Sie sind auf Seite 1von 15

535

A non-linear dynamic approach to the motion of


four-wheel-steering vehicles under various operation
conditions
Q Han1 and L Dai2*
1
Department of Mechanics, College of Traffic and Communications, South China University of Technology,
Guangzhou, People’s Republic of China
2
Industrial Systems Engineering, University of Regina, Regina, SK, Canada

The manuscript was received on 24 February 2005 and was accepted after revision for publication on 11 January 2008.

DOI: 10.1243/09544070JAUTO43

Abstract: This research aims to analyse systematically the non-linear dynamic behaviour of
four-wheel-steering (4WS) vehicles. A practical non-linear dynamic model of multiple degrees
of freedom for 4WS vehicles is established. The non-linear model presented is shown to be
feasible for the vehicles under normal and the other operation manoeuvres. Compared with
those of documented investigations, this model may be employed to analyse the motions of
the vehicle and motions of the vehicle wheels in turning and braking processes together with
the consideration of the effects of air drag and wind. Numerical simulation for the motion of
the 4WS vehicles under various operation manoeuvres is performed with the model
established. Comparison of the behaviours of 4WS and two-wheel-steering vehicles is also
presented with respect to the inputs on the steering system, the manoeuvrability, the stability,
and the relationship between the steering phase and vehicle speed.

Keywords: vehicle dynamics, manoeuvrability, non-linear motion, four-wheel-steering


vehicle, two-wheel-steering vehicle, numerical simulation, non-linear tyres

1 INTRODUCTION and negative phase steering. Lateral motion stabili-


zation for a 4WS vehicle has been reported [8].
Four-wheel-steering (4WS) vehicles show advan- Stability and non-linear behaviour of 4WS vehicles
tages of smaller turning radius, tight-space man- such as Hopf bifurcation were also found among the
oeuvrability, and reduction in driver fatigue. Re- recent studies [9, 10]. However, a systematic and
cently, companies such as General Motors and thorough investigation on the response of 4WS
Chevrolet have pushed and promoted the use of vehicles subjected to the tyre forces generated by
4WS techniques, and investigations on the behaviour the road surface, the aerodynamics resistance, and
of 4WS vehicles have attracted great attention from the loadings caused by operation conditions of the
the scientists and engineers [1–4]. Linear manoeuvr- real world is still needed.
ing equations were reported for analysing the The present research proposes an approach for
motion of 4WS vehicles [5]. Itoh et al. [6, 7] accurately and effectively analysing the motion of
developed a numerical approach for the steady state 4WS vehicles. A complete 4WS vehicle model with
turning of a 4WS tractor. The numerically deter- multiple degrees of freedom in counting non-linear
mined forces on the tractor tyres were compared effects is presented. The non-linear model estab-
with those obtained in field tests. The stability of a lished is shown to be a feasible 4WS vehicle model
4WS vehicle is crucial for the operations of positive for various manoeuvres (greater lateral accelera-
tions, possibly combined with longitudinal accelera-
*Corresponding author: Industrial Systems Engineering, Uni- tion or braking). The non-linear coupling effects of
versity of Regina, 3737 Wascana Parkway, Regina, SK, S4S 0A2, sprung and unsprung parts are taken into considera-
Canada. email: dailimli@uregina.ca tion. The model counts the joint effects of operations

JAUTO43 F IMechE 2008 Proc. IMechE Vol. 222 Part D: J. Automobile Engineering
536 Q Han and L Dai

and road conditions to the motions, such as the as shown in Fig. 1. For such a vehicle system, the
translation, surge, sway, yaw, pitch, and roll of the total mass of the vehicle is considered to consist of
vehicle body under the conditions of various opera- Ms and Mu, the masses of vehicle body and the
tion manoeuvres. The four rotational degrees of unsprung part of the vehicle respectively. Also, O9 is
freedom of the vehicle’s four wheels are taken into designated as the mass centre of the whole vehicle
account. To obtain a systematic and accurate Os as the mass centre of the vehicle body, and Ou as
analysis of the non-linear dynamics, the high-value the centre of the unsprung mass, as shown in Fig. 2.
lateral and longitudinal accelerations, the accelerat- Assume that the vehicle is symmetric about the
ing and braking processes, the aerodynamic drag of longitudinal midplane and all the three mass centres
the vehicle, and the shift of vertical loads due to the are located in the symmetric plane.
pitch and roll of the vehicle body are investigated in With the vehicle thus defined, based on Fig. 3, the
the present research. The non-linear terms are velocity relationships can be obtained from Fig. 1 as
naturally introduced into the dynamic system with
considerations of the lateral wheel forces and U~X_ ~u cosy{v siny
geometric relationships relating to the kinematic V ~Y_ ~u sinyzv cosy
motion and slip angles. Based on the analytical ð1Þ
model established, numerical simulations for the
where u and v are the longitudinal and lateral
dynamic response of the 4WS vehicle are carried out.
velocities of the vehicle in the x and y directions
The characteristics of 4WS vehicles in comparison
respectively, U and V are the vehicle velocities in the
with two-wheel-steering (2WS) vehicles are demon-
X and Y directions respectively, and y is the yaw
strated and analysed. angle. Differentiating both sides of equation (1) gives

ay ~Y€ cosy{X
€ siny~v_ zru
2 MODELLING OF 4WS VEHICLES
ax ~Y€ sinyzX
€ cosy~u_ {rv
2.1 Model of the vehicle ð2Þ

