International Conference on Advances in Electrical and Mechanical Engineering (ICAEME'2012) 18-19, 2012 Phuket (Thailand)

Active Front Independent Steering System and

its Control for Road Vehicle with Understeer
Characteristics
A.K.W. Ahmed, V. Rawat, and R.B. Bhat

inner and outer wheels [5, 6]. Such a control strategy would
Abstract—Active steering systems improve vehicle handling by require an independently controllable steering system, referred
controlling the steer angle of the wheels depending on speed. Such to here as Active Independent Front Steering (AIFS).
controls, however, have limitations as they do not attempt to utilize Furthermore, criteria must be established to quantify the
tires’ force generating potential. The present study proposes Active
Independent Front Steering (AIFS) technique with independent
ability of tire to generate further force or determine their
control for each front wheel such that the tire workload for each saturation limit. The concept of tire workload [7, 8] defined as
steered wheel is equalized. A simple control scheme has been the ratio of normal load to resultant tire force can be an
developed for a vehicle with understeer handling characteristics effective measure for implementation of the AIFS system. The
which applies the corrective input primarily to the outer wheel. The present investigation is carried out to evaluate the concept of
results show that proposed concept of AIFS can provide AIFS with a control strategy applicable to a vehicle with
performances equal to that of a conventional active steering system
while tire work loads are balanced to maximize the performance
under steer characteristics. A simple Proportional Integral
limits. (PI) controller is utilized to generate corrective steering for the
inner and outer wheels such that the target is realized while
Keywords— Active independent steering, tire workload, attempt is made to equalize the tire workloads of both the
understeer vehicle, Vehicle handling, tires. Simulation results generated for AIFS for a rounded step
and sinusoidal lane change maneuvers are compared with
I. INTRODUCTION those of AFS and uncontrolled systems to illustrate the
effectiveness of the proposed concept.
T HE concept of Active Front Steering (AFS) system has
been around since the 90’s. It has gained new momentum
with the advancement of sensors and control systems. II. SYSTEM MODEL
Conventional AFS system [1, 2] modifies the driver command
A. Vehicle System
steering input by a corrective steering angle in order to realize
a target response for that command at any forward speed. The Handling study of conventional AFS system will typically
correction is however made to both the steered wheels that are employ a bicycle model of road vehicle without regard to
designed to closely follow Ackerman steering geometry. Such lateral load shift or tire’s saturation limit. To examine the
control strategy can provide adequate performance at low proposed AIFS, it is essential to develop a 4-wheel vehicle
speed maneuvers [3, 4] until the dynamic load shift from the handling model that includes tire non-linearity and vehicle roll
inner to the outer wheel become significant [5]. During a high dynamics. Using the vehicle coordinate system shown in
lateral acceleration maneuver, the inner tire with less normal Fig.1, the normal forces for each of the tires are:
load can generate significantly less cornering force than the
outer tire with much higher normal load. The AFS induced m a x h cg m a y h cg c
correction to both wheel will therefore tend to saturate the
Front, Right wheel : Fz.f .R = Wf − +
2L 2 Tf L
inner tire while the outer tire capacity remain unexploited [6].
m a x h cg m a y h cg c
In order to maximize the performance potentials of AFS, it is Front, Left wheel : Fz.f .L = Wf − −
necessary that different corrections are introduced for the 2L 2 Tf L (1)
m a x h cg m a y h cg b
A.K.W. Ahmed is with the Mechanical and Industrial Engineering, Rear, Right wheel : Fz.r.R = Wr + +
Concordia University, Montreal, Quebec, Canada H3G 1M8 (phone: 514-848- 2L 2 Tr L
2424 Ext. 7932; fax: 514-848-3175; e-mail: waiz@alcor.concordia.ca). m a x h cg m a y h cg b
V. Rawat was with Concordia University, Montreal, Quebec, Canada H3G Rear, Left wheel : Fz.r.L = Wr + −
1M8. He is now with development division of Tata Motors, Pune India (e- 2L 2 Tr L
mail: vaibhavrawat@gmail.com).
R.B. Bhat is with the Mechanical and Industrial Engineering, Concordia
University, Montreal, Quebec, Canada H3G 1M8 (e-mail:
rbhat@alcor.concordia.ca).

