Energies 2014, 7, 3748-3763; doi:10.

3390/en7063748
OPEN ACCESS

energies
ISSN 1996-1073
www.mdpi.com/journal/energies
Article

An Acceleration Slip Regulation Strategy for Four-Wheel Drive

Electric Vehicles Based on Sliding Mode Control
Hongwen He *, Jiankun Peng, Rui Xiong and Hao Fan

National Engineering Laboratory for Electric Vehicles, Beijing Institute of Technology, Beijing 100081,
China; E-Mails: [email protected] (J.P.); [email protected] (R.X.); [email protected] (H.F.)

* Author to whom correspondence should be addressed; E-Mail: [email protected];

Tel./Fax: +86-10-6891-4842.

Received: 28 January 2014; in revised form: 15 May 2014 / Accepted: 3 June 2014 /
Published: 17 June 2014

Abstract: This paper presents an acceleration slip regulation (ASR) system for four-wheel
drive (4WD) electric vehicles, which are driven by the front and rear axles simultaneously.
The ASR control strategy includes three control modes: average distribution of inter-axle
torque, optimal distribution of inter-axle torque and independent control of optimal slip rate,
respectively, which are designed based on the torque adaptive principle of inter-axle
differential and sliding mode control theory. Furthermore, in order to accurately describe the
longitudinal tyre force characteristic, a slip rate calculation formula in the form of a state
equation was used for solving the numerical problem posed by the traditional way.
A simulation was carried out with the MATLAB/Simulink software. The simulation results
show that the proposed ASR system can fully use the road friction condition, inhibit the
drive-wheels from slipping, and improve the vehicle longitudinal driving stability.

Keywords: acceleration slip regulation (ASR); four-wheel drive (4WD); electric vehicle;
sliding mode control

1. Introduction

Although the dynamic performance of four-wheel drive (4WD) conventional vehicles is improved
compared with that of two-wheel drive vehicles, the fuel economy is relatively poor. It’s worth noting
that a purely electric vehicle can completely ignore the impact of fuel economy [1,2]. Considering the
motor has the characteristics of low speed and high torque, a pure electric vehicle that employs the 4WD
Energies 2014, 7 3749

scheme can not only improve the vehicle’s acceleration and climbing performance, but also have a
positive impact for reducing emissions and air pollution [3–5]. However, the chassis layout of a high
power motor is usually difficult due to its large volume, but a good distribution layout scheme can divide
one high power and large volume size motor into several small motors, thus realizing a flexible chassis
layout and improving the vehicle’s dynamic performance reliability [6–9].
Electric vehicles have several advantages such as a drive wheel torque that can be acquired easily,
sensitive motor response, and motor torque that can be controlled accurately, etc. Therefore, the advantages
of acceleration slip regulation (ASR) application in an electric vehicle are incomparable to those seen in
a conventional vehicle. Not only PID (Proportion Integration Differentiation) control, but also several
techniques based on nonlinear control have been applied to ASR research [10,11]. A fuzzy logic control
was introduced in ASR control in [12]. Nakakuki et al. [13] proposed the linearization feedback control
for the ASR function. Sliding mode control enables the system architecture having the switching
characteristics changes over time, forces the system into doing the small amplitude and high frequency
movement along a predetermined state trajectory with a certain characteristic, and the control theory was
studied by many scholars due to its good robustness [14–18]. The key technology of ASR is how to
balance the relationship between the driver demand torque, the tyre characteristics and the road friction
conditions [19]. This paper undertakes a specific analysis of these three elements, and an ASR strategy
which contains three control modes is designed for good roads, slippery roads and bad roads, and the
motor output torque is controlled based on sliding control theory. The simulation results show that the
proposed ASR system can choose the suitable control mode according to the different road friction
conditions and take full advantage of the road friction, thus improves the vehicle’s longitudinal
driving stability.

2. System Model

In this paper, the 4WD electric vehicle ASR system layout was as shown in Figure 1. The power battery
pack provides power to the two identical power and volume motors, and the ASR controller sends the
control command in the form of torque control to the two motors, thus the two motors simultaneously
output power externally by directly connecting to the differentials of the front and rear axles.