The 4WS vehicle studied in this research is assumed In the above equations, ax and ay are the
to be a symmetric body of mass M with four wheels, projections of the absolute acceleration of point Ou
on the moving coordinate system x, y, z, and r is the
yaw angular velocity. The unsprung part is consid-
ered as a rigid body with translational motion and
the sprung part is considered as a mass point, which
is articulated with the unsprung part by a crank. The
crank can rotate with respect to the x axis and y axis
to simulate the roll and pitch of the sprung part of
the vehicle.
Figure 1 also shows the relationship between the
roll and pitch of the sprung part of the vehicle. At an
arbitrary moment, the crank is at position AC, which
can be obtained by rotating the crank twice. The
crank is first rotated an angle w with respect to the x
axis, and it is then rotated an angle h with respect to
the y0 axis. Accordingly, the angular velocities of the
three rotations can be expressed as

r~y_

p~w_
q~h_
ð3Þ
Assuming that i, j, and k are the three unit vectors
along the coordinate axes x, y, and z, the position of
Fig. 1 A simplified model of the vehicle body the centre of mass of the sprung part of the vehicle

Proc. IMechE Vol. 222 Part D: J. Automobile Engineering JAUTO43 F IMechE 2008
Four-wheel-steering vehicles under various operation conditions 537

Fig. 2 A model of the 4WS vehicle

sprung part is composed of three parts, namely the


acceleration of Ou, the relative acceleration arc, and
the Coriolis acceleration akc, i.e. ac 5 aec + arc + akc.
From equation (2)

aec ~ðu_ {rvÞizðv_ zruÞj ð6Þ

or, in another form

d2 rOC
aec ~
dt 2
  2
h
~{ {sinðhzwÞ h_ zw_
2
   2
zcosðhzwÞ €hzw€ {sinðh{wÞ h_ {w_
Fig. 3 Two frames of coordinates  i
zcosðh{wÞ €h{w€ i
can be expressed by a vector rAC. Now   2 
zh cosw w{sinw w_
€ j
rOC ~rOA zrAC ~cizrAC ð4Þ 
h  2
z cosðhzwÞ h_ zw_
where 2
   2
rAC ~hð{sinh cosw izsinw j{cosh cosw kÞ ð5Þ zsinðhzwÞ €hzw€ zcosðh{wÞ h_ {w_
 i
zsinðh{wÞ €h{w€ k ð7Þ
The acceleration ac of the centre of mass of the

JAUTO43 F IMechE 2008 Proc. IMechE Vol. 222 Part D: J. Automobile Engineering
538 Q Han and L Dai

It should be noted that The inertia force of the sprung mass is


 
drOC hh    i F s ~{Ms Ai izAj jzAk k ð15Þ
~{ cosðhzwÞ h_ zw_ zcosðh{wÞ h_ {w_ i
dt 2
hh   The moment about the z axis can be expressed as
zh cosw w_ jz sinðhzwÞ h_ zw_
2  s _

 i M z ~{ Iz r_ {Ixz p k ð16Þ
zsinðh{wÞ h_ {w_ k ð8Þ
where Iz is the moment of inertia of the whole
s
Therefore, the Coriolis acceleration akc can be vehicle with respect to the z axis, and Ixz is the
expressed in the form product of the inertia of Ms, and angular velocities r
and p are as defined in equation (3).
drOC The moment about the x axis is
akc ~2r|
dt  
drOC M x ~{ Ixs p_ {Ixz
s _
r i ð17Þ
~2rk|
dt
h   where Ixs is the moment of inertia of the sprung mass
~{hr cosðhzwÞ h_ zw_ with respect to the x axis.
 i The moment about the y axis is
zcosðh{wÞ h_ {w_ j{2hr cosw w_ i ð9Þ
M y ~{Iys q_ j ð18Þ
Finally, the acceleration ac, of the centre of mass of
the sprung part can be written as where Iys is the moment of inertia of the sprung mass
with respect to the y axis.
ac ~aec zarc zakc ~Ai izAj jzAk k ð10Þ
2.2 Model of the tyre force
where
The additive turning of the rear wheels can be the
 2  turning in either the negative phase or the positive
h
Ai ~{ {sinðhzwÞ h_ zw_ phase. In the negative phase, the rear wheels turn in
2 the opposite direction to the front wheels whereas,
   2
€ {sinðh{wÞ h_ {w_ in the positive phase, the rear wheels turn in the
zcosðhzwÞ €
hz w
same direction as the front wheels. To express this
 i
€ {2hr cosw w_ zu_ {rv ð11Þ relationship quantitatively, the turning angle ratio Kl
zcosðh{wÞ €
h{ w
is introduced. It is defined as the ratio between the
turning angles of the rear wheel and front wheel and
  2  is given by
Aj ~h cos w w{sin w w_

