Effect of the Tire Lateral Force Saturation

on Stability Control of 4WDEV Steering

1. Niu Jing
Tianshui Normal University 741001
Tianshui Gansu China
3. Lin Cheng
sensily@163.com
National Engineering Laboratory for Electric Vehicles
2. Lu Yulin Beijing Institute of Technology 10081
Tianshui Normal University 741001 Beijing China
Tianshui Gansu China lincheng@bit.edu.cn
lu_yu_lin@163.com

Abstract: On the acceleration cornering, according to the ideal Therefore, to achieve the precise control of 4WDEV
linear tire model and the vehicle dynamic model, there is a problem handling stability, the control algorithm must be designed for
of the vehicle stability control.And the control effect is not obvious. the non-linear vehicle system, especially the non-linear of the
Based on the "uniform model" of tire, this paper establishes the tire, so as to ensure the control effect. Based on the nonlinear
multiple degrees of freedom nonlinear dynamic model of 4WDEV model of tires, the simulation model of 4WDEV is established
using Matlab/Simulink software, analyzes the stability of vehicle by using Matlab /Simulink simulation software. The influence
control, accounting the saturation of the lateral force of the tire. And of the lateral saturation characteristic of the tire on the yaw
it is proved that the fluctuation of the longitudinal velocity of the
rate of the vehicle is discussed and verified by the real vehicle
vehicle and the acceleration of the steering failure are effectively
tests. It has good accuracy, and provides a theoretical basis for
suppressed by the control strategy of the lateral force saturation. The
theoretical basis for the better implementation of the 4WDEV control
the better realization of the future control algorithm.
stability control algorithm is provided.
II. THE TIRE NON-LINEAR MODEL OF 4WDEV
key words: 4WDEV; tire; lateral force saturation
In the past we always use "friction" concept to analyze the
I. INTRODUCTION impact on handling and stability of vehicle simply on handling
and stability study. But most experiences show that, when the
Independent electric drive technology widely called as the
vehicle is speeding-up or sudden braking, the directional
"electric vehicle drive system development direction" and
"suitable for electric vehicle chassis technology", is the hot stability deteriorates significantly. Therefore, we should focus
research currently at home and abroad [1]. In the four-wheel on vehicle handling and stability control on longitudinal slip
independent electric drive electric vehicle (4WDEV) and lateral deflection of tires, at this time automobile side slip
manipulation of stable control, the laboratory of the University angle is often greater than 5 degrees, and there is a nonlinear
of Tokyo Yoici Hori and Japan's Tokyo Agricultural relationship between the lateral characteristics of the tire [4].
University Masao Nagai researched earliest, and their level is Tire model is generally divided into two categories: one is
the most advanced. Researchers of the Masao Nagai the theoretical model on behalf of GIM model, Fiala and Sakai
laboratory proposed an integrated control method of electric
models; Second, empirical models, more influential are the
vehicle direct yaw moment control (DYC), four-wheel
magic formula and Guo Konghui's "unified model".
steering and active steering, which is based on the dynamic
torque distribution [2]. Manabu Omae in Japan's Keio According to [5], "unified model" on braking and
University, has realized the traveling direction redundancy speeding-up conditions longitudinal slip rate and lateral slip
control by using dynamic torque distribution generated yaw rate of tires are defined as (1):
moment [8]. China has achieved plentiful results in adaptive
control based on dynamic torque distribution, driving anti- Vx − ω r (1)
Sx = , S y = (1 + S x ) tan α
skid, differential power steering, steering stability and ωr
simulation and experiment.
The relative longitudinal and lateral slip rate are defined as
Existing electric four-wheel independent drive control (2):
adopts more direct method of yaw moment control (DYC),
which has better control effects only in an non-extreme kxSx kySy (2)
φx = ,φ =
conditions [3]. However, in the corners acceleration conditions, μ FZ y μ FZ
vehicle handling and stability problems are more prominent,
mainly in: (1) the yaw rate is lagging and overshooting Total relative slip ratio is defined as (3):
relative to the steering wheel angle; (2) to accelerate the
steering will produce inadequate steering;(3) unilateral φ = φ x2 + φ y2 (3)
steering process, the front axle with a higher adhesion
coefficient than the rear axle, the front axle easier to enter the Among them, kx is the tire longitudinal stiffness, ky is
non-linear area, resulting in poor vehicle steering. lateral stiffness, ȝ is the friction coefficient of the tire, ‫׋‬x is the

