http://pik.sagepub.

com/
Dynamics
Engineers, Part K: Journal of Multi-body
Proceedings of the Institution of Mechanical
http://pik.sagepub.com/content/215/3/141
The online version of this article can be found at:

DOI: 10.1243/1464419011544420
2001 215: 141 Proceedings of the Institution of Mechanical Engineers, Part K: Journal of Multi-body Dynamics
Y G Liao and H I Du
system
Cosimulation of multi-body-based vehicle dynamics and an electric power steering control

Published by:
http://www.sagepublications.com
On behalf of:

Institution of Mechanical Engineers

can be found at: Dynamics
Proceedings of the Institution of Mechanical Engineers, Part K: Journal of Multi-body Additional services and information for

http://pik.sagepub.com/cgi/alerts Email Alerts:

http://pik.sagepub.com/subscriptions Subscriptions:
http://www.sagepub.com/journalsReprints.nav Reprints:

http://www.sagepub.com/journalsPermissions.nav Permissions:

http://pik.sagepub.com/content/215/3/141.refs.html Citations:

What is This?

- Sep 1, 2001 Version of Record >>

at Universiti Teknologi Malaysia on May 25, 2014 pik.sagepub.com Downloaded from at Universiti Teknologi Malaysia on May 25, 2014 pik.sagepub.com Downloaded from
Cosimulation of multi-body-based vehicle dynamics
and an electric power steering control system
Y G Liao
1
*
and H I Du
2
1
College of Engineering, Wayne State University, Troy, Michigan, USA
2
General Motors Corporation, Pontiac, Michigan, USA
Abstract: Increasing use of electric power steering (EPS) systems, which affect vehicle dynamic
behaviour, has prompted the need for a more effective method of analysing control systems coupled
with detailed vehicle models. This paper develops an integrated simulation of an EPS control system
with a full-vehicle dynamic model. Using a cosimulation technique, a multi-body-based full-vehicle
dynamic model interacting with the EPS control algorithm is concurrently simulated on a single-bump
road condition. The effects of EPS on the vehicle dynamic behaviour and handling responses resulting
from steer and road input are analysed and compared with proving ground experimental data. The
comparisons show reasonable agreement on tie rod load, rack displacement, steering wheel torque and
tyre centre acceleration. This cosimulation capability may be useful for EPS performance evaluation
and calibration, as well as for vehicle handling performance integration.
Keywords: cosimulation, electric power steering, multi-body dynamics, vehicle dynamics
NOTATION
C
h
damping of the steering wheel
C
m
damping of the motor
C
r
damping of the pinion rack mechanism
C
v
damping of the steering shaft
I
g1
inertia of the drive gear
I
g2
inertia of the driven gear
I
h
inertia of the steering wheel
I
m
inertia of the motor
I
sum
inertia of the entire mechanism
I
v
inertia of the steering column
K
t
spring rate of the torsional bar
M
r
mass of the pinion rack mechanism
n gear ratio
P
s
equation sets of DAE
r c-factor
t time
T
h
input torque
T
m
motor torque
T
sum
sum of the torque applied to the driven
mechanism
U
i
vector of the external inputs
V
m
motor speed
V
s
vehicle speed
x vector of the state variables
y vector of the external outputs
y
h
steering wheel angle input from the driver
y
rh
resultant angle of the steering wheel
y
v
angle of the lower shaft
1 INTRODUCTION
Power-assisted steering is designed to reduce the effort
exerted by the driver on the steering wheel and to reduce
the steering wheel movement for a given swivel pin
angular turn, that is, to make the input (steering wheel
movement) to output (swivel pin movement) more direct
[1]. Most passenger vehicles equipped with power
steering systems are hydraulic power assisted. A vehicle
equipped with a conventional hydraulic power steering
system consumes more fuel than a vehicle with a manual
steering system, because oil pressure must be con-
tinuously boosted.