h   dr
Kl ~ ð19Þ
{hr cosðhzwÞ h_ zw_ df
 i where dr is the turning angle of the rear wheel and df
zcosðh{wÞ h_ {w_ zv_ zru ð12Þ
is the turning angle of the front wheel.
 In Fig. 4, the velocity of the front right wheel is
h  2  
Ak ~ cosðhzwÞ h_ zw_ zsinðhzwÞ € h zw€ expressed as
2
 2   V fR ~Vzr|rOG
zcosðh{wÞ h_ {w_ zsinðh{wÞ €h{ w€ ð13Þ  
d
~uizvjzrk| ajz
2j
 
The inertia force of the unsprung mass can be d
~ u{ izðvzar Þj ð20Þ
expressed as 2r

As such, tan(df + afR) 5 (v + ar)/(u 2 d/2r). Similarly,


F u ~{Mu ½ðu_ {rvÞizðv_ zruÞj  ð14Þ
the lateral slip angles of all four wheels of the vehicle

Proc. IMechE Vol. 222 Part D: J. Automobile Engineering JAUTO43 F IMechE 2008
Four-wheel-steering vehicles under various operation conditions 539

Fig. 4 Top view of the vehicle

can be given by [14], and tyre friction [15] can be easily implemented
  in the vehicle model established in the present
vzar research, if so desired.
afR ~arctan {df
u{rd=2
 
vzar
afL ~arctan {df 3 AERODYNAMIC RESISTANCE
uzrd=2
 
v{br A straight-line driving vehicle is influenced by the
arR ~arctan {dr
u{rd=2 side wind and the wind caused by the travelling
 
v{br velocity of the vehicle. The side wind can be
arL ~arctan {dr ignored under normal weather conditions and the
uzrd=2
ð21Þ travelling wind can be considered as obeying the
formulae
Using the non-linear tyre force model [11]
  1
Ffi ~{ C1f afi {C3f a3fi Fx ~ Cx ruAx
2
 
Fri ~{ C1r ari {C3r a3ri 1
ð22Þ Fy ~ Cy rvAy
2
ð23Þ
where Ffi and Fri (i 5 R, L, i.e. the subscripts for right
and left respectively) are the tyre forces of the front where Cx and Cy are the coefficients of aerodynamic
and rear wheels respectively. This model is empirical resistance in the x and y axes respectively, and Ax
and the parameters of the model are to be and Ay are the projected areas of the vehicle.
determined in the experimental measurements
Equation (23) therefore governs the aerodynamic
corresponding to the tyre and road conditions.
resistance.
Variation in the parameters can be easily implemen-
ted in the numerical calculations in vehicle beha-
viour analysis. With this tyre model, the tyre force
4 MOTION OF THE TYRES AND NORMAL LOADS
directions are perpendicular to the centre-lines of
the tyres. This tyre force model is widely used for
4.1 Starting, accelerating, and normal driving
vehicle behaviour analysis. In this research, the
three-dimensional non-linear response of 4WS ve- In these cases considered, the governing equations
hicles under normal and other operation man- for the driving wheel (Fig. 5) can be expressed as
oeuvres of high accelerations is the main focus.
Therefore, this model is suitable for the research.
However, other tyre force models with considera- Ir v_ rL ~T {RFtrL {eFzrL
tions of the factors such as dynamic response and Ir v_ rR ~T {RFtrR {eFzrR
slip force [12], tyre saturation [13], pavement effects ð24Þ

JAUTO43 F IMechE 2008 Proc. IMechE Vol. 222 Part D: J. Automobile Engineering
540 Q Han and L Dai

wheels respectively. In equation (26), R is the tyre


radius, T represents the driving torque, Tbrkij(i, j 5 f, r)
are the braking torques, Fzij(i, j 5 f, r, L, R) are the
normal loads on the tyres, e is the offset distance of
the normal load, and Ftij(i, j 5 f, r, R, L) are the
tangential loads on the tyres. These tangential loads
are directly determined by the driving torque.
It should be noted that

0 ðT ~0Þ
Ir ~If zjIt , j~ 1 ð27Þ
2 ðT =0Þ

where It is the total moment of inertia of all the


moving parts connected to the driving wheels.
The total braking torque Tbrk is controlled by the
Fig. 5 Free-body diagram of the driving wheel
driver and distributed on the front and rear wheels
according to the rules