146
 International Conference on Advances in Electrical and Mechanical Engineering (ICAEME'2012) 18-19, 2012 Phuket (Thailand)

where δ L, R are the steering angles for left and right wheels.
The tire lateral forces (F yi ) and aligning moments (M i ) are
computed from well-established non-linear Pacejka’s Magic
Formula [9] tire model. The coefficients of the magic formula
used are those corresponding to a medium size car tire
provided in [9] and the slip angles for each of the four tires are
computed from [10]:

Fzr Fzf  b ⋅ Ω + Vy  −1 
c ⋅ Ω − Vy 
FzR FzL α f ⋅R = δ R − tan −1   ; α r ⋅R = tan  
V x + T f ⋅ Ω  V x + Tr ⋅ Ω  (3)
 b ⋅ Ω + Vy  −1 
c ⋅ Ω − Vy 
α f ⋅L = δ L − tan −1   ; α r⋅L = tan  
V x − T f ⋅ Ω  V x − Tr ⋅ Ω 

Detailed derivations for the vehicle and tire model are

presented in [5]. In this investigation, a tire’s ability to
Fig. 1 Coordinates to establish tire vertical load generate handling or cornering force is defined by tire
workload defined as:
For simplicity, the handling model is considered to be a
3DOF (longitudinal-x, lateral-y, and yaw-Ω) model as shown
Fy2 + Fx2
in Fig.2. Tire work-load = (4)
µ ⋅ Fz

For coefficient of friction μ= 1, the tire workload approaching

0.9 is considered to be the saturation limit for a tire when its
ability to generate further force is diminished.

B. Control System
Pro-Ackerman steering geometry is desirable for passenger
vehicles to minimize tire scrub and related noise and wear. On
the other hand, anti-Ackerman ratio as used in race cars [11]
has its merits where the tire capable of developing higher
lateral force is allowed to do so. The control strategy
proposed in this investigation utilizes Ackerman steering
geometry with active independent front steering (AIFS)
control. The control strategy will also allow anti-Ackerman
ratio for high speeds when necessary to realize the reference
yaw-rate, while optimizing the tire work-load.
Fig 2 Forces and moments acting on the vehicle
In order to realize a target response at any speed, an
understeer vehicle’s steering angle must be increased with
With reference to Fig 2, the final expressions for the
speed. Furthermore the corrective steering angle should be
equations of motion are:
primarily added to the outer wheel with lower workload. In

[ ]
the event the inner tire reaches saturation or the tire workload
m a x = − F y. f . R sin(δ R ) + F y. f . L sin(δ L ) approaches unity, the angle for the inner tire should be
(2) reduced while that of the outer increased further. The initial
[ ] [
m a y = F y. f . R cos(δ R ) + F y. f . L cos(δ L ) + F y.r . R + F y.r . L ] results presented in this paper are obtained for corrective
angle to the outer wheel only in order to demonstrate the
effectiveness of AIFS system. The controller thus must
 = − ∑ M  + F [
y . f . R {b cos(δ R ) − T f sin(δ R ) } ]
4
identify the outer wheel prior to any corrective action taken. A
Iz Ω  i
1  general AIFS control strategy is presented in Fig. 3 and the
[
+ Fy. f .L {b cos(δ L ) + T f sin(δ L ) } − c Fy.r .R + Fy.r .L ] present investigation considers only an understeer vehicle.

147
 International Conference on Advances in Electrical and Mechanical Engineering (ICAEME'2012) 18-19, 2012 Phuket (Thailand)

the controller corrected steering input for the inner and outer
wheels.

