This article was downloaded by: [RMIT University]

On: 07 October 2013, At: 08:47

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Vehicle System Dynamics:

International Journal of Vehicle
Mechanics and Mobility
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/nvsd20

A Motorcycle Multi-Body Model for

Real Time Simulations Based on
the Natural Coordinates Approach
Vittore Cossalter & Roberto Lot
Published online: 09 Aug 2010.

To cite this article: Vittore Cossalter & Roberto Lot (2002) A Motorcycle Multi-Body Model
for Real Time Simulations Based on the Natural Coordinates Approach, Vehicle System
Dynamics: International Journal of Vehicle Mechanics and Mobility, 37:6, 423-447

To link to this article: http://dx.doi.org/10.1076/vesd.37.6.423.3523

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information
(the “Content”) contained in the publications on our platform. However, Taylor
& Francis, our agents, and our licensors make no representations or warranties
whatsoever as to the accuracy, completeness, or suitability for any purpose
of the Content. Any opinions and views expressed in this publication are the
opinions and views of the authors, and are not the views of or endorsed by
Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and
Francis shall not be liable for any losses, actions, claims, proceedings, demands,
costs, expenses, damages, and other liabilities whatsoever or howsoever caused
arising directly or indirectly in connection with, in relation to or arising out of the
use of the Content.

This article may be used for research, teaching, and private study purposes.
Any substantial or systematic reproduction, redistribution, reselling, loan, sub-
licensing, systematic supply, or distribution in any form to anyone is expressly
 forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
Downloaded by [RMIT University] at 08:47 07 October 2013
 Vehicle System Dynamics 0042-3114/02/3706-423$16.00
2002, Vol. 37, No. 6, pp. 423±447 # Swets & Zeitlinger

A Motorcycle Multi-Body Model for Real Time

Simulations Based on the Natural
Coordinates Approach

VITTORE COSSALTER1 and ROBERTO LOT2

SUMMARY
Downloaded by [RMIT University] at 08:47 07 October 2013

This paper presents an eleven degrees of freedom, non-linear, multi-body dynamics model of a motorcycle.
Front and rear chassis, steering system, suspensions and tires are the main features of the model.
An original tire model was developed, which takes into account the geometric shape of tires and the
elastic deformation of tire carcasses. This model also describes the dynamic behavior of tires in a way
similar to relaxation models.
Equations of motion stem from the natural coordinates approach. First, each rigid body is described with
a set of fully cartesian coordinates. Then, links between the bodies are obtained by means of algebraic
equations. This makes it possible to obtain simple equations of motion, even though the coordinates are
redundant.
The model was implemented in a Fortran code, named FastBike. In order to test the code, both simulated
and real slalom and lane change maneuvers were carried out. A very good agreement between the numerical
simulations and experimental test was found. The comparison of FastBike's performance with those of some
commercial software shows that ®rst is much faster than others. In particular, real time simulations can be
carried out using FastBike and it can be employed on a motorcycle simulator.

1. INTRODUCTION

The use of computer simulations in motorcycle engineering makes it possible both to

reduce designing time and costs and to avoid the risks and dangers associated with
experiments and tests. The multi-body model for computer simulations can be built
either by developing a mathematical model of the vehicle or by using commercial
software for vehicle system dynamics. Even though the ®rst method is more dif®cult
and time consuming than the second, maximum ¯exibility in the description of the
features of the model can be obtained only by using a mathematical model. In
particular, it makes it possible to properly describe the tire behavior at large camber
1
Department of Mechanical Engineering, University of Padova, Italy.
2
Corresponding author: Roberto Lot, Department of Mechanical Engineering, University of Padova, Via
Venezia 1, 35131 Padova, Italy. Tel.: ‡39 049 8276806; Fax: ‡39 049 8276785; E-mail: [email protected];
website: www.dinamoto.mecc.unipd.it
 424 V. COSSALTER AND R. LOT

angles, whereas multi-body codes such as ADAMS, DADS or Visual Nastran lack
such a feature. Moreover, mathematical modeling has a high computation ef®ciency,
while multi-body software require a lot of time to carry out simulations.
For the reasons above, the focus of this study was to develop mathematical models
of a tire and motorcycle. The tire model properly describes the shape of the carcass
and the position of the contact point. Moreover, it takes into account the sliding of the
contact patch and the deformation of the tire carcass. The motorcycle model was
developed based on the natural coordinates approach [1], which makes it possible to
obtain simple equations of motion and hence high computation ef®ciency.

2. MOTORCYCLE AND RIDER DESCRIPTION

Downloaded by [RMIT University] at 08:47 07 October 2013

The motorcycle is modeled as a system of six bodies: the front and rear wheels, the
rear assembly (including frame, engine and fuel tank), the front assembly (including
steering column, handle-bar and front fork), the rear swinging arm and the unsprung
front mass (including fork and brake pliers). The driver is considered to be rigidly
attached to the rear assembly; front and rear assembly are linked by means of the
steering mechanism. The front suspension is a telescopic type and the rear suspension
is a swinging arm type.
This vehicle model has eleven degrees of freedom, which can be associated to the
coordinates of the rear assembly center of mass, the yaw angle, the roll angle, the
pitch angle, the steering angle, the travel of front and rear suspension and the spin
rotation of both wheels (see Fig. 1).

Fig. 1. Eleven degrees of freedom motorcycle model.

 A MOTORCYCLE MULTI-BODY MODEL 425

The following forces act on the motorcycle elements: suspensions forces due to
springs and shock-absorbers, tire forces and torques, aerodynamic forces, rider steering
torque, steer damper torque, rear and front brake torques and ®nally propulsive torque,
which is transmitted from the sprocket to the rear wheel by means of the chain.
The rider's actions on the motorcycle determine both the direction of the vehicle
and the forward speed. In this model, the rider is considered to be a rigid body
attached to the rear assembly, so that the rider's movement away from the saddle and
the corresponding control action are neglected. In this way the motorcycle's direction
is controlled only by the torque exerted on the handlebars (steering torque). The
forward speed is controlled by applying the brakes (rear and front brake torques) and
by acting on the accelerator lever (propulsive force).
Downloaded by [RMIT University] at 08:47 07 October 2013