TbrkfL ~TbrkfR ~Kbf Tbrk


TbrkrL ~TbrkrR ~ð1{Kbf ÞTbrk
ð28Þ

The driving torque T on the driven wheels is


transmitted from the torque output of the engine
by the transmission system. If the torque output of
the engine is denoted by Me, the transmission ratio
of the transmission system is given by ig, the
transmission ratio of the main reducer is designated
by i0, and the mechanical efficiency of the transmis-
sion system is represented by gT, then

Fig. 6 Free-body diagram of the driven wheel 1


T ~ i g i 0 g T Me ð N m Þ ð29Þ
2
and for the driven wheel (Fig. 6)
where
If v_ fL ~RFtfL {eFzfL pe
Me ~9549 ðN mÞ ð30Þ
If v_ fR ~RFtfR {eFzfR ne
ð25Þ
In the above equations, pe and ne are the power and
the corresponding rotating speed of the engine.
4.2 Braking Their values can be found from the corresponding
In braking, the equations of dynamics are engine characteristics curve. Taking only the max-
imum power pemax and the corresponding crank
If v_ fL ~RFtfL {eFzfL {TbrkfL rotating speed np, the external characteristics curve
of the engine of pe versus ne can be given by the
If v_ fR ~RFtfR {eFzfR {TbrkfR expression
Ir v_ rL ~RFtrL {eFzrL {TbrkrL "  2  3 #
Ir v_ rR ~RFtrR {eFzrR {TbrkrR ne ne ne
ð26Þ pe ~pemax c1 zc2 { kW ð31Þ
np np np

In the above equations, Ii (i 5 f, r) are the moments of In equation (31), the units of pe are kilowatts and
inertia of the front wheel and rear wheel, v̇ij(i, j 5 f, r, the units of ne are revolutions per minute. In the
L, R) are the angular accelerations of the four wheels, equation development, only the pure rolling of the
the subscripts L and R designating the left and right tyres is considered. The velocity of the wheel centre

Proc. IMechE Vol. 222 Part D: J. Automobile Engineering JAUTO43 F IMechE 2008
Four-wheel-steering vehicles under various operation conditions 541

and therefore the velocity of the centre of mass of


the vehicle can be given by the relations Mg a u_ {rv hcg
FzrR ~ z
2 L g L
 
VfR ~RvfR hcg v_ zruzMs =Mhp_ Ms h
{ð1{KRSF Þ { sin w
VfL ~RvfL d g Md

VrR ~RvrR Ms Ai hsg


z ð36Þ
2L
VrL ~RvrL
ð32Þ

Mg a u_ {rv hcg
FzrL ~ z
2 L g L
 
and hcg v_ zruzMs =Mhp_ Ms h
zð1{KRSF Þ { sin w
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi d g Md
 
dr 2 Ms Ai hsg
VfR ~ u{ zðvzar Þ2 { ð37Þ
2 2L
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
  ffi
dr 2 2 In the above equations, L 5 a + b, hcg is the
VfL ~ uz zðvzar Þ
2 distance from the centre of mass of the vehicle to
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
  ffi the ground, hsg is the distance between the centre of
dr 2 mass of the sprung part and the ground, and KRSF is a
VrR ~ u{ zðv{br Þ2
2 stiffness coefficient.
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
  ffi
dr 2
VrL ~ uz zðv{br Þ2
2 5 GOVERNING EQUATIONS
ð33Þ
On the basis of the models and related equations
The pitch and roll of the vehicle body actually developed, the governing equations of the vehicle
redistribute the normal loads among the wheels. are as follows. The motion in the y direction is given
This redistribution of the normal loads has a great by
influence on the vehicle performance. The normal
load on the tyre exerted by the ground can be Mu ðv_ zruÞzMs Aj {ðFfL zFfR Þ cos df
divided into two portions: the static and dynamic
{ðFrL zFrR Þ cos dr
loads. These normal loads on the four tyres can be
described by the equations zðFtfL zFtfR Þ sin df
zðFtrL zFtrR Þ sin dr ~0 ð38Þ

Mg b u_ {rv hcg
FzfR ~ { The rotation about the z axis is given by
2 L g L
  X
s _
hcg v_ zruzMs =Mhp_ Ms h Mzi ~Iz r_ {Ixz p ð39Þ
{KRSF { sin w i
d g Md
Ms Ai hsg The moment caused by the weight of the sprung
z ð34Þ
2L part is

M mg ~hð{sin h cos w izsin w j{cos h cosw kÞ



Mg b u_ {rv hcg |Ms gk~Ms hg ðsin h cosw jzsin w i Þ ð40Þ
FzfL ~ {
2 L g L
  The rotation about the x axis is given by
hcg v_ zruzMs =Mhp_ Ms h
zKRSF { sin w
d g Md  
Ixs p_ {Ixz
s _
r ~{Ms h Ak sin wzAj cos h cosw
Ms Ai hsg
{ ð35Þ zMs hg sinw{Kw w ð41Þ
2L