Electric power steering (EPS) is an advanced steering
system that uses an electric motor to provide steering
assist. It eliminates the need for a hydraulic power
steering pump, hoses, hydraulic uid, drive belt and
pulley on the engine, and therefore the total system is
lighter than a comparable hydraulic system through the
K00401 IMechE 2001 Proc Instn Mech Engrs Vol 215 Part K
141
The MS was received on 23 March 2001 and was accepted after
revision for publication on 19 June 2001.
*Corresponding author: College of Engineering, Wayne State
University, 5330 Clearview Drive, Troy, MI 48098-2428, USA.
at Universiti Teknologi Malaysia on May 25, 2014 pik.sagepub.com Downloaded from
use of compact system units. Also, since EPS is an on-
demand system that operates only when the steering
wheel is turned, the fuel efciency of a vehicle equipped
with such a system is up to 3 per cent [2] better than that
of vehicles equipped with an equivalent-output hydraulic
system. As a result, electric power steering is more
energy efcient and environmentally compatible. This
explains the great increase in EPS-equipped vehicles seen
recently.
With increasing adoption of EPS systems, several
analytical models for EPS systems have been developed.
However, these efforts focused on control algorithm
development and used only simplied representations to
approximate the mechanical characteristics of a vehicle
in developing a control system. Okamoto et al. [3]
investigated the effects of inertia and friction of the motor
used in EPS on the feeling of the driver. The feeling of
the driver was dened as the returnability (addressed
later) and vibration shaking of the steering wheel. Smith
et al. [4] proposed that power steering shudder can be
analytically described as a control system with a closed-
loop limit cycle caused by the interaction of the chassis
and steering system. A metric for determining the
shudder propensity of a vehicle was created. Patwardhan
et al. [5] studied the fundamental issues of automatic
steering control for highway driving in a general
framework. The lateral control response, which relates
to the automatic steering of a vehicle to follow a
reference along the lane centre, was simulated on a
linearized vehicle model. Kurishige et al. [6] developed a
control algorithm to reduce steering torque for an EPS-
equipped vehicle that is stationary or moving at very low
speed. They showed that an auxiliary damping controller
could prevent the undesirable steering vibration that
occurs at a frequency of around 30Hz. Later, Kurishige
et al. [7] developed an EPS system that improved steering
wheel returnability. The steering wheel returnability is
dened as: the responsive ability of the steering wheel to
return to its centre position when the driver releases
(hands-off) the steering wheel immediately after a slight
steering adjustment during high-speed driving or just
after normal steering during relatively low-speed driving.
Based on the prediction of alignment torque generated by
tyres and road surfaces, they developed a new control
algorithm that compensated the motor current as the
alignment torque increased under the limiting condition
of small lateral vehicle acceleration. Masrur et al. [8]
proposed a rotational sensorless scalar control method for
three-phase induction motors and demonstrated the
technique for an automotive electric power-assist steer-
ing system. Their prediction of steering torque correlated
well with the measurement in terms of magnitude.
Multi-body-based vehicle models have been demon-
strated to be a superior technique in realistic prediction of
vehicle dynamic behaviour [911]. Nevertheless, these
efforts focused only on the vehicle chassis mechanisms
(such as steering and suspension), without considering
the control algorithm implemented in the chassis
variables (such as steering motor torque). Most previous
research does not account for the interaction of all chassis
variables involved in monitoring the vehicle and
generating control signals. While Olson and Milacic
[12] utilized computer code to simulate a quarter-vehicle
model interacting with several control systems (steering,
brake and traction), their lower-order vehicle model
reduced the accuracy of the prediction of the vehicle
dynamic responses.
Typically, vehicle models are simplied in the control
arena, and control systems are simplied in the vehicle
dynamics arena. As debugging and rening a control
algorithm are becoming more difcult with the increasing
complexity of automotive mechanisms and other critical
vehicle chassis structures, a multi-body-based vehicle
model fully coupled with the control algorithm is
essential. However, it seems that the computational
dynamic responses of full-vehicle chassis systems
equipped with EPS have not been investigated.