3. TIRE MODEL

In motorcycles the roll angle can reach 50±55 , hence it has a signi®cant in¯uence
both on tire forces and torques and on the contact patch. In this model, the actual
shape of the tire is described in detail and the deformation of the tire carcass is taken
into account. The road±tire contact is assumed to be dot-shaped and the position of the
contact point depends on the roll angle. Tire forces and torques are applied in the
contact point. The tire forces include the vertical load N, the lateral force F and the
longitudinal force S; the tire torques include the rolling friction torque My and the yaw
torque Mz .
The tire reference frame Tw is de®ned by using 4  4 transformation matrix
notation [2], as shown in Figure 2: its origin is located in wheel center G, plane Xw Zw
is the symmetry plane of the wheel, the Xw axis is horizontal and points forwards, the
Yw axis is parallel to the wheel spin axis and points rightwards and the Zw axis
completes the reference frame. The frame T 0 has its origin located in contact point C,
the road plane X 0 Y 0 is horizontal, the X 0 axis is parallel to Xw, points forwards and has
unit vector s, the Y 0 axis points rightward and has unit vector n, the Z 0 axis is vertical
and points downwards.
As it is well known, horizontal tire forces depend on tread deformation and slide,
i.e., they depend on sideslip angle l, longitudinal slip k, camber angle j and vertical
load N as follows
S ˆ Sslip k; l; j; N †
1†
F ˆ Fslip k; l; j; N †
In several tire models [3±5] the sideslip angle and longitudinal slip are de®ned
according to wheel kinematics, without taking into account the deformation of the tire
carcass. On the contrary, in this model slip quantities are de®ned considering the
actual contact point, which moves with respect to the rim because of the deformation
 426 V. COSSALTER AND R. LOT
Downloaded by [RMIT University] at 08:47 07 October 2013

Fig. 2. Tire kinematics and tire forces.

Fig. 3. Tire deformability.

of tire carcass. Deformability of the tire carcass is taken into account as shown in
Figure 3. The contact point lies on the vertical plane which passes through the wheel
spin axis. The tire de¯ection with respect to the rim consists of radial displacement r ,
lateral displacement l and rotation x around the wheel spin axis. Moreover, it is
assumed that tire deformations do not alter the mass properties of the wheel.
 A MOTORCYCLE MULTI-BODY MODEL 427

The position of the contact point is expressed by means of its coordinates yc ; zc

with respect to frame Tw as follows
C ˆ Tw f0; yc ; zc ; 1gT 2†
Thus, the instantaneous sideslip angle is de®ned as:
VY C_
n
l ˆ ÿarctan ˆ ÿarctan 3†
VX C_
s
where VX is the forward speed, VY the lateral speed, s and n the unit vectors of axis X 0
and Y 0 respectively.
The instantaneous longitudinal slip is de®ned as:
Downloaded by [RMIT University] at 08:47 07 October 2013

VR zc y_ ‡ x†_
k ˆ ÿ1 ÿ ˆ ÿ1 ÿ 4†
VX C_
s
where VR is the rolling speed which depends both on spin velocity y_ and rotational
_
deformation rate x.
On the other hand, tire forces depend on carcass deformation and camber angle, as
shown in experimental tests [6, 7]
S ˆ Selastic x; j†
F ˆ Felastic r ; l ; j† 5†
N ˆ Nelastic r ; l ; j†
In absence of tire forces, no tire de¯ection is present and the contact point
coincides with the point of tangency between the tire surface and road plane C0. Thus,
the position of the contact point only depends on the tire shape and the coordinates of
C0 with respect to frame Tw can be de®ned as a function of the roll angle, as follows
C0 ˆ Tw f0; yt j†; zt j†; 1gT 6†
where functions yt j† and zt j† make a parametric representation of the lateral pro®le
of the carcass. In order to guarantee the condition of tangency between tire and road
plane, functions must satisfy the following relation

dzt dyt
tan j† ˆ ÿ
dj dj
Lateral and radial deformation can be calculated by subtracting expression (6)
from expression (2), obtaining
l ˆ yc ÿ yt j †
7†
 r ˆ zc ÿ zt j †
 428 V. COSSALTER AND R. LOT

This model is able to properly describe tire behavior both in steady state and
transient conditions. Indeed, by coupling Equation (1), which describe the behavior
of the contact patch during sliding, with Equation (5), which describe elasticity
properties of the tire carcass
Sslip k; l; j; N † ÿ Selastic x; j† ˆ 0
8†
Fslip k; l; j; N † ÿ Felastic r ; l ; j† ˆ 0
one obtains a description of tire behavior which is equivalent to relaxation tire models
[8±11]. To proof this, let us de®ne a linear relation between longitudinal force and
longitudinal slip
S ˆ Ks k 9†
Downloaded by [RMIT University] at 08:47 07 October 2013

and a linear relation between longitudinal force and rotational deformation

S ˆ Kx x 10†
where Ks and Kx are respectively the longitudinal slip stiffness and rotational stiffness
of tire. By substituting Equation (4) in Equation (9) and by rearranging terms, one
obtains:
!
zc y_ zc x_ zc x_
S ˆ KS ÿ1 ÿ ÿ KS ˆ KS k0 ÿ KS 11†
VX VX VX

where k0 is the steady state value of longitudinal slip, which corresponds to the steady
state value of longitudinal force S0 : The time derivation of expression (10) yields:
S_
x_ ˆ 12†
Kx
By replacing Equations (12) in Equation (11) and by rearranging the terms, one
obtains:
KS zc =Kx _
S ‡ S ˆ S0 13†
VX
which is a ®rst order relaxation equation, where relaxation length is s ˆ KS zc =Kx . The
equivalence between this tire model and the relaxation model can be found for lateral
force as well.
This approach presents several advantages with respect to relaxation models. First,
it explains the physical behavior of the tire in more detail, by highlighting both the
deformability of the carcass and the sliding of the tread. Furthermore, with this tire
model only static and steady state experimental tests are required in order to
characterize tire behavior in both static and dynamic conditions.
 A MOTORCYCLE MULTI-BODY MODEL 429

In order to complete the model it is necessary to de®ne tire torques with respect to
the contact point. The rolling resistance torque is assumed to be proportional to the
wheel load
My ˆ N d 14†
where d is the rolling friction parameter.
Yaw torque Mz is generated by lateral force F, tire trail t and twisting torque MTz as
follows [12±14]:
Mz ˆ ÿt l†F ‡ MTz j† 15†
The ®rst term depends on the sideslip angle and tends to align, the second term
depends on the roll angle and tends to self-steer.
Finally, it is not necessary to take into account overturning moment Mx , because
Downloaded by [RMIT University] at 08:47 07 October 2013

tire forces are applied in the actual contact point [3, 13, 14].

