580 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 63, NO.
2, FEBRUARY 2014
Robust Sliding Mode-Based Learning Control for
Steer-by-Wire Systems in Modern Vehicles
Manh Tuan Do, Zhihong Man, Cishen Zhang, Hai Wang, and Fei Siang Tay
Abstract—In this paper, a robust sliding mode learning control
(SMLC) scheme is developed for steer-by-wire (SbW) systems. It
is shown that an SbW system with uncertain system parameters
and unknown external disturbance from the interactions between
the tires and the variable road surface can be modeled as a
second-order system. A sliding mode learning controller can then
be designed to drive both the sliding variable and the tracking
error between the steered front-wheel angle and the hand-wheel
reference angle to asymptotically converge to zero. The pro-
posed SMLC scheme exhibits many advantages over the existing
schemes, including: 1) no information about vehicle parameter
uncertainties and self-aligning torque variations is required for
controller design; and 2) the control algorithm is capable of effi-
ciently adjusting the closed-loop response based on the most recent
history of the closed-loop stability and ensuring a robust steering
performance. Both simulations and experiments are presented to
show the excellent steering performance and the effectiveness of
the proposed learning control methodology.
Index Terms—Learning control, Lipschitz-like condition,
robustness, sliding mode, steer-by-wire (SbW).
I. I NTRODUCTION
S TEER-BY-WIRE (SbW) systems have been considered to
serve the next generation of road vehicles for improv-
ing steering performance, enhancing vehicles’ maneuverability,
and also providing drivers with better comfort and proactive
safety. The distinct characteristics of SbW systems are as
follows: 1) The mechanical linkage used to connect the hand
wheel with the steered front wheels in conventional steering
systems is removed, 2) an ac or dc motor is adopted to steer
the front wheels so that the steered front-wheel angles closely
track the hand-wheel reference angle, and 3) another motor
is coupled with the hand-wheel shaft to provide a driver with
a feeling of the interactions between the front tires and road
surface.
To date, the SbW control systems have been intensively
studied in the automotive industry [1]–[10]. The bottleneck
for the design of high-quality SbW control systems is how
the effects of uncertain vehicle dynamics and highly nonlinear
self-aligning torque variations, due to different road conditions,
Manuscript received June 28, 2012; revised November 25, 2012, April 8,
2013, and June 4, 2013; accepted August 26, 2013. Date of publication
September 5, 2013; date of current version February 12, 2014. The review of
this paper was coordinated by Dr. S. Anwar.
The authors are with the Faculty of Engineering and Industrial Sciences,
Swinburne University of Technology, Hawthorn, Vic. 3122, Australia (e-mail:
mdo@swin.edu.au; zman@swin.edu.au; cishenzhang@swin.edu.au; haiwang@
swin.edu.au; feisiang2006@hotmail.com).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TVT.2013.2280459
0018-9545 © 2013 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
Authorized licensed use See http://www.ieee.org/publications_standards/publications/rights/index.html
limited to: Vietnam National University. Downloaded on April 28,2024 at 08:47:03 for
on the steering performance can be eliminated. In [12]–[15],
a number of control methods have been proposed for SbW
systems with the strategies that the controllers are designed
based on the estimated road surface conditions and chassis
sideslip angle. In fact, the good steering performance of using
these schemes for SbW systems can be achieved only when
accurate estimates of the road surface conditions and the chassis
sideslip angle can be obtained. In [15]–[19], a few adaptive
control techniques for SbW systems have been developed for
improving the steering performance. However, how the system
states, uncertain parameters, and unknown lateral forces can be
accurately online estimated under varying road environments to
ensure a robust steering performance is still an open issue.
Recently, sliding mode control (SMC) has been employed
in SbW control systems in [20]–[26]. It has been seen that
SbW systems equipped with the SMC are able to eliminate
the effects of SbW system uncertainties and unknown complex
road conditions on the steering performance by using the upper
and lower bound information of uncertainties. Although the
steering performance with the SMC strategy is better than
the performances of other existing control techniques, how
chattering in control signals should be eliminated and how the
constraints on the system uncertainties could be eased without
degenerating the robustness and convergence performance need
to be further studied.
In this paper, we will develop a new sliding mode learning
control (SMLC) scheme for SbW systems based on [27] and
[34]–[36] to improve the steering performance against the
uncertain system dynamics and the unknown road environ-
ment. It is worth noting that the proposed SMLC algorithm
in this paper is based on the concept of the “Lipschitz-like
condition” recently proposed in [27] and [34]. The Lipschitz-
like condition describes the continuity of the gradient of the
Lyapunov functions for a large class of systems. It states that a
continuous-time system is a Lipschitz-like system if the differ-
ence between the current value of the gradient of a Lyapunov
function and its most recent value is very small as the sampling
period is sufficiently small. As shown in [27] and [34]–[36],
the controller designs based on the Lipschitz-like condition
no longer require prior information about system uncertainties
since the uncertain system dynamics are all embedded in the
Lipschitz-like condition. It will be shown from the theoretical
discussions and experimental results in this paper that, unlike
the conventional SMC, the proposed SMLC is continuous in
the state space and no chatter occurs in the closed-loop system.
In addition, the analysis on the convergence and stability of the
closed-loop SbW system equipped with the proposed SMLC
algorithm will show that the new learning control is capable of
UTC more
frominformation.
IEEE Xplore. Restrictions apply.
DO et al.: ROBUST SLIDING MODE-BASED LEARNING CONTROL FOR SbW SYSTEMS IN MODERN VEHICLES 581