Figure 1. The four-wheel drive (4WD) electric vehicle acceleration slip regulation (ASR) system layout.

Fx 2 Fx 4
w 2 w 4

w1 w 3
Fx1 Fx 3
Energies 2014, 7 3750

2.1. Vehicle Dynamics Model

Since this paper focuses on the influence of ASR on the vehicle’s longitudinal dynamic
performance, the vehicle dynamics model adopted here contains five degrees of freedom as seen in
Figure 1, the five degrees of freedom include the vehicle longitudinal motion and the rotational
movement of the four wheels. According to the Newton’s law, the motion equations of the vehicle are
derived as follows [20]:
(1) Vehicle longitudinal movement:
4
mu   Fxi (i  1, 2,3, 4) (1)
i 1

where u is longitudinal velocity of vehicle, m is vehicle mass, Fxi is tyre longitudinal force.
(2) Four wheel’s rotation movement:
 wi =Tmi  Fxi rw
I wi ω (i  1, 2,3, 4) (2)

where ωwi is wheel angular velocity, Iwi is wheel rotation inertia, rw is rolling radius of the tyre,
Tmi is motor driving torque.

2.2. Tyre Model

The tyre dynamics are usually modeled with the “Magic Formula” developed by Pacejka et al. [21]
which employs combinations of trigonometric functions to describe the tyre forces exactly. In this
model, the tyre longitudinal force is described as the complex nonlinear function of the slip rate and tyre
vertical load. Thus, the tyre longitudinal force model can be expressed as the following equation:
Fxi  D sin{C arctan[ BX  E ( BX  arctan BX )]} (3)
where B, C, D, E are the stiffness, shape, peak, curvature factor, respectively. X is the input vector for
this model. All of them are described as the function of the tyre vertical load Fz, tyre slip rate λ and
relevant fitting coefficient bi, the fitting coefficients are shown in Table 1 and the specific description
can be shown as follows:
 X  λ  Sh
C =b
 0

 D =b1 Fz 2 +b2 Fz
 b F
 BCD  (b3 Fz  b4 Fz )  e 5 z
2
(4)
 B  BCD / ( B  D )

 E =b6 Fz 2 +b7 Fz +b8

 S h =b9 Fz +b10

Table 1. “Magic Formula” fitting coefficients.

No. 0 1 2 3 4 5
bi 1.02316 21.80968 526.2336 0.09624 250.33146 0.00906
No. 6 7 8 9 10 –
bi −0.00255 0.03726 0.87693 −0.00009 −0.00033 –
Energies 2014, 7 3751

2.3. Motor Model

In this paper, permanent magnet synchronous motors were chosen for constructing the dual-motor
drive system. The motor’s rated torque is 270 Nm and the rated power is 90 kW. Motor models are
usually mainly divided into the theoretical model and the quasi- steady-state model. This paper adopted
the latter one and plotted the motor external characteristic curve according to the characteristic motor
parameters, and acquired the motor speed and peak torque according to the load signal namely the
accelerator open degree, then adopted one-order inertia to complete the dynamic correction on the motor
output torque as the following equation:
 α
 τ s +1  Tmax  n  n  nN
 m
Tm =  (5)
 α  Tmax  n   nN n >n N
 τ m s +1 n

where Tm is the motor’s output torque, Tmax is the motor’s peak torque, α is the degree of openness of
the accelerator, τm is the motor’s responsive time constant, n is the motor speed, nN is the inflection
point speed of the motor’s external curve.

2.4. Slip Rate Calculation Model

The tyre slip rate is the key factor for tyre longitudinal force calculation, the relationship among slip
rate, wheel speed and vehicle velocity can be expressed as follows:
ω wi rw  u
λi  , ω wi  0 (i  1, 2,3, 4) (6)
ω wi rw

The wheel speed ωwi and vehicle velocity u are very small when the vehicle is in transient starting,
and a numerical problem such as λ i   is likely to occur when using Equation (5), it is evidently
inconsistent with the force characteristics of tyre contact points. To solve this problem, this paper adopts
a modified slip rate calculation method which was proposed by Bernard [22]. Hence the Equation (6)
can be changed into a state Equation (7) of ωwi and u:
rω r ω sgn(u )- u
λ i + w wi λ i = w wi (i  1, 2,3, 4) (7)
σx σx

where σ x = Cs Cx , Cs is the tyre longitudinal slip stiffness, Cx is tyre shear stiffness.