978-1-5090-6161-7/17/$31.00 2017
c IEEE 1095
relative longitudinal slip rate, ‫׋‬y is the lateral relative slip rate,
Ȧ is the wheel rotation angular velocity, r is the wheel rolling
radius.

TABLE I. VEHICLE PARAMETERS

Parameter Value Parameter Value

Distance from front
Vehicle mass m/kg 1022 1.197
to center a/m
Fig.2. Variation of the tire lateral force on slip rate 20%
Yaw moment of inertia Distance from rear
1471 1.203
J/(kg.m2) to center a/m It can be seen from the simulation results that there are two
The front side lateral Front sliding
stiffness k1(N/rad)
-70000
stiffness/(N/rad)
-70000 variation intervals in the lateral characteristics of the tire.
Rear lateral deflection Rear sliding When the lateral angle is relatively small, the lateral force is
-90000 -90000 linearly related to the side angle. When the lateral angle
stiffness k2(N/rad) stiffness/(N/rad)
Track d/m 1.367
The tire friction
0.5
crosses the critical value, Force is basically no longer
coefficient increased, was saturated state. In addition, the longitudinal slip
rate of the tire has a great influence on the critical value of the
lateral deflection angle. The higher the longitudinal slip rate is,
In the tire coordinates, Fx is defined as the longitudinal
the greater the critical value of the side angle is about 8 °, The
force, Fy is the lateral force , the resultant force is F. The ratio
lateral force reaches the saturation value of 2437N and the
of longitudinal force to longitudinal maximum friction force is longitudinal slip rate is 20%, the lateral angle is about 27 °, the
Fx , the ratio of lateral force to lateral maximum friction force lateral force reaches the saturation value of 2442N, and the
is Fy , the resultant force is F , then: longitudinal slip rate has little effect on the saturation value of
the lateral force.
Fx =
Fx
, Fy =
Fy (4)
III. 4WDEV VEHICLE DYNAMICS MODEL
μ FZ μ FZ
This article only considers the movement on the horizontal
Through a lot of theoretical analysis and experimental plane of the vehicle, ignoring the vehicle pitching motion and
research, it is shown that the total dimensionless stress of the the roll movement which affects the characteristics of the
tire under the combined condition of longitudinal slip and side suspension system, and the movement of the vehicle in the
slip, F is: vertical direction mainly caused by road roughness. According
to the Ackermann-Jeantand steering model, the 4WDEV
F = 1 − e x p ( − φ − E φ 2 − ( E 2 + 1 / 1 2 )φ 3 ) (5) three-degree-of-freedom vehicle model is established, showed
in Fig.3, where the three degrees of freedom include the
The longitudinal dimensionless lateral force and the longitudinal movement of the vehicle along the X axis, the
dimensionless lateral force are distributed proportionally yaw movement around the Z axis and the horizontal
under the combined condition of longitudinal and lateral slip: movement along the Y axis.
φx φ A. Principle of 4WDEV Stability Based on Tire Non-linear
Fx = F , Fy = F y
φ φ (6) The idea of controlling vehicle handling stability at home
and abroad is mainly based on state error feedback[6] . First,
According to the vehicle parameters (as shown in Table I), the driver's intention is judged based on the steering input of
the simulation results are shown in Fig.1-2 according to the the vehicle and the motion state of the vehicle, and the ideal
non-linear model based on dynamic simulation in state of the vehicle is stabilized with the input of the driver as
Matlab/Simulink software. the control target. This target is compared with the actual
motion state of the vehicle, and the next step of the control
action is determined according to its deviation.
Vehicles are prone to "unstable" state occurs mainly in the
large lateral acceleration and a low adhesion surface [6].
Because in these cases, the tire lateral force easily reaches its
limit, which is saturated lateral force, cornering characteristics
of the wheel tends non-linear. In the design of the previous
steering stability control system, often with the simplified
Fig.1. Variation of the tire lateral force linear two-degree-of-freedom two-wheel vehicle model as the
basis, in the driver's steering wheel angle and the speed of the
input, the ideal vehicle center of mass angle and yaw rate are
achieved, and as a control target to adjust the size of the
vehicle drive torque. Thus, the real situation of the vehicle's