This paper develops an integrated simulation of an
EPS control system with a full-vehicle dynamic model.
Using a cosimulation technique, a multi-body-based full-
vehicle dynamic model interacting with the EPS control
algorithm is concurrently simulated on a single-bump
road condition. The effects of EPS on the vehicle ride and
handling responses resulting from steer and road input are
analysed and compared with proving ground experi-
mental data. The comparisons show reasonable agree-
ment on tie rod load, rack displacement, steering wheel
torque and tyre centre acceleration. This integrated
method allows expensive and time consuming prototype
testing to be focused on high-condence designs.
2 MODEL DEVELOPMENT
This section gives an overview of the general EPS
mechanism and control system. A multi-body-based full-
vehicle model consisting of the pinion rack driving
mechanism, suspensions and tyres is also described.
Then, the interface between the EPS control module and
the full-vehicle model is established in order to investi-
gate the dynamic characteristics of an EPS-equipped
vehicle.
2.1 EPS overview
EPS incorporates a steering gear, assist mechanism,
brushless motor and electronic controller to provide
responsive steering assist. Two primary inputs, the driver
torque on the steering wheel and the vehicle speed signal,
along with other system variables, are continuously fed
into an electronic control module which determines the
direction and amount of steering assist. The ow chart of
the EPS control system is shown in Fig. 1, and the
Proc Instn Mech Engrs Vol 215 Part K K00401 IMechE 2001
142 Y G LIAO AND H I DU
at Universiti Teknologi Malaysia on May 25, 2014 pik.sagepub.com Downloaded from
concept of EPS operation is further explained as follows.
As the driver turns the steering wheel, a certain angle is
exerted on the steering wheel, representing the intention
of the driver to turn the vehicle. Based on the road
conditions and the desire of the driver to turn the vehicle,
a torque sensor located between the steering wheel and
the motor detects the steering torque if there is a
discrepancy between the intended angle at the steering
wheel (i.e. the intended displacement of the rack) and the
actual angle at the front tyre. The motor current is
determined on the basis of this torque sensor signal.
Assistant torque, an input to the steering shaft, is obtained
by multiplying the motor torque by the gear ratio. This
additional torque, together with the driver steering wheel
torque, rotates the steering shaft against the reaction
torque of the tyres and the frictional loss of the steering
mechanism. This assistant torque eases the stress on the
steering column and the effort exerted by the driver.
A typical EPS mechanism is depicted in Fig. 2. Figure
3 shows the EPS block diagram that transforms the inputs
(angle of steering wheel, vehicle speed and rack force) to
outputs (rack displacement and velocity). The sum of the
torques, T
sum
, applied to the pinion rack driving mech-
anism is dened in equation (1), and the angular
acceleration of the lower shaft is dened in equation
(2). Equation (3) species the torque applied on the
steering wheel, T
h
. The c-factor is the ratio of rack
displacement to one revolution of pinion rotation, as
dened in equation (4). The block diagram of the EPS
electric control unit (ECU) coupled with the motor is
illustrated in Fig. 4a. Figures 4b to d illustrate the sub-
system block diagrams of base current, inertia and
friction compensation current respectively [13]:
T
sum
nT
m
K
t
y
h
y
v
C
v
_
y
v
C
r
r
2
_
y
v
C
m
V
m
n
1

y
v
T
sum
=I
sum
2
I
sum
I
h
I
g2
I
m
I
g1
n
2
M
r
r
2
T
h
K
t
y
h
y
v
C
h
_
y
h
I
h

y
h
3
r
rack displacement
1 revolution of pinion rotation
4
where I
sum
is the sum of inertia of the entire driving
mechanism, consisting of steering column inertia, I
v
,
Fig. 1 Flow chart of the EPS control system
Fig. 2 Layout of the EPS mechanism
K00401 IMechE 2001 Proc Instn Mech Engrs Vol 215 Part K
COSIMULATION OF MULTI-BODY-BASED VEHICLE DYNAMICS 143
at Universiti Teknologi Malaysia on May 25, 2014 pik.sagepub.com Downloaded from
drive gear inertia, I
g1
, driven gear inertia, I
g2
, motor
inertia, I
m
, and rack shaft inertia, M
r
r
2
.