The simulation results of tyre slip rate are shown in Figure 2, which are calculated by Equations (5)
and (6) respectively. Obviously, using the state equation to describe slip rate is more in line with the
vehicle real starting process.
As shown in Figure 2b, the tyre slip rate fluctuates significantly in the initial stage, this is because
rr ωi σ x in the Equation (6) can be regarded as small spring damping coefficient when the vehicle
velocity is low, namely the description of tyre slip rate is similar to that of a circular spring. Therefore, it
is necessary to modify Equation (7) as follows [15]:
Energies 2014, 7 3752

i =λ i  λ i u  ulow
 (i  1, 2,3, 4) (8)
i =λ i u  ulow

1 u -wwi rw   u  
λ i = 1+ cos  π   (i  1, 2,3, 4) (9)
2 Cs   ulow  
where λ i is the modifying factor for slip rate, i is the modified slip rate, ulow is the low vehicle
velocity threshold.

Figure 2. Comparison of two methods for tyre slip rate calculation: (a) Simulation results of
Equation (6); (b) Simulation results of Equation (7).

(a) (b)

The simulation result of the modified tyre slip rate is shown in Figure 3, where we can see that the
curve fluctuation in the initial stage was effectively smoothed, proving that the Equation (8) can better
describe the tyre slip rate, especially at low vehicle velocity.

Figure 3. The modified tyre slip rate at a low vehicle velocity.

3. Control Strategies

The cause of tyre slip phenomena is that the wheel drive torque exceeds the torque which is provided
by the road adhesion limitation, and its essence is the balanced relationships between maximum drive
torque, road adhesion conditions and the driver-desired torque. For the 4WD electric vehicle in this
paper, the maximum drive torque is the motor’s maximum torque, the road adhesion conditions are
expressed by the product of the road drive force and wheel radius, namely Troad = Fx × Rw, and the
Energies 2014, 7 3753

driver desired torque Td was obtained from the acceleration pedal signal αpedal (0–1), namely
Tdesire = αpedal × (Tf_max + Tr_max), where Tf_max and Tr_max are the maximum torque values of the front and
rear motors, and both of them are equal due to their identical characteristic parameters, namely
Tf_max = Tr_max = Tmax. So it can be simplified that Td = αpedal ×2 × Tmax and evidently Td ≤ 2 × Tmax.
The specific analysis of the balance relationships between maximum drive torque, road adhesion
conditions and driver-desired torque is given as follows:
(1) Td  2  Tmax  Troad , generally corresponds to a condition of high road adhesion. In this case, the
drive wheel slip will not happen over the entire range of the accelerator pedal, and it’s not necessary
to perform a complex torque distribution for the front and rear axles, just let αf = αr = αpedal.
Although the slip rates of the front and rear axles will be different with the same drive torque due
to the different vertical loads of front and rear axles, the vehicle dynamic performance and
driving stability will not be influenced.
(2) Td  Troad  2  Tmax , generally corresponds to the condition of middle road adhesion. In this case,
although the driver-desired torque Td is smaller than Troad, the slip phenomenon may happen on
the driving shaft with a smaller axle load due to the bad road conditions. So a reasonable
allocation of the drive torque of the front and rear axles is required to ensure Tf + Tr = Td, thereby
avoiding slipping under the condition of avoiding vehicle dynamics loss.
(3) 2  Tmax  Td  Troad , generally corresponds to the condition of low road adhesion. In this case, the
driver-desired torque Td is bigger than Troad , and the inter-axle torque distribution will cause an
unavoidable slip phenomenon. In order to make full use of the road adhesion, the independent
control for front and rear axles is the best control plan, thus the slip rates of the front and rear
axles can be controlled to be equal to the optimal rate slip. In this way Tf + Tr = Troad, and the
vehicle can acquire the biggest power at this moment.
In a word, the control flow of acceleration slip regulation strategy is as shown in Figure 4.

Figure 4. The acceleration slip regulation strategy.

Energies 2014, 7 3754

The specific control algorithm is mainly divided into three control modes as follows.