1096 2017 12th IEEE Conference on Industrial Electronics and Applications (ICIEA)
unstable state is that as the driver further increases the steering tire model to realize the lateral force saturation state pre-
wheel angle, the lateral force of the wheel begins to saturate judgment, using the front and rear axle left and right wheel
and no longer increases, but the driver can not perceive the drive torque difference Dynamic control to achieve the vehicle
state and still further increase the steering wheel angle[7-10]. to improve the yaw rate response to the control target, the
The result of the steering wheel angle at this time as the control model framework shown in Fig.4 below.
vehicle handling stability control the ideal target input will According to the setting conditions, the driver model will
produce an error.Therefore, the steering output the steering wheel angle and the drive / brake signal.

Fig.4. The frame of control mode

The input variable of the lateral force predicting module is the
Fig.3. Three degrees freedom vehicle dynamics model steering wheel angle and the longitudinal speed. The output
wheel angle at the time of saturation of the wheel's lateral variables are the active yaw moment and the lateral force
force must be taken as the input of the steering stability saturation The input of the torque distribution module is the
control system to input the ideal state of the vehicle. steering wheel angle, the longitudinal speed, the lateral
acceleration, the lateral speed, and the direction of the steering
B. Mathematical model
wheel[11]. The center of gravity angle and the actual yaw rate,
In the establishment of the differential equation of the the output variable for the vehicle demand for the active yaw
vehicle movement, it is assumed that the driving torque moment. In order to avoid the two sides of the motor capacity
provided by the four wheel motors in the steering process is is not equal, the demand of a side of the motor saturation, first
not large and there is no effect of air resistance, and the tire of all the absolute value of the demand on both sides of the
characteristic change and tire positive moment caused by the motor is limited to the current left and right sides of the motor
change of the load are neglected. The front wheel angle is can produce flooding / braking torque absolute value
generally small, ignoring the sinį , the vehicle three-degree- Minimum value, and then subtracts the required driving force
of-freedom differential equation is as follows: from the maximum drive torque currently available from the
<
mVx ( β + ωr ) = Fy 2l + Fy 2 r + ( Fy1l + Fy1r ) cos δ drive side motor as the currently remaining drive capability
(7) that can be used to accelerate, multiplied by the accelerator
•
J ω r = ( F y 1l + F y 1 r ) cos δ ⋅ a − b ( F y 2 l + F y 2 r ) + M D
pedal signal representing the driver's acceleration demand,
driven by this acceleration demand Force is superimposed on
d d
M D = ( Fx1 r − Fx1l ) cos δ ⋅ + ( Fx 2 l − Fx 2 r ) ⋅ both sides of the motor drive / braking torque, in order to
2 2 (8)
achieve priority to meet the needs of yaw moment.
Vy
β=
Vx (9) IV. SIMULATION ANALYSIS
aωr Based on Matlab / Simulink dynamic simulation platform,
α1 = −δ + β + (10)
Vx a three-degree-freedom vehicle nonlinear model with tire
bω "unified model" is established. The tire linear model is
α2 = β − r (11) established according to (12).
Vx •
In (7)-(11), Fx represents the driving force, Fy represents x ( t ) = Ax ( t ) + Bu ( t )
the lateral force, subscripts 1l and 1r represent the left front y ( t ) = Cx ( t )
wheel and the right front wheel respectively, and subscripts 2l (12)
ª2(k1 +k2) 2(ak1 −bk2) º ª −2k1 º
and 2r represent the left and right rear wheels, Ȧr is the yaw x = ( β , ωr ) « mv , mv2 −1 »
T
« mv »
rate of the vehicle, d is the wheel pitch, Į1 is the front wheel u =δ A= «
x x
» B=« x »
slip angle, ъ2 is the rear wheel slip angle, and MD is the active «2(ak1 −bk2) 2(a k1 +b k2) »
2 2
« −2ak1 »
yaw moment. Definition: y = ω r , « J ,
Jvx
» «¬ J »¼
¬ ¼, ,