2.2 Multi-body-based vehicle dynamic model
A full-vehicle model created from ADAMS [14] is
illustrated in Fig. 5. This multi-body dynamic model
comprises non-linear force-generating elements for
elastic front and rear suspension, the steering system,
including the rack and its joints with tie rods, and a tyre
model with combined longitudinal and lateral force
calculation [15]. It was originally built for vehicle
dynamic simulations. The torques (external inputs)
applied on the rear wheels move the vehicle model.
Although models of the driver, steering wheel and
steering column are not constructed in the vehicle model,
the driven path of the vehicle is controlled by the
displacement of the rack which is an output of the EPS
control system (referring to Fig. 3). The integration of the
driving mechanism with the control algorithm is
discussed in the next section. Overall, this full-vehicle
model processes 347 kinematic degrees of freedom. A
single-bump road prole is created to study the chassis
dynamic responses resulting from road surface impact
force. To simulate the higher unbalanced impact force on
the chassis, only two tyres (front and rear) on the driver
side hit the bump. The bump is 50.8mm in height and
457 mm in width, corresponding to road test conditions at
the proving ground, as displayed in Fig. 5.
3 INTEGRATED ANALYSIS PROCESS
The control algorithm in an EPS-equipped vehicle
denes the path that the vehicle is required to follow. A
code that models and analyses the EPS control system
and its effect on the dynamic characteristics of the
vehicle is required. Multi-body system codes (such as
ADAMS) are suitable for studying the multi-body
dynamic characteristics of the vehicle chassis.
Although a basic controller such as a proportional,
integral, derivative (PID) controller may be imple-
mented with some ease in ADAMS, sophisticated
control (such as the EPS system) requires Matlab
(with Simulink) [16] or some other equivalent software.
On the other hand, the reason for not constructing a full-
vehicle model, which may consist in general of several
hundred kinematic degrees of freedom, in Matlab is that
it would be too onerous a task for the user to formulate
the model.
Owing to the complexity of the EPS control algorithm
and the large number of degrees of freedom on a
full-vehicle model, it is difcult to model and analyse an
EPS-equipped vehicle using a single code. In this study,
ADAMS and Matlab are utilized together to model a full
vehicle equipped with an EPS control system. ADAMS
performs the integration of the dynamic vehicle model
and Matlab executes the control integration of the EPS
system. These two software packages are executed con-
currently in an approach called cosimulation which will
be addressed next.
Fig. 3 EPS control block diagram
Proc Instn Mech Engrs Vol 215 Part K K00401 IMechE 2001
144 Y G LIAO AND H I DU
at Universiti Teknologi Malaysia on May 25, 2014 pik.sagepub.com Downloaded from
Fig. 4 Block diagram of the EPS electric control unit (ECU) with the motor, (b) the base current within the
ECU, (c) the inertia compensation current within the ECU and (d) the friction computation current
within the ECU
K00401 IMechE 2001 Proc Instn Mech Engrs Vol 215 Part K
COSIMULATION OF MULTI-BODY-BASED VEHICLE DYNAMICS 145
at Universiti Teknologi Malaysia on May 25, 2014 pik.sagepub.com Downloaded from
3.1 Cosimulation method overview
In cosimulation, the input and output parameters of both
software packages must be dened. The coupling is
controlled by a subroutine in Matlab. This subroutine
establishes the channels for data exchange and starts
ADAMS at every time step. During the simulation, both
packages take simultaneous time steps. Each package
independently computes the solution for their respective
models. At the beginning and end of each step, the two
packages exchange input/output signal data.