3.1. Mode 1: Average Distribution of Inter-Axle Torque

This control mode equally assigns the demand torque into the front and rear axle:
T fa  Tra  Tdesire / 2 (10)
where Tfa and Tra are the torque of the front and rear motors with the Mode 1 control, respectively.

3.2. Mode 2: Optimal Distribution of Inter-Axle Torque

If the slip phenomenon occurred, this control mode will be adopted. The slip phenomenon
determining condition is as follows: the front axle slip rate λf > a or the rear axle slip rate λr > a, where a
is the slip threshold and its range is [0.1, 0.2]. Sliding model theory was used here to design the output
torque of motors, the switching function so and reaching law so were defined in the Equations (11)
and (12), respectively:
t
so  (ωr  ω f )  co  (ωr -ω f ) dt (11)
0

so  ko sgn(so ) (12)

where co is the constant switching function coefficient with Mode 2 control, ko is the non-negative reaching
coefficient with Mode 2 control, ωf and ωr are the front and rear axle angular velocity, respectively.
It can be found by substituting Equations (2) and (12) into Equation (11) after derivation:
 1 rw Iw Iw
T fo  Td  ( Fxr  Fxf )  [co (ωr  ω f )]  ko sgn( so )
 2 2 2 2 (13)
Tro  Td  T fo

where Tfo and Tro are the torque of the front and rear motors with the Mode 2 control, respectively.

3.3. Mode 3: Independent Control of Optimal Slip Rate

If the slip phenomenon still occurred after Mode 2 control was adopted, this control mode
will be adopted here to control the two motors independently and ensure the slip rate of front and
rear axle are equal. The switching functions s fi , sri and reaching laws s fi , sri were defined in
Equations (14) and (15), respectively:
s  (λ  λ )  c t (λ  λ )dt
 fi f o fi 0 f o
 t (14)
sri  (λr  λo )  cri  (λr  λo )dt
 0

s fi  k fi sgn(s fi )
 (15)
sri  kri sgn(sri )

where λo is the optimal slip rate and its range is [0.15, 0.2], cfi and cri are the constant switching
function coefficients with Mode 3 control, kfi and kri are the reaching coefficients of non-negative with
Mode 3 control.
Energies 2014, 7 3755

It can be found using the same reasoning that:

 I wω f I wω2f rw I wω2f rw
Tfi  (c fiu  c fi ω f  u) 
 c fi λo  rw Fxf  k fi sgn(s fi )
 u ur u
 (16)
 I wωr I wω2r rw I wω2r rw
Tri  (criu  cri ωr  u)  cri λo  rw Fxr  kri sgn(sri )
 u u u

4. Simulation Results and Analysis

Under the premise of ensuring the vehicle’s longitudinal stability, acceleration slip regulation is
proposed to improve the vehicle’s dynamic performance. To verify the validity and accuracy of the
control strategy, first, different control modes for fixed conditions were subjected to separate
simulation analyses, then the effective switching between different control modes for variable
condition were verified by simulation. The simulation experiments were carried out using the
MATLAB/Simulink (MathWorks, Natick, MA, USA) environment, and the model parameters are as
listed in Table 2.

Table 2. Simulation parameters.

Parameters Values Parameters Values
co 500 kri 80
cfi 50 m 5000 kg
cri 500 nN 4000
Cs 1.81 kg·m2 rw 0.447 m
Cx 1.65 kg·m2 ulow 1.83 m/s
Iw 2.035 kg·m2 τm 200
ko −100 σx 0.91
kfi −200 – –

4.1. Simulation on Average Distribution of Inter-Axle Torque

When the vehicle is traveling on a high adhesion road, the average distribution of inter-axle torque
model is used to control the load signal of the front and rear axle drive motor α f  α r  α pedal .
The simulation results are shown in Figure 5 when the signal of the accelerator pedal α pedal = 0.5 and
the peak road adhesion coefficient μH = 0.80.
As Figure 5a shows, the front and rear axle tyre slip rate varies around different stable values
because of the axle loads, but they were all in a stable slip zone, Figure 5b shows that the vehicle
acceleration began to stabilize at around 1.48 m/s2 under these conditions, then due to the motor speed
increase, the vehicle acceleration will be reduced gradually when the motor workspace transforms
from the constant torque to the constant power region.
Figure 6 shows the simulation of the distribution of the front and rear axle drive torque by axle load
under these conditions. Figure 6a shows that the driver-demanded torque is distributed to the front and
rear axles by the axle load so that the front and rear axle tyre slip rate can be equal. However, by
comparing Figures 6b and 5b, it can be known that vehicle acceleration is consistent under both torque
distribution modes, and the vehicle dynamics aren’t affected.
Energies 2014, 7 3756