TABLE I
PARAMETERS OF THE SbW S YSTEM

where σ is the rack and pinion system’s gear ratio, and r is

a scale factor accounting for the conversion from the linear
motion of the rack to the rotation of the front wheels.
The rotation of the steering motor shaft is described by the
following dynamical equation [25], [28]:
Fig. 1. Steer-by-wire (SbW) system.
Jsm δ̈sm + Bsm δ̇sm + τ12 = τsm . (2)
driving the closed-loop dynamics from the unstable domain to
the stable domain and ensuring that both the sliding variable In addition, the rotation of the steered front wheels about their
and the tracking error can asymptotically converge to zero. vertical axes crossing the wheel centers can be described by
The remainder of this paper is organized as follows. In [14], [37]
Section II, the modeling of SbW systems and the learning Jsw δ̈f + Bsw δ̇f + τF + τe = τs (3)
structure of the proposed SMLC are formulated. In Section III,
the analysis on the asymptotic error convergence and stabil- where τF is the Coulomb friction defined as
ity of the closed-loop SbW systems is presented in detail. τF = Fs sign(δ̇f ) (4)
Numerical simulations and experimental results are presented
in Sections IV and V, respectively, to illustrate the advantages with Fs being the Coulomb friction constant. Using (1) and (2),
of the proposed control scheme. Section VI gives a conclusion we can express τs as
and discusses further work.
τs = k(τsm − Jsm δ̈sm − Bsm δ̇sm ) (5)

II. P ROBLEM F ORMULATION where k = σr. Substituting (5) into (3) leads to

A. Dynamics of SbW Systems Jsw δ̈f +Bsw δ̇f +Fs sign(δ̇f )+τe = k(τsm−Jsm δ̈sm −Bsm δ̇sm ).
(6)
Fig. 1 shows an SbW system, and the corresponding system
parameters are listed in Table I. By eliminating δ̇sm and δ̈sm in (6) with the aid of (1), we obtain
It is shown that the mechanical linkage used to connect Jsw δ̈f +Bsw δ̇f +Fs sign(δ̇f )+τe = k(τsm −kJsm δ̈f −kBsm δ̇f ).
the hand wheel with the steered front wheels in conventional (7)
steering systems has been substituted by two motors, i.e.,
the steering motor and the hand-wheel motor. The role of the Rearranging (7), we have
steering motor is to steer the front wheels and ensure that the Jeq δ̈f + Beq δ̇f + Fs sign(δ̇f ) + τe = τeq (8)
front-wheel angle can closely track the hand-wheel angle.
On the other hand, the hand-wheel motor provides a driver where
with reaction torque from the interactions between the vehicle Jeq = Jsw + k 2 Jsm
tires and road surface. Beq = Bsw + k 2 Bsm
For simplicity, we assume that the backlash between the rack
τeq = kτsm . (9)
and pinion gear teeth is zero. Thus, the following relationships
hold [11], [25]: Remark 2.1: It is shown in (8) that, although the SbW system
contains a few components such as the steered front wheels, the
δ̈f δ̇f δf 1 τ12 rack and pinion gearbox, and the steering motor, the integrated
= = = = (1)
δ̈sm δ̇sm δ sm σr τs SbW system can be modeled by a second-order differential
Authorized licensed use limited to: Vietnam National University. Downloaded on April 28,2024 at 08:47:03 UTC from IEEE Xplore. Restrictions apply.
582 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 63, NO. 2, FEBRUARY 2014

did 1
= (vq − Rs iq − ωe Ld id − ωe Lmd If d ) (11)
dt Lq
3
τsm = p [Lmd If d iq + (Ld − Lq )id iq ] (12)
2
where vd and vq are the d- and q-axis stator voltages, id and
iq are the d- and q-axis stator currents, Ld and Lq are the d-
and q-axis inductances, Rs and ωe are the stator resistance and
the rotor electrical speed, Lmd is the d-axis mutual inductance,
If d is the equivalent d-axis magnetizing current, and p is the
number of pole pairs, respectively.
If id is adjusted to zero (id ∼
= 0), (12) becomes
3
τsm = pψdm iq (13)
2
where ψdm = Lmd If d is the d-axis flux linkage due to the
Fig. 2. SbW system with the SMLC scheme.
permanent magnet.
By properly controlling id and iq in the current control loop
equation. Hence, it is possible to use some advanced control
of the motor control system, one can generate the actuation
techniques to design controllers to ensure a robust steering
torque required to drive the front wheels to closely track the
performance of road vehicles [25].
hand-wheel reference angle. In fact, however, there always exist
Remark 2.2: It is known that the variable gear ratio steering
disturbances and noises due to both the flux linkage perturba-
has been widely used in many modern vehicles [10], [31]–[33].
tion and the current control error, which may be, without loss
Although the SbW system in this paper has a fixed gear ratio, it
of generality, expressed as follows:
will be shown in the later discussions that the proposed learning
controller is capable of eliminating the impacts of the variable ∗
ψdm = ψdm + Δψdm (14)
gear ratio on the steering performance since the gear ratio σ
via the parameter k = σr in (5)–(8) has been embedded in the iq = i∗q + Δiq (15)
parameters of the equivalent second-order model. ∗
where ψdm and Δψdm denote the nominal value and the
Fig. 2 shows the overall control diagram of the SbW system
perturbation of the d-axis flux linkage, and i∗q and Δiq denote
with the SMLC, where DRs and DRh denote the steering
the reference and the control error of the q-axis current control
and hand-wheel motors associated with their servo drivers,
∗ ∗ loop, respectively.
respectively, receiving the torque references τsm and τhm from
Using (14) and (15), we rewrite (13) as
the SMLC and the proportional-derivative (PD) controller,
respectively. ∗
τsm = τsm + Δτsm (16)
Remark 2.3: In Fig. 2, the hand-wheel motor provides the
driver with the feeling of the interactions between the tires and where
the road surface. Thus, a PD controller can be designed for the
3  ∗ 
hand-wheel control system using the error signal between the Δτsm = p ψdm Δiq + i∗q Δψdm + Δiq Δψdm (17)
2
steered front-wheel angle and the hand-wheel reference angle
[9]. Additionally, in this paper, the estimated self-aligning represents the lumped torque perturbation and
torque τ̂e is fed back to the hand-wheel control loop to con-
∗ 3 ∗ ∗
tinuously provide the driver with the sensation of the reaction τsm = pψ i (18)
2 dm q
torque after the tracking error vanishes.
Remark 2.4: In this paper, two permanent-magnet syn- denotes the nominal torque reference signal for the steering
chronous motors (PMSMs) are used as the steering motor and motor provided by the SMLC.
the hand-wheel motor, respectively. Compared with the control Substituting (16) in (8) leads to
of the dc-motor-based SbW systems [2], [5]–[7], the control of
∗
the ac-motor-based SbW systems by its nature exhibits the more Jeq δ̈f + Beq δ̇f + τl = kτsm (19)
complex features in terms of designing control architectures
where lumped uncertainty τl can be expressed as
and handling system nonlinearities. The detailed study will be
shown in the later sections. τl = τe + Fs sign(δ̇f ) − kΔτsm . (20)

