
manoeuvre by a fuzzy logic controller. Figures 12 to
16 show the vehicle responses of Car 201 with and
without the rear suspension auxiliary damper load
transfer mechanism. Car 201 has a front biased
weight distribution of 55 per cent front and 45 per
cent rear.

Fig. 13 Roll rate response of Car 201 at 43 km/h

Fig. 14 Resulting vehicle trajectory at 48 km/h with

2250 N s/m damper rate

5.1 Simulation results with the original vehicle

design parameters

Fig. 12 Resulting vehicle trajectory for Car 201 at

43 km/h
D08004 IMechE 2005

The parameters used in the vehicle simulation are

given in Table 4. The method is to run the simulation
both with and without the rear load transfer mechanism. The auxiliary damper rate is set to various
values to observe the handling response of the car
with the mechanism. The vehicle has an intrinsic
understeer characteristic due to the weight distribution, which is 55 per cent front and 45 per cent
Proc. IMechE. Vol. 219 Part D: J. Automobile Engineering

28

E Uzunsoy and O A Olatunbosun

Table 4 Vehicle control and simulation parameters

Vehicle speed (constant) (km/h)
Steer angle for constant-steer test (deg)
Path coordinates (m)
X
Y

Fig. 15 Front lateral load transfer of the vehicle with

and without the auxiliary dampers

Fig. 16 Rear lateral load transfer of the vehicle with

and without the auxiliary dampers

rear. However, the longitudinal velocity of the vehicle

and lateral load transfer and also the resultant tyre
slip angles determine the actual characteristics.
5.1.1 Vehicle-handling simulation at constant speed
and steer angle
Figure 12 shows the resulting paths for a constantspeed constant-steer-angle turn without the rear
load transfer mechanism and with it, using various
damping rates. The reductions in path radius provided by dierent auxiliary damper rates are shown
in the gure. The roll rate response of the vehicle is
shown in Fig. 13.
The simulation results provided by constant steer
angle and constant vehicle speed reveal important
information about Car 201. Firstly, even though the
eects due to the rear auxiliary dampers are small,
the theoretical basis explained is correct. Secondly,
Proc. IMechE. Vol. 219 Part D: J. Automobile Engineering

40
3
0, 30, 60, 90, 120,
105, 70, 40
0, 8, 0, 12, 25, 70,
50, 80

Wheelbase (m)
Front
Rear

0.742
0.908

Track (m)
Front
Rear
Mass of vehicle (including driver) (kg)
Roll inertia (kg m2)
Yaw inertia (kg m2)

1.342
1.28
300
20
200

Damper rate (N s/m)

Front
Rear
Auxiliary damper rate (N s/m)

750
750
Variable

Spring rate (N/m)

Front
Rear
Anti-roll bar rate (N m/rad)
Height of centre of gravity (m)

22 000
22 000
3000
0.285

Roll centre (m)

Front
Rear

0.0093
0.02325

Modied roll centre (m)

Rear

0.01602

however, the positive eect of the system can only

be obtained by using extremely high damper rates for
Car 201 with its original design parameters. On the
other hand, the need for these extreme damper rates
can be eliminated by modication of the motion
ratio of the tyre to the struts. For this reason, the
rocker which is responsible for transferring the tyre
motion to the struts and auxiliary dampers is
redesigned to improve the motion ratio.
5.2 Results for the modied suspension system at
constant speed and variable path
A simulation of the car traversing a winding circuit
is carried out with the original and modied suspension systems, respectively. Some of the results are
shown in Figs 14 and 15. A slight path improvement
was achieved with the modied suspension system
in Fig. 14. A sample part from the lateral load transfer
history of the vehicle (Figs 15 and 16) conrms the
initial increase in load transfer at the rear axle
(decreased load transfer at the front) with auxiliary
dampers compared with the original suspension.
This produces a transient increase in the oversteer
tendency of the vehicle, which disappears once the
roll velocity reduces to a small value.
D08004 IMechE 2005