JAUTO43 F IMechE 2008 Proc. IMechE Vol. 222 Part D: J. Automobile Engineering
542 Q Han and L Dai

The rotation about the y axis is given by newly developed P–T method [16], though other
numerical methods can also be used. Numerical
Iys q_ ~Ms hg sin h cos w values of the parameters expressed in the governing
equations and utilized in the numerical computa-
zMs hð{Ak sin hzAi cos hÞcos w{Kh h ð42Þ
tions are listed in Table 1.
where Kh and Kw are the additional resilience
moments caused by the unit roll angle and the pitch 6.1 Dynamic responses of vehicles under a linear
angle respectively. angular turning displacement input
The turning angular displacement for this case is
6 NUMERICAL ANALYSIS exhibited graphically in Fig. 7. Corresponding to
such a step input on the steering system, the motion
Implementing the governing equation derived of the vehicle is quantified with the governing
above, the motion of a 4WS vehicle can be equations derived previously.
quantified. The responses of the vehicle to three Figure 8 shows the paths of the 2WS and 4WS
types of turning displacement input are numerically vehicles for this case. Figure 9 illustrates the varia-
investigated and compared with that of a 2WS tion of the lateral velocity of the vehicle with respect
vehicle. Any existing numerical methods, such as to time. The angular velocities in the yaw plane are
the Runge–Kutta method for solving differential shown in Fig. 10 for the two types of vehicle.
equations, can be utilized for stimulating the motion Figures 11 and 12 exhibit the variations in roll
of a 4WS vehicle with implementation of the angular velocity and pitch velocity respectively with
equations developed. For good convergence and respect to time. The comparisons of the relations
high accuracy of the numerical calculations, the between yaw angular velocities and the correspond-
numerical computations are carried out with the ing yaw angular displacements are illustrated in
Fig. 13. Figure 14 compares the angular velocities in
roll and pitch planes. Comparison of the angular
Table 1 Parameters used in the calculation velocity and angular acceleration in the pitch plane
Symbol Value Units is given in Fig. 15. The figures are presented in a way
c 0.026 m to help readers to make comparisons conveniently.
h 0.2 m However, readers may need to note the different
Mu 670 kg
Ms 1160 kg scales, especially the scales of the vertical axes, and
Iz 2500 kg m2 the ranges of the variables involved in the figures.
s
Ixz 0 kg m2 The results obtained above are for the motions of
Ixz 750 kg m2
Iys 1600 kg m2 2WS and 4WS vehicles in both positive-phase
a 0.9 m steering and negative-phase steering. From the
d 1.33 m numerical calculations performed, the following
b 1.7 m
C1f 44 400 N/rad conclusions can be made corresponding to Figs 8
C3f 44 400 N/rad3 to 15.
C1r 43 600 N/rad
C3r 43 600 N/rad3 1. Taking R* as the turning radius, it can be
r 1.2258 N s2/m4
Cx 0.32 – observed from Fig. 8 that the magnitudes of
Cy 0.35 –
Ax 2.1 m2
Ay 5.7 m2
hsg 0.556 m
Kw 85 000 N/rad
Kh 76 185 N/rad
If 2.1 kg m2
It 0.136 kg m2
R 0.3 m
e 0.014 m
Kbf 0.55 –
hcg 0.5 m
KRSF 0.444 –
Me 170 Nm
ig 13.7 –
i0 0.85 –
gt 0.97 – Fig. 7. Turning angle versus time

Proc. IMechE Vol. 222 Part D: J. Automobile Engineering JAUTO43 F IMechE 2008
Four-wheel-steering vehicles under various operation conditions 543

Fig. 8 Paths of 4WS and 2WS vehicles

Fig. 9 Velocity v of the vehicle versus time

Fig. 10 Yaw angular velocity versus time

Fig. 11 Roll angular displacement versus time

JAUTO43 F IMechE 2008 Proc. IMechE Vol. 222 Part D: J. Automobile Engineering
544 Q Han and L Dai

Fig. 12 Pitch angular displacement versus time

Fig. 13 Yaw angular velocity versus yaw angular displacement

Fig. 14 Roll angular velocity versus roll angular displacement

Fig. 15 Pitch angular velocity versus pitch angular displacement

Proc. IMechE Vol. 222 Part D: J. Automobile Engineering JAUTO43 F IMechE 2008
Four-wheel-steering vehicles under various operation conditions 545