Although the slip rate can be guaranteed to be a constant based on the torque distribution by the front
and rear axle loads, in the view of the project realization, the accurate estimation of dynamic load of the
front and rear axle when the vehicle is running requires a large number of external sensors, so the
controller cost is expensive. In contrast, the average distribution of inter-axle torque control is simple and
low cost, etc., although inconsistent tyre slip rate will lead to different front and rear tyre wear, but as
long as there is enough road adhesion ability, the vehicle’s dynamic performance can’t be affected.

Figure 5. Simulation results of average distribution of inter-axle torque on good roads:

(a) The front and rear axle tyre slip rate; (b) Vehicle acceleration.

Acceleration /ms-2

(a) (b)

Figure 6. Simulation results of torque distribution by axle load on good roads: (a) The front
and rear axle tyre slip rate; (b) Vehicle acceleration.

(a) (b)

4.2. Simulation of Optimal Distribution of Inter-Axle Torque

When the driver steps on the acceleration pedal lightly on a mid-adhesion road, although the driver
demanded torque is less than the total road adhesion at that time, if the average distribution of inter-axle
torque control is still used, the drive axle which has the smaller axle load suffers a smaller ground
tangential force, and the corresponding road adhesion is smaller, which may seriously slip due to its
relatively large driving torque. Similarly, the drive axle which has smaller axle load has smaller road
adhesion, and the corresponding drive torque is smaller, resulting in it being unable to take full
advantage of the road adhesion conditions. Therefore, the design is based on an optimal distribution of
inter-axle torque to have a reasonable distribution of the drive torque between the front and rear axles.
Energies 2014, 7 3757

The simulation results of slip rate between front and rear axle under the optimal distribution of inter-axle
torque control and without torque control are shown in Figure 7, which shows the simulation results of
the slip rate with control and without control for the driver accelerator pedal signal αpedal = 0.30 and peak
road friction coefficient μH = 0.10.
Figure 7b shows that, if torque are distributed equally between the axles, the front axle with less axle
load seriously slips, and the rear axle with large axle load maintains a slip rate λ = 5% , then the vehicle
fails to make full use of the road adhesion for the axle. Figure 7a shows that after using the optimal
distribution of inter-axle torque control, the slip rate of the front and rear axles remains the same in 0.5 s
and the slip rate stabilized at around λ = 7.4% finally. Figure 7c,d are under the vehicle acceleration
control corresponding to Figure 7a,b respectively, and in can be known that the vehicle acceleration is
finally stabilized at 0.78 m/s2 under the optimal distribution of inter-axle torque control based on SMC
(Slip Mode Control), while the vehicle acceleration eventually stabilizes at 0.65 m/s2 under the average
torque distribution control. In poor road conditions, the optimal distribution of inter-axle torque control
based on SMC was significantly better than the average torque distribution control, not only ensuring the
vehicle’s longitudinal driving stability, but also improve the vehicle’s dynamic performance.

Figure 7. Simulation of stepping on the pedal lightly to accelerate on a mid-adhesion road:

(a) Optimal torque of inter-axle distribution control; (b) Average torque distribution control;
(c) Acceleration under optimal torque of the inter-axle distribution control; (d) Acceleration
under the average torque distribution control.

(a) (b)

( c) (d)

The sliding mode control process under these conditions is shown in Figure 8. Figure 8a shows the
change of the speed deviation of the front and rear axle tyres, and a deviation close to zero can be seen at
around 0.4 s. Figure 8b shows the change of front and rear axle motor control torque output basically
fluctuating at a stable value within ±10 Nm.
Energies 2014, 7 3758

Figure 8. The change of sliding mode surface and controlled variables in the sliding mode
control process: (a) Convergence process of the front and rear axle speed deviation;
(b) Output torque of the front and rear axle motor under sliding mode control.

