Downloaded from SAE International by University of Liverpool, Monday, September 10, 2018

SAE TECHNICAL
PAPER SERIES 2001-01-0596

Comparison of control methods for Electric

Vehicle Antilock Braking / Traction
Control Systems
P. Khatun, C. M. Bingham and P. H. Mellor
University of Sheffield

Reprinted From: Brake Technology, ABS/TCS, and Controlled Suspensions

(SP–1576)

SAE 2001 World Congress

Detroit, Michigan
March 5-8, 2001

400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A. Tel: (724) 776-4841 Fax: (724) 776-5760
 Downloaded from SAE International by University of Liverpool, Monday, September 10, 2018

The appearance of this ISSN code at the bottom of this page indicates SAE’s consent that copies of the
paper may be made for personal or internal use of specific clients. This consent is given on the condition,
however, that the copier pay a $7.00 per article copy fee through the Copyright Clearance Center, Inc.
Operations Center, 222 Rosewood Drive, Danvers, MA 01923 for copying beyond that permitted by Sec-
tions 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying such as
copying for general distribution, for advertising or promotional purposes, for creating new collective works,
or for resale.

SAE routinely stocks printed papers for a period of three years following date of publication. Direct your
orders to SAE Customer Sales and Satisfaction Department.

Quantity reprint rates can be obtained from the Customer Sales and Satisfaction Department.

To request permission to reprint a technical paper or permission to use copyrighted SAE publications in
other works, contact the SAE Publications Group.

All SAE papers, standards, and selected

books are abstracted and indexed in the
Global Mobility Database

No part of this publication may be reproduced in any form, in an electronic retrieval system or otherwise, without the prior written
permission of the publisher.

ISSN 0148-7191
Copyright 2001 Society of Automotive Engineers, Inc.

Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE. The author is solely
responsible for the content of the paper. A process is available by which discussions will be printed with the paper if it is published in
SAE Transactions. For permission to publish this paper in full or in part, contact the SAE Publications Group.

Persons wishing to submit papers to be considered for presentation or publication through SAE should send the manuscript or a 300
word abstract of a proposed manuscript to: Secretary, Engineering Meetings Board, SAE.

Printed in USA
 Downloaded from SAE International by University of Liverpool, Monday, September 10, 2018

2001-01-0596

Comparison of Control Methods for Electric Vehicle Antilock

Braking / Traction Control Systems

P. Khatun, C. M. Bingham and P. H. Mellor

University of Sheffield

Copyright © 2001 Society of Automotive Engineers, Inc

ABSTRACT INTRODUCTION

The alleviation of environmental problems associated Increasingly stringent regulations to reduce emissions
with personal, public and commercial transport in urban from automotive vehicles has led to a resurgence of
areas has become an important issue for both policy research activity to realise alternative electrically driven
makers and the automotive industry. Future legislation power-trains; the resulting localised control of wheel
in Europe and the USA is expected to introduce strict torque has allowed advanced, high-bandwidth anti-lock
limits in vehicle emissions, and both electric and hybrid braking/traction control (ABS/TC) strategies to be
vehicles are considered to be strong contenders for realised.
meeting low / zero emissions targets. As a result,
research into electrically driven powertrains, which have A major obstacle to the development of robust ABS/TC
similar performance attributes as ICE (Internal has always been the real-time estimation of wheel-slip
Combustion Engine) vehicles, has led to the vs. adhesion-coefficient characteristics for different tyre
development of electrically actuated wheel technologies, types and road surface conditions [1]. Recent studies
with increasing attention being focused on research into have investigated the application of observer/estimation
novel antilock braking / traction control (ABS / TCS) schemes to obtain real-time data indirectly, the extended
strategies. Kalman filter (EKF) receiving increasing attention [2].
However, published material to-date is primarily derived
This paper describes a comparison of traction control from simulation studies, with practical implementation
schemes using real - time observer based estimates of and validation of the techniques being rare. Current
µ-slip characteristics. In particular, a ‘bang-bang’ type commercial/passenger vehicles incorporating ABS/TC
strategy will be evaluated against fuzzy based control systems often employ look-up tables, which are based
schemes to facilitate torque production from the on experimental trials, and have been shown to provide
powertrain to produce optimal traction. adequate performance under many driving situations.
This type of technique however, is limited by the fixed
Simulation studies and experimental trials on a structure of the control strategy, and is often ‘de-tuned’ to
laboratory test facility and simulation studies on a accommodate worst-case scenarios, for example,
prototype electric vehicle will be used to evaluate and traction control in icy conditions with old tyres.
compare the response of the proposed techniques in real Consequently, a sub-optimal wheel-slip characteristic is
driving situations. imposed for most driving conditions.

Furthermore, the paper will demonstrate that the To-date, a feature of many TC/ABS’s for commercial IC-
application of observer techniques are appropriate for engine vehicles, is that the realisable performance is
the on-line determination of the peak adhesion often limited by the mechanical bandwidth of the active
coefficient for different tyre-road surface conditions, and actuation systems (typically 20Hz for ABS). However,
the fuzzy based techniques offer substantial potential for with the advent of electrically powered vehicles into the
optimal control of wheel traction. marketplace, the means for controlling drive-torque at
much greater bandwidths will be possible, and hence,
some of the difficulties attributed to mechanical systems
will be inherently addressed in this paper.
 Downloaded from SAE International by University of Liverpool, Monday, September 10, 2018