δst
Steering Steering Accelerations

Driver Model
Forces/Moments Velocities

Fy, Mz δst Equations of Motion

Steering Vx, Vy, Ω
Slip Angle
α
Velocities
Slip Angle
Slip Angle Model
Forces/Moments
Normal Load F a x, a y
z
Normal Load Acceleration
Tire Model
Magic Formula
Normal Load Model

Fig. 4 Four-Wheel vehicle Simulink model

The parameters used for the simulation of vehicle-

controller system are presented in Table I. A turning
simulation is carried out for driver steering command of 0.1
Fig. 3 Flow chart for AIFS control
radians (5.73 degrees) in the form of rounded step input over a
In this case, the corrective angle will be added to the outer period of 2.0 seconds. Simulations are carried out for
wheel based on the error between reference and actual conventional AFS, where the correction is added to both
measured yaw rate: wheels without altering the steering geometry, and AIFS
where the correction is added to the outer wheel only. The
∆Ω = Ω ref − Ω act (5) results in terms of vehicle yaw velocity for a constant forward
velocity of 15 m/s (54 km/h) are shown in Fig. 5.
where the reference yaw rate can be calculated from vehicle
and operating parameters of a neutrally steered vehicle as: TABLE I
V V ⋅ tan(δ st ) PHYSICAL PARAMETERS FOR FOUR-WHEEL VEHICLE MODEL
Ω ref = x = x (6)
Rref L Symbol Quantity Magnitude and Unit

A corrective steering angle factor is then obtained using a

steering gain factor: m mass of the vehicle 1530 Kg
Iz moment of inertia about z axis 3500 Kg.m2
∆δ st = K st ⋅ ∆Ω (7)
L Wheel base 2.8 m
Finally, the actively controlled AIFS steering correction b Distance of CG from front axle 1.3 m
c Distance of CG from rear axle 1.5 m
command for the outer wheel is established using:
hcg Height of CG from ground 0.4 m
Tf Half Track Width - front axle 0.7 m
δ c = k1 ∆δ st + k 2 ∫ ∆δ st (8) Tr Half Track Width - rear axle 0.7 m
k1 Controller proportional gain 4.0
k2 Controller integral gain 6.0
For the present investigation the gains are established by trial
and error and are kept constant. It must be pointed out that for
an optimal AIFS system, the gains must also be established in The results show that as expected for an understeer vehicle,
an adaptive manner along with different correction command the steady state yaw rate will be lower than the reference or
for the inner and outer wheels. ideal if uncontrolled. Hence the control system will require
adding to the steering command in order to realize the target
III. SIMULATION RESULTS response. The results in Fig. 5 show that both AFS and AIFS
Due to interdependency between, vehicle and tire models can equally realize the target response while AIFS correction
and control system, it was necessary to formulate a closed is done for the outer wheel only. The wheel steer angles
generated by the controller that lead to these responses are
loop simulation scheme as illustrated in Fig. 4, where δ st is
shown in Fig. 6.