the turning radii satisfy the inequality of


RKl ~{0:3 vRKl ~0 vRKl ~0:3 . Based on the numerical
calculations made for the present research, the
turning radius of the vehicle is also proportional
to the travelling speed of the vehicle. It may thus
be stated that the 4WS vehicle has the smallest
turning radius in negative phase steering; hence
the steering is relatively easy. This implies that the
4WS vehicles have better manoeuvrability than
2WS vehicles do. However, 4WS vehicles have the
largest turning radius in positive-phase steering,
as shown in Fig. 8. As such, the vehicles present Fig. 16 Turning angle versus time
the characteristics of insufficient turning in the
positive-phase steering case. The 4WS system is large in the case of positive-phase steering.
therefore suitable for enhancing the manoeuvr- Additionally, in comparison with 2WS vehicles,
ability at low driving speeds and for increasing the the increase in the yaw angular velocity of 4WS
stability at intermediate and high speeds. In vehicles is slower in positive-phase steering, and
knowing this, negative-phase steering should be the vehicle response time is longer with a
employed for the cases in which low speed and decrease in the yaw velocity. These are the
large turning angle are required, whereas posi- preferred turning characteristics for vehicles
tive-phase steering should be used for improving operating at high speeds.
the driving stability in the cases of high speeds. 4. For 4WS vehicles in positive- and negative-phase
2. The transverse velocities of the vehicles are not steering, the pitch and roll motions of the sprung
constant, as they should not be corresponding to portion of the vehicle are very smooth and the
the input given. This can be seen from Fig. 9. In variation in the displacement of the 4WS vehicle
fact, all the motions with the given input are not are small, especially for the negative-phase steer-
constants. The displacements of roll, yaw, and ing as shown in Figs 14 and 15. This is beneficial
pitch varied for a while and then stabilized. for improving ride comfort.
3. When a 4WS vehicle is in negative-phase steering,
the lateral forces on the front and rear tyres
generate the moments in the same rotating
6.2 Dynamic responses of vehicles under a saw or
direction with respect to the vehicle’s centre of
a half-saw angular turning displacement
mass. Therefore, the vehicle has relatively large
input
yaw angular velocities for 2WS and 4WS negative-
phase steering as time increases whereas the yaw Sawteeth angular turning is common in vehicle
angular velocity is smaller in positive-phase operations, such as in turning and lane changes.
steering, as can be seen from Figs 10 and 13. It The dynamic responses of vehicles with a saw
should be noted, however, that the lateral force angular turning displacement input are shown in
acting at the vehicle’s centre of mass is relatively Figs 16 to 19.

Fig. 17 Moving paths of the vehicles

JAUTO43 F IMechE 2008 Proc. IMechE Vol. 222 Part D: J. Automobile Engineering
546 Q Han and L Dai

Fig. 18 Yaw angular velocity versus time

Fig. 19 Lateral acceleration versus time

The dynamic responses of vehicles to a given half-


saw angular displacement input are shown in Figs 20
to 23.
Again, in reviewing the figures, readers may need
to pay attention to the different scales of the vertical
axes in the figures.
With the analysis of the results plotted in Figs 16
to 23 for the two types of saw angular turning
displacement input, together with the stability
analyses performed by the present authors for the
4WS and 2WS vehicles [10], the following conclu-
sions can be drawn.

Fig. 20 Turning angle versus time 1. When a straight-line driving vehicle passes frontal
obstacles or turns, it can be observed from the

Fig. 21 Moving paths of the vehicles

Proc. IMechE Vol. 222 Part D: J. Automobile Engineering JAUTO43 F IMechE 2008
Four-wheel-steering vehicles under various operation conditions 547

Fig. 22 Yaw angular velocity versus time

Fig. 23 Lateral acceleration versus time

motion trajectories that 4WS vehicles are more vehicle are stabilized in a short transient time as
manoeuvrable than front-wheel turning vehicles. shown in Figs 19 and 23. It may also be interest-
From Figs 17 and 18, 4WS vehicles require tighter ing to evaluate the peak values of the lateral
space manoeuvrability. The better manoeuvrabil- accelerations shown in Figs 19 and 23. The
ity and stability of 4WS vehicles over that of 2WS following can be found from the figures.
vehicles may also be seen from Figs 18 and 22. In
(a) The peak values of the lateral acceleration of
examining the peak values of the yaw angular
the 4WS vehicle in positive-phase steering are
velocities in the figures for the 4WS and 2WS
much more symmetrically and smoothly
vehicles, the following can be concluded as the
distributed during the transient period of
positive contributions to the manoeuvrability and
time in comparison with that of the 2WS
stability of the 4WS vehicles.
vehicle, as shown in Figs 19 and 23. This also
(a) The absolute peak values of the yaw angular contributes to the stability, safety, and man-
velocity of the 4WS vehicle in positive-phase oeuvrability of 4WS vehicles, although the
steering are the smallest for the two types of transient period is short.
input. (b) The total variation range of the peak values of
(b) The yaw angular velocity of the 4WS vehicle the lateral acceleration for the 4WS vehicle in
in positive-phase steering varies in a smaller positive-phase steering is smaller than that of
range in comparison with that of the 2WS the 2WS vehicle corresponding to the saw
vehicle, for the two input cases. input, during the transient period. However,
the lateral acceleration of the 4WS vehicle
2. The better manoeuvrability and stability of 4WS varies in a larger range than that of the 2WS
vehicles over that of 2WS vehicles may also be vehicle for the half-saw case owing to the
observed from Figs 19 and 23. Although the saw larger lateral force acting on the 4WS vehicle
and half-saw angular turning displacement inputs in positive-phase steering, although the range
may generate a slightly higher transient accelera- is much smaller than that in the saw input
tion, however, the lateral accelerations of the case. It should be noted, however, that the