(a) (b)

4.3. Simulation of Independent Control of Optimal Slip Rate

When the vehicle drives on s bad road and the driver-demanded torque is large, Tdesire > Troad , tyre
slip can’t be avoided with the torque distribution between the axis based on optimal distribution of
inter-axle torque control, so independent control of the optimal slip rate is required, the front and rear
axle slip rate remains the best slip rate (referred to herein λopt = 15%), in order to fully use the road
adhesion conditions and get maximum power. The signal of the driver accelerator pedal is set at
αpedal = 0.8, and the peak road adhesion coefficient is set at μ H = 0.10 , and the independent control of
optimal slip rate and the optimal distribution of inter-axle torque control results are shown in Figure 9.

Figure 9. Simulation results of stepping on the pedal heavily on snowy roads: (a) Independent
control of the optimal slip rate; (b) Optimal distribution of inter-axle control; (c) Acceleration
under independent control of optimal slip rate control; (d) Acceleration under optimal
distribution of inter-axle torque control.

(a) (b)

( c) (d)
Energies 2014, 7 3759

Figure 9b shows that the front and rear tyres seriously slip under optimal distribution of inter-axle
torque control, the driver-demanded torque is much greater than road adhesion at this time, and the
vehicle reaches an acceleration of 0.64 m/s2 in this mode, as shown in Figure 9d. As shown in Figure 9a,
independent control of the optimal slip rate control can not only stabilize the front and rear axle tyres
at the best slip point (λopt = 15%), but also maximize use of the road adhesion, to obtain an acceleration
of 0.98 m/s2 which was shown in Figure 9c.
The front and rear axle independent control process under this condition is shown in Figure 10.
Figure 10a,b are the deviation change of the drive axles tyres slip rate and optimum slip rate, respectively,
which are basically within 0.4 s, the two axles tyres can be controlled in a stable slip area. Figure 10c
shows front and rear axle output torque of the drive motor controls.

Figure 10. Front and rear axle input torque under different control modes: (a) Front axle slip
rate control deviation; (b) Rear axle slip rate control deviation; (c) Output torque under
independent control of the optimal slip rate mode.

(a) (b)

( c)

4.4. Comprehensive Simulation and Analysis of the Acceleration Slip Regulation Control Strategy

To further verify that the ASR control strategy can be effective for three different conditions to switch
under constant driver-demanded torque, changing road conditions, and driver demanded torque changes,
the road conditions remain unchanged to simulate two conditions. The simulation results are shown in
Figures 11 and 12. Where Figure 11c–f correspond to the vehicle driving characteristics under the
conditions of Figure 11a,b, respectively.
Figure 11a shows that when the driver steps on the pedal lightly (αpedal = 0.3) by a dry road (μH = 0.80)
to start with icy ground (μH = 0.10), at about 3.8 s the controller detects front axle tyre slip λ > 15%, so it
switches the average distribution mode to the axis torque distribution based on SMC mode, as shown in
Figure 11c, where “Control mode = 1” represents the conventional average torque distribution mode,
Energies 2014, 7 3760

“Control mode = 2” represents the axis torque distribution based on SMC mode, “Control mode = 3”
represents front and rear axle independent control mode. Then under the optimal distribution of inter-axle
torque control mode, the front and rear axle slip rate is basically the same at about 4.8 s, then it
stabilizes at around 6 s. Since Tdesire < Troad is always correct, before the control mode switching, vehicle
acceleration always follows the driver’s intention, steady at around 0.78 m/s2, as shown in Figure 11e.

Figure 11. Simulation results of torque distribution on changing roads: (a) Changes of slip
rate with light pedal; (b) Changes of slip rate with heavy pedal; (c) Control mode switching;
(d) Control mode switching; (e) Vehicle performance under light pedal; (f) Vehicle performance
under heavy pedal.