Let P
s
denote the equations of the overall system
modelled in ADAMS; P
s
is represented by the following
set of equations [17]:
_ x f x, U
i
, t
y hx, U
i
, t 5
where x is the vector of all of the state variables, and y is
the vector of all of the external outputs that are used to
formulate the control law. This includes velocities and
positions for each of the rigid bodies in a full-vehicle
model, as well as constraint Lagrange multiplier
equations. Generally, f () is a set of complex non-
linear differential equations (equations of motion) and
h() is a set of algebraic equations (constraint
equations). Let U
i
be the vector of all of the external
inputs which is assumed to be constant on the interval
[i*t, (i 1) *t]. Stabilized index 2 is used to formulate
and integrate equation (5) and to provide the outputs y at
the end of interval, and then the control application
computes U
i 1
for the next interval. Stabilized index 2 is
a numerical method solving differential algebraic
equations (DAEs) [18]. The index is a measure of the
difculty in solving DAE using ordinary differential
equation (ODE) solvers. Index 2 means that the number
of times needed to differentiate the system of equations to
produce a set of ODEs is 2.
3.2 Interface variables
The translational force applied on the rack causes a rack
displacement that affects the turn of the vehicle. There-
fore, the rack force is denoted as the external input,
variable U
i
in equation (5) of the vehicle model. The
resultant displacement and velocity of the rack are the
external outputs, variable y in equation (5), to the control
laws. Figure 6 illustrates the block diagram of the
integrated model and the interface between the full-
vehicle model and the EPS control module. The inputs to
the EPS system are steering wheel torque, vehicle speed
and rack force. The outputs of the EPS system are the
resultant angular position and velocity of the steering
wheel, as well as the displacement and velocity of the
rack. A simplied driver model is also constructed in the
EPS system. This driver model consists of the stiffness
(20N m/rad) and the damping (2N m s/rad) of the human
bodys upper arms, lower arms and hands [19]. An input
le (sw4_in.mat) to the driver model contains historical
experimental data of steering wheel angular position.
These experimental data were measured from the steering
wheel held by the drivers hands as the vehicle was tested
Fig. 5 Full-vehicle model with a single-bump road prole
Proc Instn Mech Engrs Vol 215 Part K K00401 IMechE 2001
146 Y G LIAO AND H I DU
at Universiti Teknologi Malaysia on May 25, 2014 pik.sagepub.com Downloaded from
on the track. The driver model produces a constant
angular position of the steering wheel that turns the
vehicle in a desired circular path corresponding to the
road curve on the proving ground. These computations
are performed by Matlab as illustrated in Fig. 6. The
shaded block denoted adams_sub, shown in Fig. 6, is a
subsystem consisting of the full-vehicle model created
from ADAMS as shown in Fig. 5.
Ideally, the rack displacement (computed from
Matlab) should be a direct input to the vehicle model
(created by ADAMS) and make the vehicle turn. Owing
to a modelling limitation in ADAMS, the time-dependent
relative distance between two bodies could not be
assigned as a variable in the cosimulation. To overcome
this modelling difculty, the rack force resulting from
displacement of the rack is used as an input to the vehicle
model. As illustrated in Fig. 6, a spring load with a
coefcient of 1500 000 N/m represents the stiffness of the
pinion rack driving mechanism and converts rack
displacement (the output from the control module) to
rack force (the input to the vehicle model). Constraining
the entire driven mechanism (steering wheel, steering
column, EPS, rack and tie rod) in a testing xture, this
coefcient was measured by a strain gauge installed on
the rack. The rack damping coefcient (4650N s/m) is
added for the purpose of convergence. A low-frequency
lter with a value of 1/(0.01s 1) is utilized to screen out
high-frequency noise which might cause numerical
divergence. This lter does not reduce the accuracy of
the prediction of high-frequency effects, which is proved
by the comparison of simulation results.