JAUTO43 F IMechE 2008 Proc. IMechE Vol. 222 Part D: J. Automobile Engineering
548 Q Han and L Dai

distribution of peak values of the lateral detailed experimental analysis. The 4WS and 2WS
acceleration of the 4WS vehicle is much vehicles analysed in the research by Itoh et al. are the
smoother and is concentrated in two loca- vehicles of identical bodies. This is also the case for
lized areas in comparison with that of the the research in this paper. The research by Itoh et al.
2WS vehicle, which has an acceleration that is therefore suitable for qualitative comparison with
varies sharply, as shown in Fig. 23. All these the present research results, which are generated
have positive influences on the stability, including considerations of the vehicle motion and
safety, and manoeuvrability of 4WS vehicles vehicle wheels in turning and braking manoeuvres,
during the transient period. together with other factors and the joint effects of
the non-linearities as described in the context of this
3. Figure 21 shows how easy the turning operation paper. Although the research by Itoh et al. is an
graphically described in Fig. 20 is for a 4WS investigation on a tractor, it is still appropriate for a
vehicle. 4WS vehicles can therefore be easily qualitative comparison with the results of the
controlled to respond to the road surface condi- present research as the research is on ‘turnability’
tions and the curvatures of the road. This results and ‘the difference between the turnabilities of 4WS
in higher manoeuvrability for the driver of a 4WS and 2WS vehicles’, and the turnability of 4WS and
vehicle in operating the vehicle. 2WS vehicles is one of the main research topics of
4. 4WS vehicles are generally more stable in opera- both the research studies.
tions (larger stable regions) in comparison with Itoh et al. found from their investigations that the
2WS vehicles. The conditions for the stability of yaw angular velocity in 4WS vehicles decreased
4WS and 2WS vehicles were given in reference because of the steering of the rear tyres in the same
[10]. direction as the front tyres and that the yaw angular
5. Under the condition of low vehicle speeds, the velocity increased by means of steering the rear tyres
lateral velocity and angular velocity are all in the in the opposite direction to that of the front tyres.
controllable ranges for 2WS vehicles and 4WS This agrees exactly with the conclusions made on
vehicles in both positive- and negative-phase the basis of the present model, with the numerical
steering. In the cases of intermediate or high results exhibit in Figs 10, 13, 18, and 27. The
vehicle speeds, however, 4WS vehicles in positive- conclusions of Itoh et al. on the higher manoeuvr-
phase steering provide the highest manoeuvr- ability of 4WS vehicles in positive-phase steering
ability and therefore higher safety. over that of negative-phase steering and the sharp
turning characteristics of 4WS vehicles in negative-
On the basis of the analysis above, it may be phase turning also agree with the discussions given
observed that the model established in the present on the numerical results that are generated by the
research can be used to analyse the three-dimen- present vehicle model.
sional non-linear behaviour of a vehicle under Additionally, the higher manoeuvrability and
various operating and environmental conditions. In stability of 4WS vehicles over that of 2WS vehicles
fact, such a non-linear dynamic model with multiple are recognized by Furukawa et al. [4]. This also
degrees of freedom for the 4WS vehicles including matches the conclusions and results made on the
the effects of the non-linear tyre force, the air drag basis of the model and numerical simulations of the
and the wind, together with motions of the vehicle present research.
and vehicle wheels in turning and braking man-
oeuvres has not been found in the current literature.
In the current literature, there are limited research 7 CONCLUDING REMARKS
results available for dynamic responses of 4WS
vehicles; very few archived documents show the This research analyses the non-linear behaviour of a
characteristics of the 4WS vehicles with a compar- 4WS vehicle via an accurate and effective approach.
ison with those of 2WS vehicles. The numerical With the development of the approach, the kine-
results obtained from the model in the present matics of the sprung vehicle body and the unsprung
research show a good coherence with the results part of the vehicle can be conveniently studied. A
reported in the limited investigations on the non-linear vehicle model with multiple degrees of
dynamic behaviours of 4WS and 2WS vehicles found freedom coupling the sprung and unsprung compo-
from the archived documents. A study on the nents of the vehicle is developed such that the
behaviours of 4WS and 2WS vehicles was reported effects of non-linear tyre forces, aerodynamic resis-
by Itoh et al. [6], with a validation on the basis of a tance, negative- and positive-phase steering, turning