(a) (b)

( c) (d)

( e) (f)

Figure 11b shows that, when the driver steps on the pedal heavily (αpedal = 0.8) by a dry road ground
(μH = 0.80) starting with icy ground (μH = 0.10), the controller detects that the front axle slips at about
2.1 s, mode switches from the average distribution of the torque to distribution of inter-axle torque,
Energies 2014, 7 3761

but Tdesire >Troad at this time, so distribution of inter-axle torque control is unable to prevent tyre
slippage, so it continues to switch from distribution of inter-axle torque mode to the front and rear
axles independent control mode, as shown in Figure 11d. Since Tdesire >Troad at this time, the front and
rear axle independent control mode can take advantage of the road adhesion to the optimum dynamic,
the acceleration eventually stabilizes at around 0.98 m/s2, as shown in Figure 11f.
To further verify the torque distribution strategy between the axis effective switching between the
three control modes, we simulated under the same road conditions (μH = 0.10), the changing driver-
demanded torque conditions, as shown in Figure 12.
Figure 12 shows that, when the vehicle is traveling on the icy road (αpedal = 0.1) with pedal signal
αpedal = 0.1, the driver-demanded torque is small, the front and rear drive axles are both under road
adhesion conditions, and the controller controls the front and rear axle drive motors under average
torque distribution mode. Then at about 4 s, the pedal changes to αpedal = 0.3, then the slip rate of the front
axle has a slip trend, and then the controller switches to the axis torque distribution based on SMC mode.
When the pedal depth further increases until αpedal = 0.8, the total ground adhesion is not enough,
so slippage can’t be avoided by torque distribution between the axles, so the controller switches to the
front and rear axles independent control mode. With increasing speed, the motor transforms from
constant torque region to constant power region, under the constant accelerator pedal, the
driver-demanded torque is decreasing, then the mode switches to “the axis torque distribution based on
SMC mode” again, as shown in Figure 12b with the oval label. Figure 12d shows the change of vehicle
acceleration under these conditions, and the vehicle dynamics can follow the driver’s intentions well.
It can thus be seen that the proposed inter-axis torque distribution control strategy can effectively
switch between different conditions, ensuring the vehicle dynamics while improving the vehicle's
longitudinal driving stability.

Figure 12. Simulation results under the same road with the changing driver accelerator
pedal: (a) Driver accelerator pedal signal; (b) The change of control mode under inter-axle
torque distribution control strategy; (c) Change of slip rate; (d) Vehicle acceleration.

Time/s

(a) (b)

( c) (d)
Energies 2014, 7 3762

5. Conclusions

(1) Aiming at the 4WD electric vehicle, which was driven by front and rear independent motors, a
model of the ASR system was established.
(2) Compared with the conventional method of slip rate calculation, using the state equation of slip
rate can be more accurate to describe the tyre slip process in a low vehicle velocity situation.
(3) An ASR control strategy which contains three torque distribution mode was designed, namely
average distribution of inter-axle torque for high road adhesion, optimal distribution of
inter-axle torque for middle road adhesion and independent control of optimal slip rate for low
road adhesion. Several simulations were carried out with MATLAB/Simulink, and the
simulation results with some comparisons show that, the proposed strategy could realize the
transformation among different control modes, thus fully use the road adhesion conditions,
make the vehicle’s dynamic performance to follow the driver’s wishes. As a result, the vehicle
longitudinal drive stability and dynamic performance are ensured.

Acknowledgments

This work was supported by the National High Technology Research and Development Program of
China (2012AA111603, 2013BAG05B00) in part, the Program for New Century Excellent Talents in
University (NCET-11-0785) and Beijing Institute of Technology Post Graduate Students Innovation
Foundation in part. The author would also like to thank the reviewers for their corrections and
helpful suggestions.