B. Steering AC Motor Torque Perturbation For further analysis, (19) can be expressed in the state space
form as follows:
The model of the PMSM is described as [29], [30]
did 1 ż1 = z2
= (vd − Rs id + ωe Lq iq ) (10)
dt Ld ż2 = a(t) + bu(t) (21)
Authorized licensed use limited to: Vietnam National University. Downloaded on April 28,2024 at 08:47:03 UTC from IEEE Xplore. Restrictions apply.
DO et al.: ROBUST SLIDING MODE-BASED LEARNING CONTROL FOR SbW SYSTEMS IN MODERN VEHICLES 583

∗
where z = [ z1 z2 ] T = [ δ f δ̇f ]T , u(t) = τsm , Inequality (31) implies that the difference between the value
b = (k/Jeq ), and of the gradient of the Lyapunov function and its approximation
is diminutive as the time interval τ is sufficiently small.
1
a(t) = − (Beq δ̇f + τl ). (22) Remark 2.7: It is shown in (27) and (28) that the control
Jeq signal u(t) is continuous for s(t) = 0. However, it is shown
Remark 2.5: Lumped uncertainty τl , including the self- in the Appendix that Δu(t) is also continuous at all the points
aligning torque, the Coulomb friction, and the steering motor of s(t) = 0. Thus, the proposed SMLC in (27) is continuous at
torque perturbation, is bounded but unknown. It will be shown every time instant in the state space.
that prior knowledge of the bounds of the lumped uncertainty In the next section, the asymptotic convergence and the
τl is not required in the proposed controller design. Such an stability analysis of the proposed SMLC are discussed in detail.
advantage makes the proposed control scheme exhibit strong
robustness against uncertainties. III. C ONVERGENCE A NALYSIS

C. Proposed Sliding Mode Learning Controller Theorem 3.1: Consider the SbW system in (19), if the con-
trol input in (27) with the correction term in (28) is used, the
The tracking error is defined as tracking error ε(t) defined in (23) will asymptotically converge
1 to zero.
ε(t) = δf (t) − θh (t) = δf (t) − θhr (t) (23) Proof: Considering a Lyapunov candidate function
Nθ
where parameter Nθ denotes the ratio between the hand-wheel 1
V (t) = s(t)2 (32)
angle θh and the steering angle δf . 2
A sliding variable is then defined as
the time derivative of V (t) can be expressed in terms of
s(t) = ε̇(t) + λε(t) (24) (25)–(28) as

where λ > 0. V̇ (t) = s(t)ṡ(t)

Thus, the time derivative of (24) is expressed as
= s(t) (f (t) + bu(t))
ṡ(t) = a(t)+bu(t)− θ̈hr (t)+λ(δ̇f (t) − θ̇hr (t)) = f (t) + bu(t) = s(t) [f (t) + bu(t − τ )] − s(t)bΔu(t)
(25)  
 
= V̇ (t, t − τ ) − αV̇ˆ (t − τ ) − β V̇ˆ (t − τ ) (33)
where
f (t) = a(t) − θ̈hr (t) + λ(δ̇f (t) − θ̇hr (t)). (26) where

In this paper, the sliding mode learning controller is proposed V̇ (t, t − τ ) = s(t) [f (t) + bu(t − τ )] . (34)
as follows:
Remark 3.1: Considering the continuity of V̇ (t), if the time
u(t) = u(t − τ ) − Δu(t) (27)
interval τ is chosen to be sufficiently small, we can always find
where correction term Δu(t) is defined as a positive number M  1 such that the following inequality is
    satisfied [27]:
ˆ (t − τ ) + β V̇ˆ (t − τ ) , for s(t) = 0
1
α V̇     1  ˆ 
Δu(t) = bs(t)   
0, for s(t) = 0. V̇ (t, t − τ ) − V̇ (t − τ ) < V̇ (t − τ ) (35)
M
(28)
τ is the time delay, V̇ˆ (t − τ ) is the estimated of the first- for V̇ (t, t − τ ) = 0, V̇ (t − τ ) = 0, and V̇ˆ (t − τ ) = 0. Inequal-
order derivative of the Lyapunov function candidate V (t − ity (35) is called the Lipschitz-like condition [27], [34]–[36]. It
τ ) = 0.5s(t − τ )2 , defined as describes that, for a large class of systems with the continuity
of their V̇ (t), the difference between the current value of the
V (t) − V (t − τ )
V̇ˆ (t − τ ) = (29) gradient of the Lyapunov function and its most recent value is
τ infinitesimal as the time interval τ is sufficiently small. The sig-
and the control parameters α and β to be determined. nificances of the Lipschitz-like condition for control system de-
Remark 2.6: The minimal value of the time delay τ is equal signs are in two folds: First, the uncertain system dynamics are
to the sampling period. If τ is sufficiently small, it is reasonable all embedded in the left-hand side of (35); thus, for controller
to assume that designs with the aid of the Lipschitz-like condition, knowledge
    of the upper and lower bounds of the system uncertainties is no
sign V̇ˆ (t − τ ) = sign V̇ (t − τ ) (30) longer required. Second, the Lipschitz-like condition provides
    a strategy for us to design a learning controller that recursively
 ˆ  ˆ 
V̇ (t − τ ) − V̇ (t − τ ) < μ V̇ (t − τ ) (31) updates the control signal based on the most recent stability
history of the closed-loop system. This point can be seen from
for V̇ˆ (t − τ ) = 0, V̇ (t − τ ) = 0, and 0 < μ  1. the following convergence proof of the closed-loop system.
Authorized licensed use limited to: Vietnam National University. Downloaded on April 28,2024 at 08:47:03 UTC from IEEE Xplore. Restrictions apply.
584 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 63, NO. 2, FEBRUARY 2014