Proc. IMechE Vol. 222 Part D: J. Automobile Engineering JAUTO43 F IMechE 2008
Four-wheel-steering vehicles under various operation conditions 549

torque output of the vehicle engine, and mechanical 3 Nagai, M., Ueda, E., and Moran, A. Nonlinear
efficiency of the transmission system can be inte- design approach to four-wheel-steering system
grally taken into consideration in vehicle motion using neural networks. Veh. System Dynamics,
1995, 24, 329–342.
analysis. Such an approach for non-linear vehicle
4 Furukawa, Y., Yuhara, N., Sano, S., Takeda, H.,
behaviour analysis has not been found in the current and Matsushita, Y. A review of four-wheel steering
literature. The non-linear model shows the feasibility studies from the viewpoint of vehicle dynamics and
and efficiency for analysing the behaviour of the control. Veh. System Dynamics, 1989, 18, 151–186.
vehicle under normal and other operation condi- 5 You, S. S. and Chai, Y. H. Multi-object control
tions with high accelerations. The numerical results synthesis: an application to 4WS passenger vehi-
generated by the model established show good cles. Mechatronics, 1999, 9, 363–390.
agreement with those of the existing investigations 6 Itoh, H., Oida, A., and Yamazaki, M. Numerical
available in the literature. With a specified operation simulation of a 4WD–4WS tractor turning in a rice
field. J. Terramechanics, 1999, 36, 91–115.
manoeuvre, the model established allows calcula-
7 Itoh, H., Oida, A., and Yamazaki, M. Measurement
tions of the vehicle path, and the force and moments
of forces acting on 4WD–4WS tractor tires during
acting on the vehicle in three-dimensional aspects. steady-state circular turning in a rice field. J.
Considering the sensitivity of the safety and stability Terramechanics, 1995, 32, 263–283.
of 4WS vehicles to the operation conditions, e.g. 8 Abe, M., Kano, Y., Suzuki, K., Shibahata, Y., and
negative-phase steering at high speeds would be Furukawa, Y. Side-slip control to stabilized vehicle
dangerous because of the very high yaw rates lateral motion by direct yaw moment. JSAE Rev.,
produced, it is significant to quantify the kinetic 2001, 22, 413–419.
and dynamic parameters for the proper operation of 9 Liu, Z., Payre, G., and Bourassa, P. Nonlinear
oscillations and chaotic motions in a road vehicle
a 4WS vehicle. The advantages of 4WS vehicles over
system with driver steering control. Nonlinear
2WS vehicles in terms of the ease of turning, Dynamics, 1996, 9, 281–304.
manoeuvrability, stability, smaller turning radius in 10 Dai, L. and Han, Q. Stability and Hopf bifurcation
negative-phase turning, tight-space manoeuvrabil- of a nonlinear model for a four-wheel-steering
ity, ride comfort, and safety are evident based on the vehicle system. Commun. Nonlinear Sci. Numer.
quantitative analysis and numerical simulations Simulation, 2004, 9, 331–341.
presented in this research. 11 Bakker, E., Pacejka, H. B., and Lidener, L. A new
tire model with an application in vehicle dynamics
studies. In Proceedings of the Fourth Autotechnol-
ACKNOWLEDGEMENTS ogies Conference, Monte Carlo, 1989, SAE paper
890087, pp. 83–95.
12 Svendenius, J. and Gäfvert, M. A. Semi-empirical
The authors wish to acknowledge, with thanks, the
dynamic tire model for combined-slip forces. Veh.
financial support from the Natural Sciences and
System Dynamics, 2006, 44, 189–208.
Engineering Research Council of Canada, the Can- 13 Raper, R. L., Bailey, A. C., and Burt, E. C. Inflation
ada Foundation for Innovation, the National Natural pressure and dynamic load effects on soil deforma-
Science Foundation of China (10272046), and the tion and soil–tire interface stresses. Trans. ASAE,
National Natural Science Foundation of Guangdong 1995, 38, 685–690.
Province (020858). 14 Allen, R. W., Chrstos, J. P., and Rosenthal, T. J. A
tire model for use with vehicle dynamics simula-
tions on pavement and off-road surfaces. Veh.
REFERENCES System Dynamics, 1997, 27, 318–322.
15 Deur, J., Asgari, J., and Hrovat, D. A 3D brush-type
1 Cho, Y. H. and Kim, J. Design of optimal four- dynamic tire friction model. Veh. System Dynamics,
wheel steering system. Veh. System Dynamics, 1995, 2004, 42, 133–173.
24, 661–682. 16 Dai, L. and Singh, M. C. A new approach with
2 Akita, T. and Satoh, K. Development of 4WS piecewise-constant arguments to approximate and
control algorithms for an SUV. JSAE Rev., 2003, 4, numerical solutions of oscillatory problems. J.
441–448. Sound Vibr., 2003, 263, 535–548.

JAUTO43 F IMechE 2008 Proc. IMechE Vol. 222 Part D: J. Automobile Engineering

Das könnte Ihnen auch gefallen