Author Contributions

Hongwen He and Jiankun Peng built the dynamic model of ASR for 4WDEV and performed the
simulations, Rui Xiong designed the control modes, Hao Fan debugged the algorithm. All authors
carried out data analysis, discussed the results and contributed to writing the paper.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Wang, J.; Wang, Q.; Jin, L.; Song, C. Independent wheel torque control of 4WD electric vehicle for
differential drive assisted steering. Mechatronics 2011, 21, 63–76.
2. Kang, J.; Yoo, J.; Yi, K. Driving control algorithm for maneuverability, lateral stability, and
rollover prevention of 4WD electric vehicles with independently driven front and rear wheels.
IEEE Trans. Veh. Technol. 2011, 60, 2987–3001.
3. Dasgupta, K. Analysis of a hydrostatic transmission system using low speed high torque motor.
Mech. Mach. Theory 2000, 35, 1481–1499.
4. Gasbaoui, B.; Nasri, A. A Novel 4WD Electric Vehicle Control Strategy Based on Direct Torque
Control Space Vector Modulation Technique. Intell. Control. Autom. 2012, 3, 236–242.
Energies 2014, 7 3763

5. Xue, X.; Cheng, K.; Ng, T.; Cheung, N.C. Multi-objective optimization design of in-wheel
switched reluctance motors in electric vehicles. IEEE Trans. Ind. Electron. 2010, 57, 2980–2987.
6. Tong, Q. Simulation Study on the Longitudinal Dynamics of a Full Drive Vehicle with
Two Separated Motors. Master’s Thesis, Beijing Institute of Technology, Beijing, China, 2013.
7. Jalali, K.; Uchida, T.; McPhee, J.; Lambert, S. Development of a Fuzzy Slip Control System for
Electric Vehicles with In-wheel Motors. SAE Int. J. Altern. Powertrains 2012, 1, 46–64.
8. Liu, W.; He, H.; Peng, J. Driving Control Research for Longitudinal Dynamics of Electric
Vehicles with Independently Driven Front and Rear Wheels. Math. Probl. Eng. 2013, 2013,
doi:10.1155/2013/408965.
9. Zhao, Z. Study of acceleration slip regulation strategy for four wheel drive hybrid electric car.
Chin. J. Mech. Eng. 2011, 47, 83–98.
10. Sakai, S.; Hori, Y. Advantage of electric motor for anti-skid control of electric vehicle. Eur. Power
Electron. Drives J. 2001, 11, 26–32.
11. Yin, D.; Oh, S.; Hori, Y. A novel traction control for EV based on maximum transmissible torque
estimation. IEEE Trans. Ind. Electron. 2009, 56, 2086–2094.
12. Chen, F.W.; Liao, T.L. Nonlinear linearization controller and genetic algorithm based fuzzy logic
controller for ABS systems and their comparison. Int. J. Veh. Des. 2000, 24, 334–349.
13. Nakakuki, T.; Shen, T.; Tamura, K. Adaptive control approach to uncertain longitudinal tire slip in
traction control of vehicles. Asian J. Control. 2008, 10, 67–73.
14. Liu, J.K. Sliding Mode Control Design and MATLAB Simulation, 2nd ed.; Tsinghua University
Press: Beijing, China, 2012.
15. Ameodeo, M.; Ferrara, A.; Terzaghi, R.; Vecchio, C. Wheel Slip Control via Second–Order
Sliding-Model Generation. IEEE Trans. Intell. Trans. Syst. 2010, 11, 122–131.
16. Bashash, S.; Fathy, H.K. Transport-based load modeling and sliding mode control of plug-in
electric vehicles for robust renewable power tracking. IEEE Trans. Smart Grid 2012, 3, 526–534.
17. Subudhi, B.; Ge, S.S. Sliding-Mode-Observer-Based Adaptive Slip rate Control for Electric and
Hybrid Vehicles. IEEE Trans. Intell. Trans. Syst. 2012, 13, 1617–1626.
18. Kim, J.; Park, C.; Hwang, S.; Hori, Y. Control algorithm for an independent motor-drive vehicle.
IEEE Trans. Veh. Technol. 2010, 59, 3213–3222.
19. Austin, L.; Morrey, D. Recent advances in antilock braking systems and traction control systems.
Proc. Inst. Mech. Eng. Part D 2000, 214, 625–638.
20. Peng, J.K.; He, H.W.; Feng, N.L. Simulation Research on an Electric Vehicle Chassis System
Based on a Collaborative Control System. Energies 2013, 6, 312–328.
21. Pacejka, H.B.; Bakker, E. The magic formula tyre model. Veh. Syst. Dyn. 1992, 21, 1–18.
22. Clover, C.; Bernard, J. Longitudinal tire dynamics. Veh. Syst. Dyn. 1998, 29, 231–260.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).