4 SIMULATION RESULTS AND COMPARISON
WITH EXPERIMENTS
The simulation result and its correlation with experi-
mental data are discussed in this section. The simulation
vehicle model, or numerical model, is a production-type
vehicle, while the prototype vehicle was fabricated by
mounting a competitors vehicle body frame on our new
design of chassis (suspension, driving mechanism and
EPS). Therefore, the test vehicle does not truly
correspond to the simulation model. A prototype vehicle
was tested several times on the proving ground. This
vehicle was driven at a speed of 32km/h on a circular test
track with a diameter of 65.8 m. As vehicle lateral
acceleration reached 0.25g, the vehicle was driven over a
single bump with tyres on the driver side (front and rear).
The location and dimension of this single bump are
illustrated in Fig. 5. For vehicle performance evaluation,
only the left-side tyres were driven over this single bump
which generated a larger unbalanced load on the vehicle
chassis from road surface impact. Accelerometers
mounted on several locations inside the vehicle acquired
the dynamic characteristic of the vehicle chassis under
these driving and road conditions.
Fig. 6 Block diagram of the integrated model
K00401 IMechE 2001 Proc Instn Mech Engrs Vol 215 Part K
COSIMULATION OF MULTI-BODY-BASED VEHICLE DYNAMICS 147
at Universiti Teknologi Malaysia on May 25, 2014 pik.sagepub.com Downloaded from
A full-vehicle dynamic model integrated with the EPS
control algorithm is simulated under the same driving and
road conditions as the test conducted on the proving
ground. The vehicle hits the bump at 1.85s on the left
front tyre (referring to Fig. 7) and then on the left rear
tyre. The entire simulation time is 3s and the sampling
rate is 0.005s, which is about 5 times as fast as the
required controller bandwidth. The vehicle model and
EPS control algorithm exchange input/output signal data
every 0.005s, so that 600 time steps are executed.
The comparisons between simulation results and test
data are focused on the time period when the vehicle
impacts on the single bump. Figure 7 shows the predicted
and test results of left front tie rod load in the time
domain. At the rst valley (1.9s), the model under-
predicts 10.3 per cent of the test (2327.4 as against
2595.3 N). However, at the rst peak (1.92s), the model
overpredicts 38.7 per cent of the test (3337.7 as against
2405.5 N). The left front tyre centre accelerations in the
directions of fore/aft and vertical are shown in Figs 8a
and b respectively. At the rst valley illustrated in Fig. 8a,
the model overpredicts 61.2 per cent of the test (10.8 as
against 6.7g). The model also overpredicts 25.2 per
cent of the test (13.9 as against 11.1g) at the rst peak. As
shown in Fig. 8b, the model consistently underpredicts
29.5 per cent at the rst peak of 1.83s (5.5 as against
7.8g), 21.7 per cent at the rst valley of 1.935 s (14.4 as
against 18.4 g) and 25.8 per cent at the second peak of
1.98 s (11.2 as against 15.1g). Figure 9 illustrates the
resultant torque on the steering wheel. At steady state
(starting from 0.7s), the model overestimates 0.96 N m
(more than the test) of the steering wheel resultant torque.
The vehicle model with the EPS function disabled is
simulated to investigate the inuence of EPS on vehicle
performance. When the EPS is turned off, no assist torque
is generated and therefore larger torque and acceleration
(than with EPS turned on) occur on the steering wheel.
The accelerations on the centre of the left front tyre are
evaluated for comparison. As shown in Fig. 10a, after
the vehicle hits the bump, the tyre centre acceleration in
the direction of fore/aft has a larger magnitude at every
peak and valley in the case of disabled EPS. The vehicle
model with EPS off also has more uctuations and needs
a longer settling time. This phenomenon is also observed
in the load on the left front tie rod, as illustrated in
Fig. 10b.