4. MULTI-BODY MODEL

The mathematical model of the motorcycle was developed based on the natural
coordinates approach [1]. Natural coordinates consist of cartesian coordinates of
points or direction cosines of vectors belonging to the bodies of the system. With this
approach, kinematic relationships and equations of motion are very simple. However,
the number of variables required for describing a system is larger than the number of
degrees of freedom and so additional constraint equations must be introduced.
The equations were derived using Maple1, a software which makes it possible to
perform symbolic manipulation ef®ciently and to avoid calculation errors. Moreover,
it generates automatically the Fortran code.

4.1. Kinematic Description

Equations of motion were derived in the inertial reference frame XYZ: axes X and Y
are horizontal and lie on the road level, the Z axis is vertical and points downwards;
the unit vectors of inertial frame are, respectively, cx , cy and cz .
A body-®xed frame Ti is attached to each rigid body. The elements of the
transformation matrix are used as generalized coordinates, i.e., the con®guration of
each body is described by means of the coordinates of origin and direction cosines of
the body-®xed frame (see Fig. 4).
The rear tire reference frame Tw1 has its origin in the center of the wheel
G1 ˆ fx1 ; y1 ; z1 ; 1gT and is de®ned as shown in Section 3, as well as the reference
frame T 01 . Moreover, the rear wheel ®xed-frame T1 is obtained from frame Tw1
by a rotation of spin angle y1 around Yw1 axis. It is useful to de®ne the follow-
ing unit vectors: s1 ˆ fsx1 ; sy1 ; 0; 0gT parallel to both Xw1 and X10 axes,
 430 V. COSSALTER AND R. LOT
Downloaded by [RMIT University] at 08:47 07 October 2013

Fig. 4. Description of multi-body system using basic points and unit vectors.

w1 ˆ fwx1 ; wy1 ; wz1 ; 0gT parallel to axis Yw1 , v1 ˆ fvx1 ; yy1 ; vz1 ; 0gT parallel to axis
Zw1 and n1 ˆ fÿsy1 ; sx1 ; 0; 0gT parallel to axis Y 01 .
The rear assembly ®xed-frame T2 has its origin in the swinging arm pin joint
P2 ˆ fx2 ; y2 ; z2 ; 1gT ; plane X2Z2 is parallel to plane X1 Z1 , the X2 axis is perpendicular
to the steering axis, points forwards and has unit vector u2 ˆ fux2 ; uy2 ; uz2 ; 0gT , the Y2
axis has unit vector w2 ˆ w1 and ®nally the Z2 axis is parallel to the steering axis and
has unit vector v2 ˆ fvx2 ; yy2 ; vz2 ; 0gT .
The front assembly ®xed-frame T3 has the origin in the point P3 ˆ fx3 ; y3 ; z3 ; 1gT ,
which is the intersection between the steering axis and its perpendicular plane passing
through P2 . The X3 Z3 plane is parallel to the symmetry plane of the front wheel, the
X3 axis is perpendicular to the steering axis, points forwards and has unit vector
u3 ˆ fux3 ; uy3 ; uz3 ; 0gT , the Y3 axis is parallel to the front wheel spin axis and has unit
vector w3 ˆ fwx4 ; wy4 ; wz4 ; 0gT , ®nally the Z3 axis has a unit vector v3 ˆ v2 .
The front tire reference frame Tw4 has its origin in the center of the wheel
G4 ˆ fx4 ; y4 ; z4 ; 1gT and is de®ned as shown in Section 3, as well as the reference
frame T 04 . Besides, the front wheel ®xed-frame T4 is obtained from frame Tw4 by a
rotation of spin angle y4 around Yw4 axis. The following unit vectors are de®ned:
s4 ˆ fsx4 ; sy4 ; 0; 0gT parallel to both Xw4 and X40 axis, w4 ˆ w3 parallel to Yw4 axis,
v4 ˆ fvx4 ; yy4 ; vz4 ; 0gT parallel to Zw4 axis and n 4 ˆ fÿsy4 ; sx4 ; 0; 0gT parallel to Y 04 axis.
 A MOTORCYCLE MULTI-BODY MODEL 431

The swinging arm ®xed-frame T5 has its origin in the rear wheel center G1 , the X5
axis is parallel to vector G1 P2 and has unit vector u5 ˆ fux5 ; uy5 ; uz5 ; 0gT , the Y5 axis
has unit vector w5 ˆ w1 and the Z5 axis has unit vector v5 ˆ fvx5 ; yy5 ; vz5 ; 0gT .
The front unsprung mass ®xed-frame T6 has the origin on the center of mass
G6 ˆ T4 fGx6 ; Gy6 ; Gz6 ; 1gT ; X6 ; Y6 and Z6 axes are parallel respectively to X3 ; Y3 and
Z3 and their unit vectors are u6 ˆ u3 , w6 ˆ w4 , v6 ˆ v2 .
The con®guration of the motorcycle is described by means of a set of n ˆ 45
coordinates, including the coordinates of points G1 , P2 , P3 , G4 , direction cosines of
unit vectors s1 , v1 , w1 , u2 , v2 , u3 , s4 , v4 , w4 , u5 , v5 and spin rotations of both wheels:
q ˆ fx1 ; y1 ; z1 ; sx1 ; sy1 ; wx1 ; wy1 ; wz1 ; vx1 ; vy1 ; vz1 ; y1 ; x2 ; y2 ; z2 ; ux2 ; uy2 ; uz2 ; vx2 ; vy2 ;
vz2 ; x3 ; y3 ; z3 ; ux3 ; uy3 ; uz3 ; x4 ; y4 ; z4 ; sx4 ; sy4 ; wx4 ; wy4 ; wz4 ; vx4 ; vy4 ; vz4 ; y4 ; ux5 ;
Downloaded by [RMIT University] at 08:47 07 October 2013