It is expected that, in the next generation of control system Using (31) in (44), we can obtain
1    
designs, the Lipschitz-like condition will play a very essential
  
role to relax many constraints on uncertain system dynamics in V̇ (t) < V̇ˆ (t − τ ) + V̇ˆ (t − τ ) + μ V̇ˆ (t − τ )
existing robust control technologies. M
 
 
Adding the term V̇ (t − τ ) − V̇ (t − τ ) to (33), we can ex- − αV̇ˆ (t − τ ) − β V̇ˆ (t − τ )
press V̇ (t) as     
1    
V̇ (t) = V̇ (t, t − τ ) − V̇ (t − τ ) + V̇ (t − τ ) − αV̇ˆ (t − τ ) < − 1 + μ + α V̇ˆ (t − τ ) − β V̇ˆ (t − τ ) . (45)
  M
 
− β V̇ˆ (t − τ )
  With the chosen parameter α in (39), (45) satisfies
 
≤ V̇ (t, t − τ ) − V̇ (t − τ ) + V̇ (t − τ )
  V̇ (t) < 0. (46)
 
− αV̇ˆ (t − τ ) − β V̇ˆ (t − τ ) . (36)
Thus, the closed-loop SbW system with the SMLC in (27)
Substituting (35) into (36) yields is asymptotically stable, and the SMLC control law in (27)
ensures that both the sliding variable s(t) and the tracking error
1  ˆ 
 ˆ

ˆ

 ε(t) asymptotically converge to zero.
V̇ (t) < V̇ (t−τ )+ V̇ (t−τ )−αV̇ (t−τ )−β V̇ (t−τ ) .
M
(37)
IV. N UMERICAL S IMULATIONS
1) For the case V̇ (t − τ ) > 0:
Rewriting (37), we have In this paper, we consider an electric vehicle with its bicycle
     model. The dynamics of the yaw motion of the vehicle is given
1    
V̇ (t) < V̇ (t − τ ) + − α V̇ˆ (t − τ ) − β V̇ˆ (t − τ ) . by [15], [37]
M −Cf −Cr Cr lr −Cf lf
(38) β̇ mVCG −1 + 2
mVCG β
Cf
mVCG
= −Cf lf2 −Cr lr2
+ C f lf δf
γ̇ Cr lr −Cf lf γ
If control parameter α is chosen such that lz lz VCG lz
(47)
1 1
<α<1− −μ (39)
M M where the vehicle body slip angle β at the center of gravity
(CG) and the yaw rate γ are the state variables, respectively;
(38) satisfies
the constant longitudinal velocity VCG is about 10 m/s at the
V̇ (t) < V̇ (t − τ ). (40) CG; the vehicle mass m is 2000 kg; the moment of inertia Iz
of the vehicle about the CG is 1300 kg · m2 ; the distances of
Inequality (40) indicates that the learning controller (27) always the front- and rear-wheel axles from the CG are lf = 1.2 m
makes the value of V̇ (t) decrease when V̇ (t − τ ) > 0. and lr = 1.05 m, respectively; and the front and rear cornering
Suppose that, at time t = t0 , V̇ (t0 ) = 0. Then, at time t = stiffness coefficients Cf and Cr are respectively chosen as
t0 + τ , (33) can be expressed as 
  4000 (snowy road), for t ≤ 15 s
  Cf = Cr = . (48)
V̇ (t0 + τ ) = V̇ (t0 + τ, t0 ) − αV̇ˆ (t0 ) − β V̇ˆ (t0 ) . (41) 12 000 (dry road), for t > 15 s

Please note that the dynamics of the yaw motion of the

If control parameter β is chosen to satisfy the following vehicle in (47) is derived under the assumption that the tire slip
condition: angle is less than 4◦ . Thus, the nonlinear self-aligning torque τe
   
 ˆ  ˆ  in the linear region can be approximated as follows [37]:
V̇ (t0 + τ, t0 ) − αV̇ (t0 ) < β V̇ (t0 ) (42)

γlf
(41) becomes τe = −Cf (tp + tm ) β + − δf (49)
VCG
V̇ (t0 + τ ) < 0. (43) where tp (= 0.023 m) and tm (= 0.016 m) are the pneumatic
and mechanical trails, respectively.
The discussions from (38) to (43) have shown that the The nominal parameters of the dynamical model of the
proposed learning controller in (27) is capable of reducing steered front wheels in (3) are chosen as Jsw = 3.8 kg · m2 ,
V̇ (t) from the positive value to the negative value, i.e., the Jsm = 0.0035 kg · m2 , Bsw = 10 Nms/rad, and Bsm =
closed-loop trajectory can be driven from the unstable region 0.018 Nms/rad; the parameters σ and r in (1) are set as σ = 3
(V̇ (t) > 0) to the stable region (V̇ (t) < 0). and r = 6; and the Coulomb friction constant in (4) is chosen
2) For the case V̇ (t − τ ) < 0: as Fs = 30 Nm.
One can rewrite (36) as The hand-wheel reference angle is generated by the follow-
1  ˆ 
 ˆ

ˆ

 ing function:
V̇ (t) < V̇ (t−τ )+ V̇ (t−τ )−αV̇ (t−τ )−β V̇ (t−τ ) .
M
(44) θhr (t) = 0.4 sin(0.7πt) rad. (50)
Authorized licensed use limited to: Vietnam National University. Downloaded on April 28,2024 at 08:47:03 UTC from IEEE Xplore. Restrictions apply.
DO et al.: ROBUST SLIDING MODE-BASED LEARNING CONTROL FOR SbW SYSTEMS IN MODERN VEHICLES 585

Fig. 5. Comparison of the tracking errors.