Rear suspension auxiliary roll damping

6 CONCLUSIONS
Car 201 was designed to get round the twisting
Formula SAEFormula Student circuits as quickly as
possible and safely. The load transfer mechanism
was designed for the rear suspension system to help
achieve this objective. The simulation results of
the original Car 201 discussed above show that the
original design of the vehicle is not suitable enough
for this purpose. Car 201 is an understeering vehicle
which is not ideal in a racing car. However, this car
has been designed for the students who have little
opportunity to experience driving the car on the
track. Therefore, an understeering car which is easier
to control at the limit of traction may be the best
choice in the circumstances. The design of the suspension system very much aects the ride and handling properties of the vehicle. While a well-designed
rear load transfer mechanism can build an oversteering tendency on an understeering vehicle, the rear
suspension geometry of Car 201 with its long suspension links and narrow dierential box has been
shown not to be a good base for the rear load transfer mechanism. However, while keeping most of the
existing design, it has been possible to nd a cost
eective solution through the redesign of the load
transfer rocker to provide a more suitable wheel-todamper velocity ratio. It has been shown that the
rocker geometry is very important in load transfer
and the velocity diagram can be used to obtain a
suitable design.
Although a design change can also be made for
the front suspension geometry, possible changes may
be limited due to the initially designed leg placement
of the driver. Also, this study was especially focused
on the rear suspension system.
Under these conditions, the rear load transfer
mechanism is analysed and by a proper modication
of the rockers and axle, the benets of the system
are shown. As a result, it is shown that the rear load
transfer mechanism can be used in future models of
the Formula Student cars and, indeed, in road-going
cars, with a proper optimization of the struts and
anti-roll bar rates and the design of the links and
rockers, to improve the handling performance.

2 Uzunsoy, E. and Olatunbosun, O. A. A generic

fuzzy pilot and path planning for vehicle handling
dynamics simulation studies. In Proceedings of the
SAE International Conference on Future Transportation
Technology, Costa Mesa, California, 2003, SAE paper
2003-01-2263.
3 Uzunsoy, E. Application of fuzzy control in vehicle
handling simulation and path planning. PhD Dissertation, University of Birmingham, Birmingham,
2003.
4 Goodyear Tyres, Formula SAE tyre data: 20.06.5-13.
http://www.goodyear.com.
5 Fenton, J. Handbook of Automotive Design Analysis,
1976 (Butterworth, London).
6 Ellis, J. R. Vehicle Dynamics, 1969 (London Business
Books Limited, London).
7 Ellis, J. R. Vehicle Handling Dynamics, 1994 (Mechanical Engineering Publications Limited, London).
8 Uzunsoy, E. and Olatunbosun, O. A. A new vehicle
handling dynamics simulation tool: HANDSIM. In
Proceedings of the 7th Annual Research Symposium
of Postgraduate Research, University of Birmingham,
May 2001, pp. 2630.

APPENDIX
Notation
a
b
C

d
F
b
F
x
F
s
F
v
F
y
WF ,WF
x
y
h

cg

rc

REFERENCES
1 Uzunsoy, E. and Olatunbosun, O. A. A fuzzy logic
application on vehicle path planning and yaw control. In Proceedings of the 9th Annual Research
Symposium of Postgraduate Research, University of
Birmingham, 7 May 2003, pp. 5054.
D08004 IMechE 2005

29

I
xx
I
zz
K
K
s
K

distance from the front axle to the

centre of gravity (m)
distance from the rear axle to the
centre of gravity (m)
roll damping of the suspension
(N s/m)
distance between the roll centre and
the centre of gravity (m)
braking force (N)
tractive force (N)
spring force (N)
vertical force (N)
lateral force (N)
forces applied about the x and y axes
respectively (N)
height of the centre of gravity of the
vehicle (m)
height of the roll centre of the
vehicle (m)
moment of inertia about the x axis
moment of inertia about the z axis
actual spring stiness (N/m)
vertical rate of each of the left and
right springs (N/m)
eective spring stiness at wheel
(N/m)

Proc. IMechE. Vol. 219 Part D: J. Automobile Engineering

30

K
w
m
m
s
M
x
M
y
p, p
r, r
s
t
, U
U

E Uzunsoy and O A Olatunbosun

roll stiness of the suspension

total vehicle mass (kg)
sprung mass (kg)
total moment about the x axis (N m)
total moment about the y axis (N m)
roll acceleration and velocity
respectively of the vehicle
yaw acceleration and velocity
respectively of the vehicle
lateral separation of the springs (m)
track (m)
vehicle longitudinal acceleration
(m/s2) and velocity (m/s) respectively

Proc. IMechE. Vol. 219 Part D: J. Automobile Engineering

, V
V

vehicle lateral acceleration (m/s2)

and velocity (m/s) respectively

d
w

steer angle (rad)

roll angle (deg)

Subscripts
f
L
r
R

front
left
rear
right

D08004 IMechE 2005