uy5 ; uz5 ; vx5 ; vy5 ; vz5 gT 16†

The motorcycle has only f ˆ 11 degrees of freedom, thus it is necessary to
formulate a set of m ˆ n ÿ f ˆ 34 independent constraint equations:
fj ˆ 0; j ˆ 1 . . . m 17†
By imposing the unit length condition to all unit vectors, the following 11
independent constraint equations are obtained:
f1 ˆ s1
s1 ÿ 1 f2 ˆ w1
w1 ÿ 1 f3 ˆ v1
v1 ÿ 1
f4 ˆ u2
u2 ÿ 1 f5 ˆ v 2
v 2 ÿ 1 f6 ˆ u3
u3 ÿ 1
17:1ÿ11†
f7 ˆ s4
s4 ÿ 1 f8 ˆ w4
w4 ÿ 1 f9 ˆ v4
v4 ÿ 1
f10 ˆ u5
u5 ÿ 1 f11 ˆ v5
v5 ÿ 1
By imposing the orthogonal conditions to every couple of unit vectors which belong
to the same reference frame, 15 more independent constraint equations are obtained:
f12 ˆ s1
w1 f13 ˆ s1
v1 f14 ˆ v1
w1
f15 ˆ u2
w1 f16 ˆ v2
u2 f17 ˆ v2
w1
f18 ˆ u3
v2 f19 ˆ u3
w4 f20 ˆ v2
w4 17:12ÿ26†
f12 ˆ s4
w4 f22 ˆ s4
v4 f23 ˆ v4
w4
f24 ˆ u5
w1 f25 ˆ v5
w1 f26 ˆ w5
v5
The remaining 8 constraint equations are the following:
 vector G1 P2 must be perpendicular to the f27 ˆ G1 P2
w1 (17.27)
rear wheel spin axis Y1
 the magnitude of vector G1 P2 must be f28 ˆ G1 P2
G1 P2 ÿ l2f (17.28)
equal to the swinging arm length lf
 vector v5 must be perpendicular to the f29 ˆ G1 P2
v5 (17.29)
vector G1P2
 432 V. COSSALTER AND R. LOT

 the magnitude of vector P2P3 must be f30 ˆ P2 P3
P2 P3 ÿ l223 (17.30)

equal to l23
 vector P2P3 must lie on the X2Z2 plane f31 ˆ P2 P3
w1 (17.31)
(thus it must be perpendicular to the f32 ˆ P2 P3
v2 (17.32)
vectors w1 and v2)
 the point R3 ˆ G4 ÿ l1 u3 must lie on the f33 ˆ P3 R3
w4 (17.33)
steering axis Z3
(thus it must be perpendicular to vectors f34 ˆ P3 R3
u3 (17.34)
w4 and u3)
It is worth pointing out that the natural coordinates approach made it possible to obtain
simple constraint equations, which are quadratic with respect to the coordinates.
Downloaded by [RMIT University] at 08:47 07 October 2013

4.2. Lagrange's Equations

Due to the presence of constraints, the Lagrange's equations become
d @K @K X m
@fj
ÿ ‡ lj ÿ Qi ˆ 0; i ˆ 1: : n 18†
dt @ q_ i @qi jˆ1 @qi

where K is the kinetic energy, li are the Lagrange multipliers and Qi the generalized
forces.
By coupling the de®nition of kinetic energy to the transformation matrix notation,
the kinetic energy of ith rigid body is
Z Z
1 1
Ki ˆ P_ 2 dm ˆ fx; y; z; 1gT_ Ti T_ i fx; y; z; 1gT dm 19†
2 m 2 m
where fx; y; z; 1gT are the coordinates of point P with respect to frame Ti . Assuming
that the origin of the reference frame is the center of mass of the body and expanding
the previous equation, one obtains:
2 3
u_ 2i _i
u_ i
w_ i u_ i
v_ i u_ i
G
Z 6 _i7
1 6 w_ i
u_ i w_ 2i w_ i
v_ i w_ i
G 7
Ti ˆ fx; y; x; 1g6
6
7 fx; y; x; 1gT dm
7
2 m 4 v_ i
u_ i v_ i
w_ i v_ 2 _
v_ i
Gi 5
i
G_ i
u_ i G _ i
w_ i G_ i
v_ i _2
G i
Z Z Z Z
1 _2 1 2 2 1 2 2 1 2
ˆ Gi dm ‡ u_ i x dm ‡ w_ i y dm ‡ v_ i z2 dm
2 m 2 m 2 m 2 m
Z Z Z
‡ u_ i
w_ i xy dm ‡ u_ i
v_ i xz dm ‡ w_ i
v_ i yz dm
Zm Zm Zm
_ _
‡ u_ i
Gi x dm ‡ v_ i
Gi y dm ‡ w_ i
Gi z dm _
m m m
 A MOTORCYCLE MULTI-BODY MODEL 433

By substituting the integral terms in the previous equation with moments and products
of inertia with respect to the center of mass, the kinetic energy of each rigid body can
be calculated as a function of the elements of transformation matrix Ti , as follows
1 _2 1 ÿ 2
1 ÿ

Ki ˆ m i G i ‡ Ix;i ÿu _ i ‡ w_ 2i ‡ v_ 2i ‡ Iy;i u_ 2i ÿ w_ 2i ‡ v_ 2i
2 4 4
1 ÿ

‡ Iz;i u_ 2i ‡ w_ 2i ÿ v_ 2i ‡ Cxz;i u_ i
v_ i ‡ Cxy;i u_ i
w_ i ‡ Cyz;i w_ i
v_ i 20†
4
If the body center of mass does not coincide with the origin of the reference frame, it
is necessary to replace G _ i ˆ T_ i fGxi ; Gyi ; Gzi ; 1gT in the
_ i ˆ f_xi ; y_ i ; z_ i ; 1gT with G
previous equation. Thus, the kinetic energy of the whole system is:
1 h i
_ 2 ‡ 1 Iy1 s_ 2 ÿ w_ 2 ‡ v_ 2 ‡ y_ 1 s1
v_ 1 ÿ s_ 1
v1 † ‡ y_ 2 ‡ 1 Id1 w_ 2
Downloaded by [RMIT University] at 08:47 07 October 2013

K ˆ m1 G 1 1 1 1 1 1
2 2 2
1 ÿ
ÿ

‡ m2 G _ 2 ‡ 1 Ix2 ÿu_ 2 ‡ w_ 2 ‡ v_ 2 ‡ 1 Iy2 u_ 2 ÿ w_ 2 ‡ v_ 2
2 2 1 2 2 1 2
2 4 4
1 ÿ

‡ Iz2 u_ 22 ‡ w_ 22 ÿ v_ 22 ‡ Cxz2 u_ 2
v_ 2 ‡ Cxy2 u_ 2
w_ 2 ‡ Cyz2 w_ 2
v_ 2
4
1 ÿ
ÿ

‡ m3 G _ 2 ‡ 1 Ix3 ÿu_ 2 ‡ w_ 2 ‡ v_ 2 ‡ 1 Iy3 u_ 2 ÿ w_ 2 ‡ v_ 2
3 3 4 3 3 4 3
2 4 4
1 ÿ 2 2 2