C. Control model frame

In order to realize the pre-determination of the lateral force C=[0,1].
In the simulation process, assuming the pavement level,
saturation of the front wheel and to improve the vehicle's yaw
with reference to GB / T 6323.2-1994, the front wheel gives a
rate response, the control model uses the front wheel nonlinear
45° angle step signal, respectively, for the vehicle side and

2017 12th IEEE Conference on Industrial Electronics and Applications (ICIEA) 1097
longitudinal slip rate of 20% (general ABS working time). The
simulation results are shown in Fig.5-8. In these figures
“Consider” means that we consider the effects of the tire
lateral force on stability control of vehicle. And “Non-
consider” means that we do not consider the effects of the tire
lateral force.
In the case of pure side, the input vehicle speed signal is
20km/h, and the simulation time is 100s. The simulation starts
when the front wheel lateral force is saturated.
Enter the longitudinal slip rate of 20% signal, the vehicle Fig.9. Pure side speed comparison
from 20km/h to the longitudinal slip rate of 20%, to maintain
the vehicle driving force unchanged, the simulation starts

Fig.10. Curve acceleration speed comparison

Fig.5. Pure side lateral yaw velocity TABLE II. COMPARISON OF SIMULATION RESULTS
Time t/s 0 20 40 60
Pure side Ȧ/˄rad/s˅ 0 682 1320 1876
Longitudinal slip rate 20% Ȧ/˄ 0 973 1816 2814
rad/s˅

when the front wheel lateral force is saturated.And the

simulation time is 100s.
It can be seen from the simulation results of Fig. 5-8 that
the speed of the vehicle has little effect on the yaw rate of the
vehicle and the yaw rate of the vehicle is basically convergent.
Fig.6. Pure side vehicle speed
However, without considering the saturation characteristics of
the tire lateral force, the vehicle yaw rate increases
substantially linearly as the driver turns the steering wheel,
and the greater the longitudinal slip rate, the faster the vehicle
yaw rate increases. The results are shown in Table II. below.
V. REAL VEHICLE TEST VERIFICATION AND ANALYSIS
This test uses a high-speed distributed drive electric
vehicle platform for real vehicle test, the vehicle parameters
shown in Table I. above.
Fig.7. Slip 20% lateral yaw velocity
A. Steering wheel angle step input test
From the vehicle steering power characteristics curve
shows that the low-speed steering power control effect is more
obvious, so select the 20km / h speed steering wheel angle
step input test conditions to consider the lateral force
saturation characteristics of the control strategy for vehicle
traffic stability the validity of sex. During the test, when the
vehicle reaches the target vehicle speed, the test record
personnel begin to record the data, the driver to the fastest
Fig.8. Slip 20% vehicle speed speed of the steering wheel to reach the target angle (80°) and
to maintain the test vehicle to a steady state, test speed error
control in ±10%. The test results are shown in Fig.9.