148
 International Conference on Advances in Electrical and Mechanical Engineering (ICAEME'2012) 18-19, 2012 Phuket (Thailand)
0.6
(a)
0.5
Yaw Velocity, Ω (rad/s)
0.4 Ideal
No control
0.3
AFS
AIFS
0.2
0.1
0
0 5 10
Time (sec)
Fig. 5 Vehicle yaw rate for a step steering input
8
Steering input and Output, δ (deg)
6 The simulated yaw rates of the vehicle for AFS, AIFS, and
Driver steering commnad
4 AFS-Right, Outer
AFS-Left, Inner
AIFS-Right, Outer
2 AIFS-Left, Inner
0 magnitude of this deviation.
0 5 10
Time (sec)
Time (Sec)
Fig. 6 The steering input and angle generated at each wheel
The results show that the AIFS correction was achieved by
increasing the outer (right) wheel steer angle alone to 8
degrees, while for AFS both the the inner and outer were
increased above 7 degrees. The AFS by design maintained the
Ackerman steering ratio and the increase was necessary to
realize the target at this speed. However, due to lack of normal
load on the inner (left) wheel, the workload on the inner wheel
of AFS system was the highest as shown in Fig. 7. These
results showing workload for each tire during the maneuver
indicate that the inner of AFS has reached saturation while
that of the outer wheel is the lowest. On the other hand, for
AIFS, the workload of both tires is close to each other. With
independent controllability of AIFS system, it is possible for
the controller to reduce the steer angle of the inner (left) wheel
while increase that of the outer (right) wheel further to realize
the target. And in doing so, the workload will be equalized.
This will leave each of the tires a reserve to develop further
cornering or braking force without any of the tires reaching
saturation.
In order to examine the performance of the proposed
concept in a lane change maneuver, next set of results are
generated for a sinusoidal steering input of peak to peak
magnitude 0.03 rad (1.72 degrees) over a period of 2π sec.
The simulations are carried out for a fairly high constant
forward velocity of 25 m/sec. (90 km/h). The target, ideal or
the reference yaw rate for the vehicle is defined by equation
(6) corresponding to a neutrally steered vehicle configuration.
1
Tire Work Load, Compare
0.8
AFS-Right, Outer
0.6
AFS-Left, Inner
AIFS-Right, Outer
0.4 AIFS-Left, Inner
15
0.2
0
0 5 10 15
Time (sec)
Fig. 7 Comparison of tire workload for AFS and AIFS
without control are shown in Fig. 8(a) in comparison to the
ideal response. As the results show, both AFS and AIFS are
capable of providing the target response, whereas the
uncontrolled system will generate significantly lower yaw rate
at this speed unless the steering angle is increased by the
driver. The resulting trajectory or the path of the vehicle is
shown in Fig. 8(b) for the three systems which illustrate the
15 Similar to step input response, the AIFS realized the target
response in lane change by modifying the steer angle of the
outer wheel alone. However, for a sinusoidal input or a lane
change maneuver, the right wheel is the outer wheel only for
half the cycle after which the left wheel will become the outer
wheel. The steering angles generated by the AFS and AIFS in
order to realize the responses in Fig. 8 are shown in Fig. 9.
The results show that steering angle for AFS was increased
from driver command of 1.72 degrees peak to peak while the
increase was applied to both the left and right wheels. The
AIFS realized the same target response by applying totally
different strategy where the corrective angle was applied to
the right wheel for first half of the cycle while the left being
the inner wheel was kept at the level of driver command. For
the second half of the cycle, the right wheel was brought back
to the level of driver command while the required increase in
the cornering force was realized by the increase in the steering
angle of the left wheel. The results presented in Fig. 9 further
indicate that the switching between the inner and outer wheel
takes place prior to reaching the half cycle. It is attributed to
the attempt of controller to maintain target yaw rate and the
point during the cycle where no correction is required to
realize the target.
The performance measure in terms of tire work load of left
and right wheels of AFS and AIFS systems for the lane
change
149
 International Conference on Advances in Electrical and Mechanical Engineering (ICAEME'2012) 18-19, 2012 Phuket (Thailand)

0.4
(a)
Ideal AFS-Right
0.1

Tire Work Load, Compare

No-control AFS-Left
Yaw Velocity, Ω (rad/s)
0.2 AIFS-Right
AFS
0.05
AIFS AIFS-Left

0 0

-0.05
-0.2
-0.1
-0.4
0 2 4 6 8 0 2 4 6 8
Time (sec) Time (sec)
(b)
Fig. 10 Tire workload for a lane change maneuver
20