‡ Iz3 u_ 3 ‡ w_ 4 ÿ v_ 3 ‡ Cxz3 u_ 3
v_ 3 ‡ Cxy3 u_ 3
w_ 4 ‡ Cyz3 w_ 4
v_ 3
4
1 h i
‡ m4 G _ 2 ‡ 1 Iy4 s_ 2 ÿ w_ 2 ‡ v_ 2 ‡ y_ 4 s4
v_ 4 ÿ s_ 4
v4 † ‡ y_ 2 ‡ 1 Id4 w_ 2
4 4 4 4 4 4
2 2 2
1 ÿ
ÿ

‡ m5 G _ 2 ‡ 1 Ix5 ÿu_ 2 ‡ w_ 2 ‡ v_ 2 ‡ 1 Iy5 u_ 2 ÿ w_ 2 ‡ v_ 2
5 5 1 5 5 1 5
2 4 4
1 ÿ

‡ Iz5 u_ 25 ‡ w_ 21 ÿ v_ 25 ‡ Cxz5 u_ 5
v_ 5 ‡ Cxy5 u_ 5
w_ 1 ‡ Cyz5 w_ 1
v_ 5
4
1 ÿ
ÿ

‡ m6 G _ 2 ‡ 1 Ix6 ÿu_ 2 ‡ w_ 2 ‡ v_ 2 ‡ 1 Iy6 u_ 2 ÿ w_ 2 ‡ v_ 2
6 3 4 3 3 4 3
2 4 4
1 ÿ

‡ Iz6 u_ 23 ‡ w_ 24 ÿ v_ 23 ‡ Cxz6 u_ 3
v_ 3 ‡ Cxy6 u_ 3
w_ 4 ‡ Cyz6 w_ 4
v_ 3 21†
4
where the terms relative to wheels (i ˆ 1 and i ˆ 4) are slightly different from the
terms relative to other bodies because of the axial symmetric structure of the wheels
(Ix;i ˆ Iz;i ˆ Id;i and Cxz;i ˆ Cyz;i ˆ Cxy;i ˆ 0) and because spin velocity y_ 1 ; y_ 4 has
been used.
The generalized forces expression can be obtained from the virtual work dW of the
forces acting on the vehicle
Xm
dW ˆ Qi dqi 22†
iˆ1
 434 V. COSSALTER AND R. LOT

In order to determine virtual works, it is necessary to calculate the virtual rotation dYi
of each rigid body with respect to its reference frame Ti . By extending the concept of
angular velocity matrix [2] to virtual rotation matrix dY ˆ TT dT and by extracting
the components of virtual rotation from dY, the following virtual rotation operator
can be de®ned:
dY Ti † ˆ fvi
dwi ; ui
dvi ; wi
dui ; 0gT 23†
Virtual work contains the following terms:
dW ˆ dWg ‡ dWS ‡ dWA ‡ dWt ‡ dWB ‡ dWt;F ‡ dWt;T ‡ dWP 24†
 The virtual work due to the gravity force:
X
6
Downloaded by [RMIT University] at 08:47 07 October 2013

dWg ˆ mi g
dGi 24:1†

iˆ1
where g ˆ f0; 0; g; 1gT is the gravity acceleration.
 The virtual work due to front suspension force FSf , which acts between the front
assembly and front wheel, and virtual work due to rear suspension force FSr , which
acts between the rear assembly and swinging arm:
dWs ˆ FSf v2
dP3 ÿ dR3 † ‡ ts FSr cy
‰dY T2 † ÿ dY T5 †Š 24:2†
where ts ˆ @yr =@zr is the velocity coef®cient between spring de¯ection zr and arm
rotation yr .
 The virtual work due to drag, side and lift aerodynamics forces FA ˆ
Tw1 fFD ; FS ; FL ; 0gT , which are applied on point CA ˆ T2 fXCA ; 0; ZCA ; 1gT :
dWA ˆ FA
dCA 24:3†
 The virtual work due to rider steering torque t and steer damper torque tD , which
are applied between the rear and front assembly:
dWt ˆ t ‡ tD †cz
‰dY T3 † ÿ dY T2 †Š 24:4†
 The virtual work due to rear brake torque MBr , which acts between the rear wheel
and swinging arm, and the virtual work due to front brake torque MBf , which acts
between the front wheel and front unsprung mass:
dWB ˆ MBr cy
‰dY T1 † ÿ dY T5 †Š ‡ MBf cy
‰dY T4 † ÿ dY T6 †Š 24:5†
 The virtual work due to rear tire force FT1 ˆ T01 fS1 ; F1 ; ÿN1 ; 0gT , which is applied
on rear contact point C1 ˆ Tw1 f0; yc1 ; zc1 ; 1gT , and the virtual work due to front tire
force FT4 ˆ T04 fS4 ; F4 ; ÿN4 ; 0gT , which is applied on front contact point C4 ˆ
Tw4 f0; yc4 ; zc4 ; 1gT :
dWt;F ˆ FT1
dG1 ‡G1 C1 FT1
T1 dY T1 †‡FT4
dG4 ‡G4 C4  FT4
T4 dY T4 †
24:6†
 A MOTORCYCLE MULTI-BODY MODEL 435

Fig. 5. Geometry of the chain transmission.

Downloaded by [RMIT University] at 08:47 07 October 2013

 The virtual work due to rear tire torque MT1 ˆ T01 f0; My1 ; Mz1 ; 0gT and front tire
torque MT4 ˆ T04 f0; My4 ; Mz4 ; 0gT :
dWt;M ˆ MT1
T1 dY T1 † ‡ MT4
T4 dY T4 † 24:7†
 The virtual work due to the propulsive torque, which is transmitted from the drive
sprocket to the wheel by means of the chain. As shown in Figure 5, the drive
sprocket center is R ˆ T2 fRX ; 0; RZ ; 1gT , whereas the chain angles are:
   
G1 R
s1 rc ÿ rp
yc1 ˆ arctan ÿ arcsin
G1 R
v1 jG1 Rj
   
G1 R
u2 rc ÿ rp
yc2 ˆ arctan ‡ arcsin
G 1 R
v2 jG1 Rj

The chain tension FC ˆ T1 fTc sin yc1 †; 0; Tc cos yc1 †; 0gT acts between point
P7 ˆ T1 frc cos yc1 †; 0; rc sin yc1 †; 1gT and point P8 ˆ T2 RX ‡ rp cos yc2 †; 0; RZ ‡
rp sin yc2 †; 1gT , thus the virtual work is
dWp ˆ Fc
dP7 ÿ dP8 † ÿ Tc rc dy1 24:8†

Explicit Lagrange's equations are not shown because of their large number, while
their compact form is the following:

_ q
F q; q;  ; k; t† ˆ M _ q_ ‡ FT k ÿ Q ˆ 0
q‡M 25†

where M is the mass matrix, F is the Jacobian matrix of constraint equations (17), k is
the vector of Lagrange multipliers and Q is the vector of generalized forces. Due to
the natural coordinates approach, the mass matrix is very sparse and has only 9% non-
zero elements; moreover the evaluation of Equation (25) require less than 2,000
multiplications and less than 1,000 additions.
 436 V. COSSALTER AND R. LOT

4.3. Tire Equations

As seen in Section 3, tire deformation is described by means of three coordinates,
hence for both the rear and front tires the following six coordinates should be de®ned:
q0 ˆ fyc1 ; zc1 ; x1 ; yc4 ; zc4 ; x4 gT 26†
The tire behavior must be described by means of as many equations as coordinates.
Equation (8) can be re-written as follows
p1 ˆ Sslip;1 k1 ; l1 ; j1 ; N1 † ÿ Selastic;1 x1 ; j1 † ˆ 0
p2 ˆ Sslip;4 k4 ; l4 ; j4 ; N4 † ÿ Selastic;4 x4 ; j4 † ˆ 0
ÿ
27:1ÿ4†
p3 ˆ Fslip;1 k1 ; l1 ; j1 ; N1 † ÿ Felastic;1 r;1 ; l;1 ; j1 ˆ 0
ÿ

p4 ˆ Fslip;4 k4 ; l4 ; j4 ; N4 † ÿ Felastic;4 r;4 ; l;4 ; j4 ˆ 0
Downloaded by [RMIT University] at 08:47 07 October 2013

Equations (3), (4) and (7) make it possible to express slip quantities and tire
deformations as a function of generalized coordinates, whereas camber angles can be
calculated as follows:
j1 ˆ arcsin wz1 †
28†
j4 ˆ arcsin wz4 †
The remaining equations are obtained by imposing the contact between the tire and
road plane Z ˆ 0, as follows:
p5 ˆ C1 †z ˆ z1 ‡ wz1 yc1 ‡ vz1 zc1 ˆ 0
27:5ÿ6†
p6 ˆ C4 †z ˆ z4 ‡ wz4 yc4 ‡ vz4 zc4 ˆ 0

It is worth pointing out that Equations (27.1±4) are differential equations because
slip quantities (3) and (4) contain time derivation of coordinates x and x0 . On the
contrary, Equations (27.5±6) are algebraic.

4.4. State Space Formulation

Lagrange's Equation (25), constraint Equation (17) and tire Equation (27) form a set
of 85 second order differential-algebraic simultaneous equations (DAEs) of index 3
[15], with the following unknowns: 51 generalized coordinates and 34 Lagrange
multipliers.
In order to obtain a 1 index DAEs problem, algebraic constraint Equation (3)
should be replaced by differential equations using the Baumgarte stabilization method
[16], as follows:
 ‡ 2&o/_ ‡ o2 /
/0 ˆ / 29†
where constant  and o are properly chosen.
 A MOTORCYCLE MULTI-BODY MODEL 437

The DAEs problems of index 1 can be numerically solved using the DASSL solver
[17], however the transformation of DAEs into a set of ordinary differential equations
(ODEs) makes it possible to increase integration speed. For this purpose, the
Lagrange multipliers are replaced with the following differential expression:
k ˆ c ‡ t0 c_ 30†
where constant t0 is properly chosen. Moreover, tire Equation (27) should be replaced
by the following set of ODEs
T
p0 ˆ fp1 ; p2 ; p3 ; p4 ; p5 ‡ t0 p_ 5 ; p6 ‡ t0 p_ 6 g 31†
In addition, the 2nd order Lagrange's Equation (25) should be reduced to a 1st order
ODEs. The system is then described by means of the following 2n ‡ m ‡ 6 ˆ 130
Downloaded by [RMIT University] at 08:47 07 October 2013

state variables
T
y ˆ fx; v; c; x0 g 32†
and the following state space equations
8 9
>
> F > >
< =
v ÿ q_
_ t† ˆ
G y; y; 0 ˆ0 33†
> / >
>
: >
;
p0
Although the number of equations is rather high with respect to the number of degrees
of freedom, each equation is simple and the evaluation of expression (33) require less
than 3,000 multiplications and less than 2,000 additions. These equations have been
implemented in a Fortran code, using the implicit solver DASSL for numerical
integration.

5. COMPARISON BETWEEN COMPUTER SIMULATIONS

AND EXPERIMENTAL MEASUREMENTS

In order to validate the multi-body model, some experimental tests were carried out on
an Aprilia RSV 1000 motorcycle; they were then compared to the simulation results.
The geometrical and inertial characteristics of the motorcycle and the non-linear
elastic and damping characteristics of the suspensions were measured at the
Department of Mechanical Engineering (DIM) at the University of Padua [18, 19].
Tire parameters were also measured with department's equipment [20], whereas the
driver inertia properties were estimated as shown in reference [21]. The charac-
teristics of the motorcycle are given in Appendix and in Figures 10, 11 and 12.
The motorcycle was equipped with a measurement system: roll and yaw rate,
steering angle, spin velocity of both wheels and steering torque were measured and
 438 V. COSSALTER AND R. LOT

stored on a data recorder [19]. Data post-processing made it possible to calculate

vehicle forward speed and roll angle as well.
In order to reproduce the experimental maneuvers by means of numerical
simulations, steering torque t was calculated according to measured steering torque
tm and measured roll angle jm , as follows:
t ˆ tm ‡ kj jm ÿ j† 34†
where j is the simulated roll angle and kj the control gain. The chain propulsive force
and the front brake torque were calculated based on measured speed um , as follows:
S ˆ mr ‰u_ m ‡ ku um ÿ u†Š
8
< T c ˆ r1 S MFf ˆ 0; S0 acceleration† 35†
Downloaded by [RMIT University] at 08:47 07 October 2013