The discrepancies between the simulation result and
the experiment are primarily due to the test vehicle
(prototype vehicle) not truly corresponding to the
simulation model (production-type vehicle). To reduce
the time for new vehicle development, it is quite common
to fabricate the prototype vehicle by mounting a
competitors vehicle body frame on the proposed design
of chassis. Since the objective here is to evaluate the
capability of the cosimulation technique, the correlation
was done using the available prototype vehicle. Overall,
the simulation results correlate quite well with the
experimental data in terms of the occurrences of spikes
and their magnitudes.
Although the predictions were not fully validated with
the corresponding test vehicle, this study has investigated
the capability of the cosimulation in virtual vehicle
development. According to the predicted and test data,
the EPS control algorithm has a signicant impact on the
vehicle dynamic behaviour. To improve the dynamic
responses of an EPS-equipped vehicle, further investiga-
tions will focus on the EPS control algorithm, such as
motor inertia, motor speed and gear ratio. These
parameters affect the motor current which generates the
assistant torque in the steering.
Fig. 7 Load on the left front tie rod
Proc Instn Mech Engrs Vol 215 Part K K00401 IMechE 2001
148 Y G LIAO AND H I DU
at Universiti Teknologi Malaysia on May 25, 2014 pik.sagepub.com Downloaded from
Fig. 8 (a) Left front tyre centre fore/aft acceleration and (b) left front tyre centre vertical acceleration
Fig. 9 Resultant torque on the steering wheel
K00401 IMechE 2001 Proc Instn Mech Engrs Vol 215 Part K
COSIMULATION OF MULTI-BODY-BASED VEHICLE DYNAMICS 149
at Universiti Teknologi Malaysia on May 25, 2014 pik.sagepub.com Downloaded from
5 CONCLUSIONS
Owing to the complexity of the EPS control algorithm
and the large number of degrees of freedom on a full-
vehicle dynamic model, it is difcult to model and
analyse an EPS-equipped vehicle using a single code.
The cosimulation provides a more complete representa-
tion of the vehicle and the control system by selectively
using the strengths of each application. This study applies
a cosimulation technique to investigate the inuence of
EPS on the vehicle ride and handling responses. The
cosimulation executes two computer codes concurrently
to resolve the multi-body dynamics (full-vehicle model)
and control algorithm (EPS system).
A multi-body-based full-vehicle dynamic model inte-
grated with an EPS control algorithm has been developed
that predicts the dynamic behaviour of an EPS-equipped
vehicle resulting from steer and road input. The inte-
grated vehicle model is simulated on a circular path with
a single-bump road condition and the dynamic responses
of vehicle chassis and steering system resulting from road
surface impact are evaluated and compared with proving
ground experimental data. The comparisons show
reasonable agreement on tie rod load, rack displacement,
steering wheel torque and tyre centre acceleration.
To investigate the inuence of EPS on vehicle
performance, the vehicle model with the EPS function
disabled is also simulated. When the vehicle (with
disabled EPS) impacts on a bump, tyre centre accelera-
tions have larger magnitude at every peak and valley. It is
also more oscillatory and has a longer settling time.
This study also evaluates the capability of the
cosimulation technique in integrating a multi-body-based
full-vehicle dynamic model with a complex control
Fig. 10 Comparison of EPS on and off (a) on the left front tyre centre fore/aft acceleration and (b) on the left
front tie rod load
Proc Instn Mech Engrs Vol 215 Part K K00401 IMechE 2001
150 Y G LIAO AND H I DU
at Universiti Teknologi Malaysia on May 25, 2014 pik.sagepub.com Downloaded from
algorithm. The integrated method allows engineers to
determine whether the overall design of the vehicle and
EPS meets design targets without going through
extensive prototype testing which is expensive and time
consuming. This simulation capability can be used for
EPS performance evaluation and calibration.