IV. CONCLUSION
Y Co-ordinate (m)

15
The concept of AIFS with appropriate controller can be
designed for an understeer vehicle to realize target response to
10 a handling command similar to conventional AFS system.
AIFS can however, enhance the performance limits of AFS by
5 increasing the tire’s capability as it tends to equalize the
workloads of the tires. In contrast to an understeer vehicle, an
0
oversteer vehicle will require a more complex control strategy
0 50 100 150 200 250 as the steer angle must be reduced with speed to realize target.
X Co-ordinate (m)
Fig. 8 Yaw rate and vehicle trajectory for lane change maneuver
REFERENCES
[1] Ackermann, J., Odenthal, D. and Bunte, T. “Advantages of Active
Driver steering commnad
Steering input and Output, δ (deg)

Steering for Vehicle Dynamics Control”; 32nd International Symposium

1
AFS-Right
AFS-Left
0.5
AIFS-Right
AIFS-Left
0 Safety”; Proceedings of the Institution of Mechanical Engineers Part D:
-0.5
-1
0 2 4 6
Time (sec)
Fig. 9 Steering angle generated by AFS and AIFS systems
maneuver is presented in Fig. 10. Although the tire work load
is low for this maneuver, it presents a comparative analysis for
the two active steering systems considered. Although the
steering angle of both left and right wheels of AFS system is
very similar (Fig. 9), there is a 30 % difference between the
peak tire workloads of the two wheels as shown in Fig. 10.
For the proposed AIFS system, the tire workloads of both
wheels remain close to each other throughout the cycle.
These results clearly show that in a severe lane change
maneuver, one of the wheels of AFS will approach saturation
while the other will have the least work load. For the same
maneuver, the AIFS will yield lower workload and hence will
be able to handle more severe maneuver than the AFS system.
on Automotive Technology and Automation, Vienna, Austria, 1999, pp.
263-270.
[2] Mokhiamar, O. and Abe, M. “Active Wheel Steering and Yaw Moment
Control Combination to Maximize Stability as well as Vehicle
Responsiveness during Quick Lane Change for Active Vehicle Handling
Journal of Automobile Engineering, 2002, Vol. 216 (2), pp. 115-124.
[3] Velardocchia, M., Morgando, A. and Sorniotti, A. “Four-Wheel-Steering
Control Strategy and its Integration with Vehicle Dynamics Control and
Active Roll Control”; SAE paper no. 2004-01-1061.
[4] J.Y. Zhang, J.W. Kim, K.B. Lee and Y.B. Kim"Development of an
active front steering (AFS) system with QFT control", Int. J of
Automotive technology, 2008, Vol.9, No.6, pp.695-702.
8 [5] Vaibhav Rawat, "Active Independent Front Steering for Yaw-Rate
Control and Tire Work-Load Equalization in Road Vehicles", MASc
thesis, 2007, Concordia University, Montreal, Canada.
[6] A.K.W. Ahmed, V. Rawat, R.B. Bhat, “Active front steering system for
road vehicle and its limitations” Accepted for Int. Science, Technology
and Engineering Conference, 2012,Dubai, UAE.
[7] Velardocchia, M., Morgando, A. and Sorniotti, A. “Four-Wheel-Steering
Control Strategy and its Integration with Vehicle Dynamics Control and
Active Roll Control”, 2004, SAE paper no. 2004-01-1061.
[8] Mokhiamar, O. and Abe, M. “Effects of Model Response on Model
Following Type of Combined Lateral Force and Yaw Moment Control
Performance for Active Vehicle Handling Safety”; JSAE Review, 2002,
Vol. 23 (4), pp. 473-480.
[9] Pacejka, H.B., Bakker, E., and Nyborg, L. “Tyre Modeling for Use in
Vehicle Dynamics Studies”; SAE paper no. 870421, 1987.
[10] Elbeheiry, E.M., Zeyada, Y.F. and Elaraby, M.E. “Handling Capabilities
of Vehicles in Emergencies Using Coordinated AFS and ARMC
Systems”; Vehicle System Dynamics, 2001, Vol. 35 (3), pp. 195-215.
[11] Gaffney III, E. F. and Salinas, A.R. "Introduction to Formula SAE
Suspension and Frame Design"; SAE paper no. 971584, 1997.

150