rc
:
Tc ˆ 0 MFf ˆ ÿr4 S; S<0 braking†

where S is the longitudinal thrust, mr the generalized mass, u the simulated speed and ku
the control gain. Rear brake was not used in either the real or simulated maneuvers.
Figure 6 shows the comparison of the experimental measurements with the
numerical simulation for a lane change maneuver. The lane change width was 3.6 m
and the lane change length was 40 m. It was not possible to measure the trajectory of
the motorcycle, so the experiments were compared with simulations by analyzing
steering torque (Fig. 6a), vehicle speed (Fig. 6b), roll angle (Fig. 6c) and steering
angle (Fig. 6d). The ®gure shows that at the beginning of the maneuver the rider is
driving straight and increasing speed. When he starts to apply positive steering torque
(point A), the vehicle begins to capsize on the left-hand side. Afterwards, when the
steering torque is zero (point B) the magnitude of roll angle is still increasing; when
the steering torque reaches its minimum (point C), the roll angle is increasing and the
vehicle begins to capsize on the right-hand side. Then, the rider straightens the vehicle
(from point D) and ®nally decreases the speed.
The agreement between experimental and simulated data is very good: the overall
error (RMS) of steering torque is less than 3% of its peak value, the overall error of
vehicle speed is less than 0.5% of its peak value, the overall error of roll angle is about
the 9% of its peak value and the overall error of steering angle is about 26% of its peak
value. The steering angle has the maximum error, because of some steering oscillations
that are present in the simulation but that were not found in the experimental test.
Figure 7 shows the comparison of a real slalom maneuver with a simulated one,
by representing steering torque (Fig. 7a), vehicle speed (Fig. 7b), roll angle (Fig. 7c)
and steering angle (Fig. 7d). The pylon distance is 14 m and the vehicle speed is about
13.5 m/s. During the slalom maneuver, both roll and steering angles are delayed in
phase from steering torque of about 90 . Once again, the agreement between
experimental and simulated data is very good: the overall error of steering torque is
 Downloaded by [RMIT University] at 08:47 07 October 2013

A MOTORCYCLE MULTI-BODY MODEL

Fig. 6. Lane change maneuver: comparison between experimental measurements and numerical simulations.
439
 Downloaded by [RMIT University] at 08:47 07 October 2013

440
V. COSSALTER AND R. LOT

Fig. 7. Slalom maneuver: comparison between experimental measurements and numerical simulations.
 Downloaded by [RMIT University] at 08:47 07 October 2013

A MOTORCYCLE MULTI-BODY MODEL

Fig. 8. Lane change maneuver: comparison of numerical simulations carried out using different multi-body software.
441
 442 V. COSSALTER AND R. LOT
Downloaded by [RMIT University] at 08:47 07 October 2013

Fig. 9. CPU time on a processor AMD 800 MHz.

less than 10% of its peak value, the overall error of vehicle speed is less than 3% of its
peak value, the overall error of roll angle is about the 15% of its peak value and the
overall error of steering angle is about the 13% of its peak value.

6. COMPARISON OF THE PERFORMANCES OF THE MULTI-BODY MODEL

WITH PERFORMANCE OF MULTI-BODY COMMERCIAL SOFTWARE

In this section simulations carried out using FastBike are compared with simulations
carried out using Dads1 and Visual Nastran1.
The features of Visual Nastran and Dads motorcycle models are about the same as
FastBike. It is worth pointing out that these multi-body software do not have a suitable
tire model, so it was necessary to implement the tire model presented in [14] and [13].
In this model the tire is rigid and has a toroidal shape.
The Figure 8 shows simulations of a lane change maneuver carried out using
different codes. The agreement between the data is excellent, both for the steering
torque (Fig. 8a) and roll angle (Fig. 8b).
Even if commercial software for multi-body analysis greatly reduces the time
needed for modeling systems, the time required for simulation is greater. Figure 9
compares the CPU time needed to carry out 1 s of simulation on a AMD K7 800 MHz
processor. The only code that allows real time simulation is Fast bike, which is about
10 times faster than Dads and about 100 times faster than Visual Nastran.

7. CONCLUSIONS

An original mathematical model of a tire and motorcycle was presented.

 A MOTORCYCLE MULTI-BODY MODEL 443

The tire model was developed in order to describe tire behavior at a large camber
angle. The shape of the tire and position of the contact point were described in detail.
The model is based on the physical description of tire forces genesis: the sliding of the
contact patch generates tire forces, which produce a deformation of the carcass of the
tire. By taking into account simultaneously both phenomena, an accurate description
of tire properties is obtained. It was demonstrated that this model is equivalent to
relaxation tire models.
The motorcycle multi-body model has eleven degrees of freedom and includes the
main features of a motorcycle, taking into account the non-linear properties of tires
and suspensions. The very good agreement between the numerical simulations and
experimental tests demonstrates the feasibility and correctness of the model.
The equations of motion were developed based on the natural coordinates
Downloaded by [RMIT University] at 08:47 07 October 2013

approach. This method made it possible to obtain simple equations of motion and
hence high computation ef®ciency was obtained. The comparison of the per-
formances of the FastBike code with the performance of DADS and Visual Nastran
showed that the ®rst is much faster than the others. In particular, real time simulations
can be carried out using FastBike and it can also be used on a motorcycle simulator.
For the same reason, it can be useful for solving optimization problems.

ACKNOWLEDGEMENTS

The authors would like to thank A. Doria for his suggestion regarding the organization
of the paper and D. Bortoluzzi and N. Ruffo for their contribution during the
experimental tests.
This research was partially supported by funds from the Italian Ministry for
Universities and for Scienti®c and Technological Research (MURST 40% funds).

REFERENCES

1. Jalon, J.C. de and Bayo, E.: Kinematic and Dynamic Simulation of Multibody Systems. Springer, 1994.
2. Sush, H. and Radcliffe, C.W.: Kinematics and Mechanism Design. Wiley, New York, 1978, Chapter 3.
3. Sakay, H.: Study on Cornering Properties of Tire and Vehicle. Tire Science and Technology 18 (1990),
pp. 136±139.
4. Pacejka, H.B. and Bakker, E.: The Magic Formula Tyre Model. Vehicle System Dynamics 21 (1991),
pp. 1±18.
5. Pacejka, H.B. and Sharp, R.S.: Shear Force Development by Pneumatic Tyres in Steady State
Conditions: A Review of Modelling Aspects. Vehicle System Dynamics 20 (1991), pp. 121±176.
6. Wang, Y.Q., Gnadler, R. and Schieschke, R.: Vertical Load-De¯ection Behaviour of a Pneumatic Tyre
Subjected To Slip And Camber Angles. Vehicle System Dynamics 25 (1996), pp. 137±146.
7. Berritta, R., Cossalter, V. and Doria, A.: Identi®cation of The Lateral and Cornering Stiffness af
Scooter Tyres Using Impedance Measurements. Proc. 2nd International Conference on Identi®cation
in Engineering Systems, Swansea, UK, 1999, pp. 669±678.
 444 V. COSSALTER AND R. LOT