REFERENCES
1 Heisler, H. Vehicle and Engine Technology, 1999, pp. 226
234 (Society of Automotive Engineers, Warrendale,
Pennsylvania).
2 Nakayama, T. and Suda, E. The present and future of
electric power steering. Int. J. Veh. Des., 1994, 15(35),
243254.
3 Okamoto, K., Chikumr, I., Saio, N. and Miyazaki, H.
Improvement of drivers feel of electric power steering.
SAE paper 890079, 1989.
4 Smith, J. M., Ferries, G. R. and Arbanas, R. L. An
analytical control systems approach to steering shudder.
SAE paper 951254, 1995.
5 Patwardhan, S., Tan, H. and Guldner, J. A general
framework for automatic steering control: system analysis.
In Proceedings of American Control Conference, June
1997, pp. 15981602.
6 Kurishige, M., Kifuku, T., Inoue, N., Zeniya, S. and
Otagaki, S. A control strategy to reduce steering torque for
stationary vehicles equipped with eps. SAE paper 1999-01-
0403, 1999.
7 Kurishige, M., Wada, S., Kifuku, T., Inoue, N.,
Nishiyama, R. and Otagaki, S. A new eps control strategy
to improve steering wheel returnability. SAE paper 2000-
01-0815, 2000.
8 Masrur, M. A., Hampo, R. J. and Miller, J. M. Rational
sensorless scalar control of three-phase induction motors
and its application to automotive electric power assist
steering. Proc. Instn Mech. Engrs, Part D, Journal of
Automobile Engineering, 2000, 214(D1), 3344.
9 Choi, G. J., Yoo, Y. M., Lee, K. P. and Yoon, Y. S. Real-
time multibody vehicle dynamic analysis method using
suspension composite joints. Int. J. Veh. Des., 2000, 24(2),
259273.
10 Hegazy, S., Rahnejat, H. and Hussain, K. Multi-body
dynamics in full-vehicle handling analysis under transient
manoeuvre. Veh. Syst. Dynamics, 2000, 34(1), 124.
11 Trom, J. D., Lopez, J. L. and Vanderploeg, M. J.
Modeling of a mid-size passenger car using a multibody
dynamics program. In ASME Design Engineering Tech-
nical Conference, 1986, paper 86-Det-39 (American
Society of Mechanical Engineers, New York).
12 Olson, W. W. and Milacic, D. Development of anti-lock
braking traction and control systems of the advanced
technology demonstrator II using DADS simulation code.
Int. J. Veh. Des., 1996, 17(3), 295316.
13 Koyo Technical Manual, 2000 (Koyo Inc.).
14 ADAMS/Solver Reference Manual, Version 10.0, 2000
(Mechanical Dynamics, Inc.).
15 Pacejka, H. Tyre and Its Impact on Vehicle Dynamic
Behavior, 1997 (Pottendam).
16 MATLAB Users Manual, Version 5.3, 1999 (MathWorks,
Inc., Natick, Massachusetts).
17 McNally, P. J. and Gray, S. The adams/controls and
adams/plant products: an example combining the power of
Adams 9 and Matlab 5. In ADAMS Users Conference,
1997, paper.
18 Campbell, S. L. and Zhong, Y. Jacobian reuse in explicit
integrators for higher index DAEs. Appl. Numer. Math.,
1997, 25(4), 391412.
19 Robbins, D. H. Anthropometry of motor vehicle occupants.
Vol. 2. Mid-sized male. Report of US Department of
Transportation, National Highway Trafc Safety Adminis-
tration and Report UMTRI-83-53-2 of The University of
Michigan, Transportation Research Institute, 1983.
K00401 IMechE 2001 Proc Instn Mech Engrs Vol 215 Part K
COSIMULATION OF MULTI-BODY-BASED VEHICLE DYNAMICS 151
at Universiti Teknologi Malaysia on May 25, 2014 pik.sagepub.com Downloaded from