Fig. 3. Transient responses of SMLC. (a) Sliding variable. (b) Tracking
performance.
where μ = 0.001 and M = 100. It is clearly shown that the
inequalities in the Lipschitz-like condition in (31) and (35) are
both satisfied.
Next, the effectiveness of the proposed SMLC is illustrated
in comparison with the H-infinity control and the conventional
SMC. The H-infinity controller for the SbW system is designed
as follows [40]:

∗ Jeq Beq ε
τsm = δ̇f + θ̈hr − c1 ε̇ − c2 ε + L (51)
k Jeq ε̇

0
where L = − P ; control parameters ϕ, c1 , and c2 are set
ϕ2
to 0.1, 1, and 150, respectively; and matrix P is given by

25.2224 0.0103
P = . (52)
0.0103 0.1669

The conventional SMC with the boundary layer (BL-SMC)

for the SbW system is given by [25], [26]
1  
∗
Fig. 4. Validation of the Lipschitz-like condition. (a) Inequality (31), where τsm = sat(s) J eq (θ̈hr + λ|ε̇|) + B eq |δ̇f | + τ l (53)
μ = 0.001. (b) Inequality (35), where M = 100. k

The sampling period is chosen to be equal to the time delay
ΔT = τ = 0.001 s, and the initial value of the front-wheel
angle δf (0) = −0.1 rad To simulate the effects of the self-
aligning torque variations due to different road environments
on the steering performance, two different road conditions are
considered as in (48), where the snowy road surface is for
the first 15 s and the dry road surface is for the next 15 s,
respectively.
First, the control parameters in (28) are chosen as α = 0.95
and β = 0.395. The transient responses of the sliding variable
and the tracking performance in the first 5 s have been shown
in Fig. 3(a) and (b), respectively. It is shown that the asymptotic
convergence values of both the sliding variable and the tracking
error of the closed-loop SbW system with the proposed SMLC
have been achieved.
To further confirm the concept of the Lipschitz-like condition
in (31) and (35), both the left side and the right side of (31) and
(35) have been plotted out in Fig. 4(a) and (b), respectively,
Authorized licensed use limited to: Vietnam National University. Downloaded on April 28,2024 at 08:47:03 UTC from IEEE Xplore. Restrictions apply.
where λ = 12, and the sat(s) function is defined as

sign(s), for |s| ≥ 0.8
sat(s) = (54)
s
0.8 , for |s| < 0.8
and the upper bounds of τl , θ̈hr , Jeq , and Beq are chosen
as τl = 8 Nm, θ̈hr = 10 rad/s2 , J eq = 10 kg · m2 , and B eq =
30 Nms/rad, respectively.
Fig. 5 shows the comparison of the tracking errors between
the steered front-wheel angle and the hand-wheel reference
angle of the SbW system with the proposed SMLC (27), the
H-infinity control (51), and BL-SMC (53), respectively. It is
shown that the H-infinity controller is unable to handle the
large variation of the road surface condition; the BL-SMC,
however, has greatly improved the tracking performance and
behaves with strong robustness against the variation of the road
conditions. Furthermore, the proposed SMLC has demonstrated
the best tracking performance with the smallest tracking error.
586 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 63, NO. 2, FEBRUARY 2014

Fig. 6. Implemented SbW system.

V. E XPERIMENTAL R ESULTS
Fig. 6 shows the SbW platform in the Robotics and Mecha-
tronics Laboratory at Swinburne University of Technology. Two
ac motors (Mitsubishi HF-SP102) driven by the servo drivers
(Mitsubishi MR-J3-100A) are used as the steering motor and
the hand-wheel motor, respectively, in the SbW platform. A
gearhead of 10 : 1 ratio is adopted to amplify the steering motor
drive torque. Two angle sensors are installed to measure the
hand- and front-wheel angles, respectively.
The Advantech PCI-1716 multifunction card is interfaced to
the desktop computer for real-time control with the Real-Time
Windows Target toolbox in MATLAB/Simulink. The nominal
parameters of the motors and the rack and pinion gearbox of the
SbW platform are the same as the parameters in the simulation
section.
For the comparison of the steering performances with differ-
ent control techniques, we use the signal in (50) as the reference
signal for the steered front wheels to follow. The sampling
period is selected as equal as the time delay ΔT = τ = 0.001 s.
It has been noted that the dynamical model of the yaw motion
of the vehicle in the simulation section is derived in the linear
region of the nonlinear self-aligning torque τe . However, the
proposed SMLC has no constraint on the operating region of Fig. 7. (a) Steering performance (H-infinity control). (b) Tracking error
the nonlinear self-aligning torque τe . To test the robustness of (H-infinity control). (c) Control input (H-infinity control).
the SbW system against the changes of the self-aligning torque,
a voltage signal Vτe is input to the steering motor to produce
a nonlinear torque disturbance for modeling the self-aligning Fig. 8(a)–(c) shows the steering performance, tracking error,
torque τe [38], [39] and control input, respectively, with the BL-SMC (53). It is
shown that, although the steering performance with the BL-
 SMC is much better than that with the H-infinity controller,
η1 tanh(2δf ), for t ≤ 15 s the steering angle could not track the reference angle well after
V τe = (55)
η2 tanh(δf ), for t > 15 s the road surface condition varied. This is because the preset
value of the boundary layer parameter in (54) was not ad-
where η1 = 1 and η2 = 5.6 to ensure that the values of the self- justed for t > 15 s; the steady-state error was therefore largely
aligning torque τe in the first 15 s and in the second 15 s are increased.
different. Fig. 9(a)–(c) shows the experimental results of the steering
Fig. 7(a)–(c) shows the performance of the SbW system with performance, tracking error, and control input, respectively,
the H-infinity controller (51). It is shown that, after the road with the proposed SMLC (27) where the control parameters
surface condition changed (t > 15 s), the H-infinity controller are chosen as α = 0.5 and β = 0.1, respectively. It is shown
was unable to drive the front wheels to follow the reference that the steering performance has been significantly improved
signal well. with a very small tracking error, as compared with the steering
Authorized licensed use limited to: Vietnam National University. Downloaded on April 28,2024 at 08:47:03 UTC from IEEE Xplore. Restrictions apply.
DO et al.: ROBUST SLIDING MODE-BASED LEARNING CONTROL FOR SbW SYSTEMS IN MODERN VEHICLES 587

Fig. 8. (a) Steering performance (BL-SMC). (b) Tracking error (BL-SMC). Fig. 9. (a) Steering performance (SMLC). (b) Tracking error (SMLC).
(c) Control input (BL-SMC). (c) Control input (SMLC).