8. De Vries, E.J.H. and Pacejka, H.B.: Motorcycle Tyre Measurements and Models. Proc. 15th IAVSD
Symposium: The Dynamics of Vehicles on Road and Tracks. Budapest, Hungary, 1997, pp. 280±298.
9. Maurice, J.P. and Pacejka, H.B.: Relaxation Length Behavior of Tyres. Vehicle System Dynamics 27
(1997), pp. 339±342.
10. Zegelaar, P.W.A. and Pacejka, H.B.: Dynamic Tyre Responses to Brake Torque Variations. Vehicle
System Dynamics 27 (1997), pp. 65±79.
11. Guo, K., Liu, Q. and Yangpin, H.: A Non-Steady Tire Model for Vehicle Dynamic Simulation and
Control. Proc. AVEC International Symposium on Advanced Vehicle Control: AVEC'98, Nagoya,
Japan, 1998.
12. Fujioka, T. and Goda, K.: Tire Cornering Properties at Large Camber Angles: Mechanism of the
Moment around the Vertical Axis. JSAE Review 16 (1995), pp. 257±261.
13. Berritta, R., Cossalter, V., Doria, A. and Lot, R.: Implementation of a Motorcycle Tyre Model in a
Multi-Body Code. Tire Technology International, March 1999.
14. Cossalter, V., Doria, A. and Lot, R.: Steady Turning of Two Wheel Vehicles. Vehicle System Dynamics
Downloaded by [RMIT University] at 08:47 07 October 2013

31 (1999), pp. 157±181.

15. Gear, C.W.: Differential-Algebraic Equation Index Transformations. SIAM Journal on Scienti®c and
Statistical Computing, 9 (1988), pp. 39±47.
16. Baumgarte, J.: Stabilization of Constraints and Integrals of Motion in Dynamical Systems. Computer
Methods in Applied Mechanics and Engineering 1 (1972), pp. 1±16.
17. Petzold, L.R.: A Description of DASSL: A Differential/Algebraic System Solver. In: R.S. Stepleman
(ed.): IMACS Transactions on Scienti®c Computation 1, 1982, pp. 430±432.
18. Da Lio, M., Doria, A. and Lot, R.: A Spatial Mechanism for the Measurement of the Inertia Tensor:
Theory and Experimental Results, ASME Journal of Dynamic Systems, Measurement and Control, 121
(March 1999), pp. 111±116.
19. Bortoluzzi, D., Doria, A., Lot, R. and Fabbri, L.: Experimental Investigation And Simulation Of
Motorcycle Turning Performance. 3rd International Motorradkonferenzen, Monaco, 2000.
20. Cossalter, V., Da Lio, M. and Berritta, R.: Studio e Realizzazione di una Macchina per la
Determinazione delle Caratteristiche di Rigidezza e Smorzamento di un Pneumatico Motociclistico. V
Convegno di Tribologia, Varenna, 8±9 Ottobre 1998 (in italian).
21. Bartlett, R.: Introduction to Sports Biomechanics. -E & FN Spon-London, 1997.

APPENDIX
MOTORCYCLE CHARACTERISTICS

Motorcycle Geometric and Mechanical Properties

m1 16.2 kg Rear wheel mass
Ia1 0.66 kgm2 Rear wheel axial inertia
Id1 0.33 kgm2 Rear wheel diametrical inertia
m2 223 kg Rear assembly mass (including rider)
(Gx2 ; Gy2 ; Gz2 ) (0.255, 0.000, ÿ0.0202) m Coordinates of rear assembly CoM with
respect to frame T2
Ix2 ; Iy2 ; Iz2 (24.4, 26.2, 30.3) kgm2 Rear assembly moments of inertia
Cxz2 ; Cyz2 ; Cxy2 (0.0, 0.0, 0.0) kgm2 Rear assembly products of inertia
 A MOTORCYCLE MULTI-BODY MODEL 445
Downloaded by [RMIT University] at 08:47 07 October 2013

Fig. 10. Suspension properties.

l23 0.730 m Distance between rear arm pin and steer

axis
m3 8.75 kg Front assembly mass
(Gx3 ; Gy3 ; Gz3 ) (0.023, 0.000, ÿ0.098) m Coordinates of front assembly CoM with
respect to frame T3
Ix3 ; Iy3 ; Iz3 (0.29, 0.14, 0.21) kgm2 Front assembly moments of inertia
Cxz3 ; Cyz3 ; Cxy3 (0.0, 0.0, 0.0) kgm2 Front assembly products of inertia
10 Nms Damping coef®cient of steering damper
m4 12.0 kg Front wheel mass
Ia4 0.47 kgm2 Front wheel axial inertia
Id4 0.22 kgm2 Front wheel diametric inertia
l1 0.034 m Front wheel offset
ZF;0 0.517 m Center of wheel position (with respect to
frame T3 ) when the suspension is
completely extended
lf 0.535 m Rear arm length
m5 10.0 m Rear arm mass
 446 V. COSSALTER AND R. LOT
Downloaded by [RMIT University] at 08:47 07 October 2013

Fig. 11. Front tire properties.

(Gx5 ; Gy5 ; Gz5 )

(0.275, 0.000, ÿ0.052) m Coordinates of rear arm CoM with
respect to frame T5
Ix5 ; Iy5 ; Iz5 (0.20, 0.80, 0.80) kgm2 Rear arm moments of inertia
y5,0 ÿ.165 rad Rear arm rotation (respect frame T2 )
when the suspension is completely
extended
zr ˆ 0:13526
y5 ÿ 0:138
y25 ÿ 0:036
y35 Relation between spring travel zr and
arm rotation y5
m6 7.00 kg Unsprung front mass
Gx6 ; Gy6 ; Gz6 (ÿ0.029, 0.000, ÿ0.189) m Coordinates of unsprung mass CoM
with respect to frame T3
Ix6 ; Iy6 ; Iz6 (0.22, 0.18, 0 .07) kgm2 Unsprung mass moments of inertia
rp 0.041 m Sprocket radius
rc 0.104 m Wheel sprocket radius
 A MOTORCYCLE MULTI-BODY MODEL 447
Downloaded by [RMIT University] at 08:47 07 October 2013

Fig. 12. Rear tire properties.

(ap ; bp ) (0.080, 0.030) m X±Z coordinates of sprocket center with

respect to frame T2
CD A 0.28 Ns2/m2 Drag force coef®cient (FD ˆ CD A
u2 )

Global Properties
m 276.8 kg Total mass
p 1.421 m Wheel base
e 0.43 rad Castor angle
h 0.636 m Height of the center of mass
b 0.675 m Horizontal position of the center
(with respect to the rear wheel)