1098 2017 12th IEEE Conference on Industrial Electronics and Applications (ICIEA)
B. Corners accelerating test the driver's model identification and the communication speed
This study mainly investigates the influence of the of the driver model and the handling stability control system.
longitudinal slip rate on the stability of the vehicle. According
to the vehicle speed sensor and the crankshaft speed sensor
signal, the vehicle speed is monitored in real time when the
longitudinal slip rate is 20%, the position of the accelerator References
pedal is changed, and the steering wheel angle is gradually [1]Chen Qingquan. Modern electric vehicles, motor drives and power
adjusted 120 ° to keep the steering wheel angle unchanged, electronics technology [M]. Chen Qingquan academician, Beijing:
start recording data, data recorded after the deceleration stop, Mechanical Industry Press, 2005.7.
the test results shown in Fig.10. [2]Wang Y Q,Motion Control of Electric Vehicle by Distribution of Traction
Forces[J],Trans.IEE-Japan, Vo 1.116,No.3,1996: 279-284.
It can be seen from Fig. 9-10 that the vehicle speed is [3]OmaeM,Shimizu H and Hashimoto N.Traction-force-distribution control of
basically stable between 20 ~ 23km / h and the yaw rate of the EV with in-wheel motors for lane keeping[C/CD].Proceedings of the 18th
vehicle is slightly better when the control strategy considering International Electric Vehicle Symposium,2001.
the saturation characteristic of the lateral force of the tire is [4]JIN Chi, XIONG Lu, YANG Pengfei. A control allocation algorithm for
In-wheel-motor driven vehicles with coupling tire
involved. In the case of acceleration, the effect of longitudinal characteristics[C/CD].Proceedings of the International Symposium on
slip is obvious, and the fluctuation of vehicle speed is more Advanced Vehicle Control,2012,Korea,Seoul.
obvious than that of pure side, but it is also stable between 45 [5]Guo Konghui, Liu Qing. Theoretical Model of Unsteady Tire Cornering
Tire Considering Complex Deformation of Carcass. Automotive
~ 63km / h. The response of vehicle yaw rate is good. Engineering, 1997,19 (2): 65 ~ 71.
[6]HE Peng, SHAO Yang. Stability Control and Optimal Power Allocation
VI. CONCLUSIONS Method for Four-wheel Independent Drive Electric Vehicle [J]. Journal of
1) Considering the complexity of the control system and the Hebei University of Technology, 2007,36 (4): 26-32.
[7]XIONG Lu, YU Zhuoping, WANG Yang, etal. Vehicle dynamics control
control precision, the traditional four-wheel independent of four in-wheel motor drive electric vehicle using gain scheduling based
wheel motor drives the vehicle steering stability control on tire cornering stiffness estimation[J]. Vehicle System Dynamics, 2012,
system, which uses the linear model to control the yaw rate of 50(6):831-846.
[8]KANGHYUN N, HIROSHI F, YOICHI H. Design of an adaptive sliding
the target vehicle increases in the vehicle longitudinal slip rate mode controller for robust yaw stabilisation of in-wheel-motor-driven
When there is a large error, and with the steering wheel to electric vehicles[J]. International Journal of Vehicle Designˈ2015ˈ
further increase the angle, the error will intensify; 67(1)˖98-113ˊ
2) Under the condition of large lateral acceleration and low [9]Lin Cheng, Peng Chunlei, Cao Wanke. Sliding mode variable structure
control for independent driving of electric vehicle stability [J]. Automotive
adhesion coefficient, when the vehicle has obvious Engineering, 2015, 37 (2): 132 -138.
longitudinal slippage, the linear stability control system of [10]Zhang Jinzhu, Zhang Hongtian, Sun Yuantao. Yaw moment control for
linear model is used to control the yaw rate of the target electric vehicle stability [J]. Journal of Electric and Control Technology,
2012, 16 (6): 75-80.
vehicle will appear unsteady unsteady Working condition, the [11]YANG Pengfei, XIONG Lu, YANG Chen, etal. Simulation and
control effect is not good; evaluation of stability control strategies of four in-wheel-motored electric
3) In order to improve the controllability of 4WDEV vehicle, vehicle[C /CD]//T he 2nd IF T0MM Asian Conference on Mechanism
the non-linear model based on nonlinear tire model and multi and Machine Science. 2012, Tokyo , Japan.
degrees freedom model is optimal. The expected value of the
control target should be set in the tire side force saturation
State is appropriate, of course, the robustness and accuracy of
this system is also worth exploring;
4)In order to obtain the better handling stability of the four-
wheel electric drive vehicle, we will need to further improve

2017 12th IEEE Conference on Industrial Electronics and Applications (ICIEA) 1099