performance of the H-infinity control in Fig. 7(a)–(c) and where Jh = 0.08 kg · m2 , Bh = 0.15 Nms/rad, and the param-
the steering performance of the BL-SMC in Fig. 8(a)–(c), eter Nθ in (23) is set as Nθ = 12.
respectively. Such an excellent steering performance of the The PD controller is designed together with the estimated
SMLC is largely due to the excellent learning capability of self-aligning torque τ̂e as an input to drive the hand-wheel
the SMLC that has ensured that the control gain can be motor to generate torque to model the interactions between the
effectively adjusted in time as the road surface condition front tires and the road surface
changes. τ̂e
∗
Now, we consider the SMLC for the SbW system with τhm = kp ε + kd ε̇ + (57)
Nθ
the real reference signal generated by a driver with the hand
wheel in Fig. 6. The corresponding control parameters and where control parameters kp = 15 and kd = 2.
the experiment setting are the same as those in Fig. 9(a)–(c). The estimated self-aligning torque τ̂e is output by the hand-
The hand-wheel model is described by the following wheel motor with the corresponding voltage input given by
equation:
Vτ̂e = ξ tanh(δf ) (58)
Jh θ̈h + Bh θ̇h + τhm = τh (56) where parameter ξ = 1.2.
Authorized licensed use limited to: Vietnam National University. Downloaded on April 28,2024 at 08:47:03 UTC from IEEE Xplore. Restrictions apply.
588 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 63, NO. 2, FEBRUARY 2014

VI. C ONCLUSION
In this paper, a new SMLC technique has been developed
for SbW systems with uncertain dynamics and varying road
conditions. It is shown that the developed control scheme is
capable of learning the closed-loop dynamics from its history
and then driving both the sliding variable and the tracking error
to converge to zero asymptotically. Such a learning control
process ensures that strong robust steering performance can
be achieved. Both the numerical and experimental results have
further confirmed the excellent steering performance of the
proposed learning control methodology. Further work on the
sliding mode learning-based observers for diagnosis of SbW
systems is being investigated by the authors.

A PPENDIX
V ERIFICATION OF THE C ONTINUITY
OF THE P ROPOSED SMLC

When s(t) asymptotically converges to zero as time t → ∞,

both s(t) and s(t − τ ) are the infinitesimals of the same order.
We can therefore have the following relationship between s(t)
and s(t − τ ) as time t → ∞:
s(t)
lim =ρ (59)
s→0 s(t − τ )

where 0 < |ρ| ≤ 1. From (28), we have

lim Δu(t)
s(t)→0

1  ˆ  
 
= lim αV̇ (t−τ )+β V̇ˆ (t−τ )
s(t)→0 bs(t)

1 α 2  β 
= lim s (t)−s2 (t−τ ) + s2 (t)−s2 (t − τ )
s(t)→0 bs(t) 2τ 2τ
  
1 s2 (t)  s2 (t) 
= lim lim α s2 (t)− 2 + lim βs2 (t)− 2 
s(t)→0 2bτ s(t) s(t)→0 ρ s(t)→0 ρ
  
1 1  1
= lim lim αs2 (t) 1− 2 + lim βs2 (t)1− 2 
s(t)→0 2bτ s(t) s(t)→0 ρ s(t)→0 ρ
  
αs(t) 1 βs(t)  1
= lim 1− 2 + lim  1− 2  = 0. (60)
s(t)→0 2bτ ρ s(t)→0 2bτ ρ

Thus, the proposed control is continuous at s(t) = 0. Consider-

ing the discussions both here and in Theorem 3.1 in Section III,
we conclude that the proposed SMLC is continuous at every
time instant in the time domain.

Fig. 10. (a) Steering performance (SMLC). (b) Tracking error (SMLC).
(c) Front-wheel control input (SMLC). (d) Hand-wheel control input.
Fig. 10(a)–(d) shows the steering performance, tracking er-
ror, front-wheel control input, and hand-wheel control input,
respectively. It is shown that the proposed SMLC scheme has
exhibited excellent steering performance with strong robustness
against the varying road conditions.
Authorized licensed use limited to: Vietnam National University. Downloaded on April 28,2024 at 08:47:03 UTC from IEEE Xplore. Restrictions apply.
R EFERENCES
[1] T. J. Park, C. S. Han, and S. H. Lee, “Development of the electronic
control unit for the rack-actuating steer-by-wire using the hardware-in-
the-loop simulation system,” Mechatronics, vol. 15, no. 8, pp. 899–918,
Oct. 2005.
[2] R. Pastorino, M. A. Naya, J. A. Perez, and J. Cuadrado, “Geared PM
coreless motor modelling for driver’s force feedback in steer-by-wire
systems,” Mechatronics, vol. 21, no. 6, pp. 1043–1054, Sep. 2011.
[3] F. Bolourchi, J. D. Demerly, and S. Amberkar, “A control system method-
ology for steer by wire systems,” presented at the SAE World Congress
Exhibition, Detroit, MI, USA, 2004, SAE Tech. Paper 2004-01-1106.
DO et al.: ROBUST SLIDING MODE-BASED LEARNING CONTROL FOR SbW SYSTEMS IN MODERN VEHICLES
[4] D. Odenthal, T. Bunte, H. D. Heitzer, and C. Eicker, “How to make steer-
by-wire feel like power steering,” in Proc. 15th Trienn. World Congr. Int.
Fed. Autom. Control, Barcelona, Spain, 2002, pp. 1–6.
[5] N. Matsunaga, J. Im, and S. Kawaji, “Control of steering-by-wire system
of electric vehicle using bilateral control designed by passivity approach,”
J. Syst. Des. Dyn., vol. 4, no. 1, pp. 50–60, Jan. 2010.
[6] B. H. Nguyen and J. H. Ryu, “Direct current measurement based steer-by-
wire systems for realistic driving feeling,” in Proc. IEEE Int. Symp. Ind.
Electron., 2009, pp. 1023–1028.
[7] N. Bajcinca, R. Cortesao, and M. Hauschild, “Robust control for steer-by-
wire vehicles,” Auton. Robots, vol. 19, no. 2, pp. 193–214, Sep. 2005.
[8] Z. Gao, J. Wang, and D. Wang, “Dynamic modeling and steering per-
formance analysis of active front steering system,” Proc. Eng., vol. 15,
pp. 1030–1035, 2011.
[9] M. Bertoluzzo, G. Buja, and R. Menis, “Control schemes for steer-
by-wire systems,” IEEE Ind. Electron. Mag., vol. 1, no. 1, pp. 20–27,
Spring 2007.
[10] S. W. Oh, H. C. Chae, S. C. Yun, and C. S. Han, “The design of a controller
for the steer-by-wire system,” JSME Int. J. Ser. C Mech. Syst., vol. 47,
no. 3, pp. 896–907, Sep. 2004.
[11] Y. Hori, “Future vehicle driven by electricity and control-research on
four-wheel-motored ‘UOT electric march II’,” IEEE Trans. Ind. Electron.,
vol. 51, no. 5, pp. 954–962, Oct. 2004.
[12] V. Krishnaswami and G. Rizzoni, “Vehicle steering system state estima-
tion using sliding mode observers,” in Proc. 34th IEEE Conf. Decision
Control, New Orleans, LA, USA, 1995, vol. 4, pp. 3391–3396.
[13] H. Ohara and T. Murakami, “A stability control by active angle control
of front-wheel in a vehicle system,” IEEE Trans. Ind. Electron., vol. 55,
no. 3, pp. 1277–1285, Mar. 2008.
[14] P. Yih, J. Ryu, and J. C. Gerdes, “Vehicle state estimation using steering
torque,” in Proc. Amer. Control Conf., Boston, MA, USA, 2004, vol. 3,
pp. 2116–2121.
[15] Y. Yamaguchi and T. Murakami, “Adaptive control for virtual steering
characteristics on electric vehicle using steer-by-wire system,” IEEE
Trans. Ind. Electron., vol. 56, no. 5, pp. 1585–1594, May 2009.
[16] T. Fukao, S. Miyasaka, K. Mori, N. Adachi, and K. Osuka, “Active
steering systems based on model reference adaptive nonlinear control,”
Veh. Syst. Dyn., vol. 42, no. 5, pp. 301–318, Dec. 2004.
[17] P. Seflur, D. Dawson, J. Chen, and J. Wagner, “A nonlinear tracking
controller for a haptic interface steer-by-wire systems,” Proc. 41st IEEE
Conf. Decision Control, vol. 3, pp. 3112–3117, 2002.
[18] A. E. Cetin, M. A. Adli, D. E. Barkana, and H. Kucuk, “Implementation
and development of an adaptive steering control system,” IEEE Trans.
Veh. Technol., vol. 59, no. 1, pp. 75–83, Jan. 2010.
[19] A. Roshanbin and M. Naraghi, “Vehicle integrated control—An adaptive
optimal approach to distribution of tire forces,” in Proc. IEEE ICNSC,
2008, pp. 885–890.
[20] J. Guldner, V. I. Utkin, and J. Ackermann, “A sliding mode control ap-
proach to automatic car steering,” in Proc. Amer. Control Conf., 1994,
vol. 2, pp. 1969–1973.
[21] J. Ackermann, J. Guldner, W. Sienel, R. Steinhauser, and V. I. Utkin,
“Linear and nonlinear controller design for robust automatic steer-
ing,” IEEE Trans. Control Syst. Technol., vol. 3, no. 1, pp. 132–143,
Mar. 1995.
[22] T. Kawabe, M. Nakazawa, I. Notsu, and Y. Watanabe, “A sliding mode
controller for wheel slip ratio control system,” Veh. Syst. Dyn., vol. 27,
no. 5/6, pp. 393–408, Jun. 1997.
[23] R. Kazemi and A. A Janbakhsh, “Nonlinear adaptive sliding mode con-
trol for vehicle handling improvement via steer-by-wire,” Int. J. Autom.
Technol., vol. 11, no. 3, pp. 345–354, Jun. 2010.
[24] F. Zou, D. Song, Q. Li, and B. Yuan, “A new intelligent technology of
steering-by-wire system by variable structure control with sliding mode,”
in Proc. Int. Joint Conf. Artif. Intell., 2009, pp. 857–860.
[25] H. Wang, H. Kong, Z. Man, M. T. Do, Z. Cao, and W. Shen, “Sliding
mode control for steer-by-wire systems with AC motors in road vehicles,”
IEEE Trans. Ind. Electron., vol. 61, no. 3, pp. 1596–1611, Mar. 2014.
[26] Z. Man, Robotics, 2nd ed. Singapore: Prentice-Hall, 2005.
[27] Z. Man, S. Khoo, X. Yu, and J. Jin, “A new sliding mode-based learning
control scheme,” in Proc. 6th IEEE ICIEA, 2011, pp. 1906–1911.
[28] S. Haggag, D. Alstrom, S. Cetinkunt, and A. Egelja, “Modeling, control,
and validation of an electro-hydraulic steer-by-wire system for articulated
vehicle applications,” IEEE/ASME Trans. Mechatronics, vol. 10, no. 6,
pp. 688–692, Dec. 2005.
[29] K. K. Shyu, C. K. Lai, Y. W. Tsai, and D. I. Yang, “A newly robust
controller design for the position control of permanent-magnet syn-
chronous motor,” IEEE Trans. Ind. Electron., vol. 49, no. 3, pp. 558–565,
Jun. 2002.
Authorized licensed use limited to: Vietnam National University. Downloaded on April 28,2024 at 08:47:03 UTC from IEEE Xplore. Restrictions apply.
589
[30] Y. X. Su, C. H. Zheng, and B. Y. Duan, “Automatic disturbances re-
jection controller for precise motion control of permanent-magnet syn-
chronous motors,” IEEE Trans. Ind. Electron., vol. 52, no. 3, pp. 814–823,
Jun. 2005.
[31] J. Kim, T. Kim, B. Min, S. Hwang, and H. Kim, “Mode control strategy for
a two-mode hybrid electric vehicle using electrically variable transmission
(EVT) and fixed-gear mode,” IEEE Trans. Veh. Technol., vol. 60, no. 3,
pp. 793–803, Mar. 2011.
[32] Y. Morita, K. Torii, N. Tsuchida, M. Iwasaki, H. Ukai, N. Matsui,
T. Hayashi, N. Ido, and H. Ishikawa, “Improvement of steering feel of
electric power steering system with variable gear transmission system
using decoupling control,” in Proc. 10th IEEE Int. Workshop Adv. Motion
Control, 2008, pp. 417–422.
[33] Y. Yao, “Vehicle steer-by-wire system control,” presented at the SAE
World Congress Exhibition, Detroit, MI, USA, 2006, SAE Tech. Paper
2006-01-1175.
[34] Z. Man, S. Khoo, X. Yu, C. Miao, J. Jin, and F. Tay, A New Design of
Sliding Mode Control Systems, Sliding Modes After the First Decade of
the 21st Century. Berlin, Germany: Springer-Verlag, 2012, pp. 151–167.
[35] M. T. Do, Z. Man, C. Zhang, and J. Zheng, “Sliding mode learning control
for nonminimum phase nonlinear systems,” in Proc. 8th IEEE ICIEA,
2013, pp. 1290–1295.
[36] M. T. Do, Z. Man, C. Zhang, and J. Jin, “A new sliding mode-based learn-
ing control for uncertain discrete-time systems,” in Proc. 12th ICARCV,
2012, pp. 741–746.
[37] P. Yih and J. C. Gerdes, “Modification of vehicle handling characteristics
via steer-by-wire,” IEEE Trans. Control Syst. Technol., vol. 13, no. 6,
pp. 965–976, Nov. 2005.
[38] A. Baviskar, J. R. Wagner, D. M. Dawson, D. Braganza, and P. Setlur, “An
adjustable steer-by-wire haptic-interface tracking controller for ground
vehicles,” IEEE Trans. Veh. Technol., vol. 58, no. 2, pp. 546–554,
Feb. 2009.
[39] P. Setlur, J. R. Wagner, D. M. Dawson, and D. Braganza, “A trajectory
tracking steer-by-wire control system for ground vehicles,” IEEE Trans.
Veh. Technol., vol. 55, no. 1, pp. 76–85, Jan. 2006.
[40] N. C. Shieh, P. C. Tung, and C. L. Lin, “Robust output tracking control of
a linear brushless DC motor with time-varying disturbances,” Proc. Inst.
Elect. Eng.—Elect. Power Appl., vol. 149, no. 1, pp. 39–45, Jan. 2002.
Manh Tuan Do received the B.E. degree (with
first-class honors) in electrical and automatic con-
trol engineering from Hanoi University of Science
and Technology, Hanoi, Vietnam, in 2008. He is
currently working toward the Ph.D. degree with
the Faculty of Engineering and Industrial Sciences,
Swinburne University of Technology, Hawthorn,
Australia.
From 2008 to 2010, he was a Laboratory Systems
Engineer with Intel Products Corporation. His re-
search interests include sliding mode control and its
applications, nonlinear control systems, robotics, and robust control of road
vehicles.
Zhihong Man received the B.E. degree from Shang-
hai Jiao Tong University, Shanghai, China, in 1982;
the M.Sc. degree from the Chinese Academy of
Sciences, Beijing, China, in 1987; and the Ph.D. de-
gree from The University of Melbourne, Melbourne,
Australia, in 1994.
From 1994 to 1996, he was a Lecturer with the
School of Engineering, Edith Cowan University,
Joondalup, Australia. From 1996 to 2001, he was a
Lecturer and then a Senior Lecturer with the School
of Engineering, University of Tasmania, Hobart,
Australia. From 2002 to 2007, he was an Associate Professor of computer
engineering with Nanyang Technological University, Singapore. From 2007 to
2008, he was a Professor and the Head of Electrical and Computer Systems
Engineering with Monash University Sunway Campus, Subang Jaya, Malaysia.
Since 2009, he has been with Swinburne University of Technology, Hawthorn,
Australia, as a Professor and the Head of Robotics and Mechatronics. His
research interests are in nonlinear control, signal processing, robotics, neural
networks, and vehicle dynamics and control.
590 IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 63, NO. 2, FEBRUARY 2014

Cishen Zhang received the B.Eng. degree in com-
puter engineering from Tsinghua University, Beijing,
China, in 1982 and the Ph.D. degree in electri-
cal engineering from The University of Newcastle,
Callaghan, Australia, in 1990.
During 1970–1978, he was an Electrician with
Changxindian (February Seven) Locomotive Man-
ufactory, Beijing, China. During 1983–1985, he
carried out research on control systems with Delft
University of Technology, Delft, The Netherlands.
In 1989–2002, he was with the Department of Elec-
trical and Electronic Engineering, The University of Melbourne, Melbourne,
Australia, as a Lecturer, Senior Lecturer, and Associate Professor, and during
2002–2010, he was with the School of Electrical and Electronic Engineering
and the School of Chemical and Biomedical Engineering, Nanyang Tech-
nological University, Singapore, as a Reader. Since November 2010, he has
been a Professor and the Head of Electrical and Electronic Engineering with
the Faculty of Engineering and Industrial Sciences, Swinburne University of
Technology, Hawthorn, Australia. His research interests include control, signal
processing, and medical imaging.
Fei Siang Tay received the B.E. degree with (first-
class honors) in electrical and computer systems
engineering from Monash University Sunway
Campus, Subang Jaya, Malaysia, in 2008. He is
currently working toward the Ph.D. degree with
the Faculty of Engineering and Industrial Sciences,
Swinburne University of Technology, Hawthorn,
Australia.
His research interests include automatic control,
robotics, fuzzy systems, neural networks, nonlin-
ear systems, applied mathematics, and dynamical
modeling.

Hai Wang was born in Hefei, China, in 1984. He

received the B.S. degree in electrical engineer-
ing from Hebei Polytechnic University, Tangshan,
China, in 2007 and the M.S. degree in electrical en-
gineering from Guizhou University, Guiyang, China,
in 2010. He is currently working toward the Ph.D.
degree with Swinburne University of Technology,
Hawthorn, Australia.
His current research interests include sliding mode
control with applications to robots and road vehi-
cles, neural networks, nonlinear systems, and vehicle
dynamics and control.

Authorized licensed use limited to: Vietnam National University. Downloaded on April 28,2024 at 08:47:03 UTC from IEEE Xplore. Restrictions apply.