TRACTION CONTROL / ANTILOCK BRAKING in slip will reduce traction, and consequently induce an
SYSTEMS unstable acceleration of the wheel until the drive torque
is reduced.
The tractive force between a tyre and the road surface
(Figure 1) is proportional to the normal load, Fz, the The objective of an antilock braking system (ABS) is to
constant of proportionality being termed the adhesion manipulate the tractive force applied to the driven wheels
coefficient, µ. The adhesion coefficient µ, is the ratio of in order to limit the slip, σ, between the road surface and
tire brake force at the tire road interface and the normal the tyre, and consequently only operate within the stable
load acting on the tire, i.e: region of the µ-σ characteristic. A major obstacle in the
FL practical design of ABS control schemes is the
µ= (1) determination in real-time of µ-slip characteristics, i.e.
FZ µmax and σ. For example, Table 1 shows typical values of
slip required to obtain the maximum adhesion coefficient,
for various road conditions.
Max. adhesion Optimum
Road condition coefficient σ)
slip (σ
µmax)
(µ
Dry Road 0.85 0.35
Slippery / Wet Road 0.4 0.2
Snowy / Icy Road 0.2 0.1
Table 1. Max. values of µ for various road surfaces [4].
Figure 1. Wheel model [3]. To overcome this problem, many manufacturers employ
a slip-limiting control scheme to account for ‘worst-case’
The value of µ is highly dependent on tyre characteristics conditions, typically icy roads. Although this constitutes
(compound, wear, ageing etc) and road surface the safest criteria to adopt for design purposes, it can
conditions (dry, icy, gravel, tarmac etc), although it can impose conservative limits on the traction available
be primarily regarded as a function of the relative slip, σ, under less severe conditions [5].
between the two contacting surfaces. By definition slip
is the normalised difference between the wheel speed A second obstacle often encountered by designers of
and the contact point: TC/ABS schemes, is the availability of appropriate test-
tracks/skid-pans or other experimental rolling road
v − Rω facilities to evaluate proposed algorithms and provide
σ= repeatable conditions for comparative studies.
v (2)
Experimental Test Facility
where v is the vehicle velocity, R is the effective radius of
the driven wheel, and ω is the angular velocity of the In order to address these issues, a low cost experimental
test bench is used to facilitate the design and preliminary
wheel. Typical µ-σ characteristics are shown in Figure 2.
evaluation of TC/ABS schemes. The test facility
C o n tr o l R e g io n illustrated in Figure 3, comprises of an electric traction
1 .0
µ m ax
drive connected to a standard three-phase induction
Adhesion coefficient (µ)

0 .8 machine, which is used to mimic the road load.

D ry R o a d
0 .6

0 .4
W et R oad
0 .2
Snow y R oad

λ0 0 .3 0 .6 0 .9
S lip (σ )

Figure 2. Example µ-slip characteristics for various road Figure 3. DSP based experimental test facilities.
conditions [4].
The experimental electric traction drive (Figure 4) is a
It is clear from Figure 2 that increasing slip can increase 6kW continuous, 12kW peak rated brushless permanent
the tractive force between the tyre and road surface by magnet motor supplied via a power controller from a 72V
virtue of an increase in µ. However, once the peak of the high performance Nickel Cadmium traction battery . The
characteristic is encountered, µmax, any further increase
 Downloaded from SAE International by University of Liverpool, Monday, September 10, 2018

power electronic controller is constructed using three

independent MOSFET / IGBT H-bridges [6] and is -10 Pushover
capable of full four-quadrant operation. The electrical -8 torque

Torque (Nm)
machine and power converter has a combined efficiency
-6 60V
of 92% at rated power. In the prototype electric vehicle,
the traction motor is connected to the wheels via a step- -4 50V
down gearstage. However, for the proposed test-bench, -2
40V
the brushless permanent magnet motor is directly 0
coupled to a 2.2kW, 50Hz, 3-phase, 6-pole, induction
2
machine, the angular velocity of the connecting shaft
being monitored via an optical shaft encoder interface. 4
The induction motor is connected to the utility supply via 6
a variac, which allows the supply voltage to be adjusted. 1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 -0.6 -0.8 -1.0
To facilitate the implementation of the developed control P.U.Slip
schemes, a custom state-of-the-art, DSP hardware
platform has been designed, based on the TMS320C31 Figure 5. Induction machine torque-slip characteristics.
floating-point processor, to integrate the system control
This procedure can continue up to the pushover torque
algorithms with sensor interfaces, man-machine
limit of the induction machine. A subsequent increase in
interface’s, monitoring, and supervisory tasks. A
applied torque will then further decrease the slip and
foreground/background distribution of 'non-time-critical' force the system into an unstable operating regime
and ‘time-critical’ functions has been employed to ensure where the induction machine is not able to oppose the
the controller remains a high integrity process. applied torque, and a runaway condition is encountered
with the brushless motor accelerating uncontrollably.
This is analogous to the classical µ-σ characteristics
encountered in traction control systems when the slip at
µmax is exceeded. An algorithm which automatically
controls the torque developed by the brushless PM motor
to maintain the induction machine in the ‘stable-slip’
region of operation, is therefore also likely to be a good
candidate to control the torque applied to the wheels of
an electric vehicle to maintain tractive force in the stable
region of the µ-σ characteristic. The torque-slip
characteristic can be obtained from the standard
equivalent circuit of an induction machine:

3V 2 R 2 / s
τ ind = (3)
Figure 4 . All – Electric Racing Vehicle. ω sync (( R1 + R2 / s) 2 + ( x1 + x2 ) 2 )
The key principle of operation of the proposed test facility where:V is the applied voltage, R1 is the stator
for TC/ABS investigations is based on the resistance, R2 is the referred rotor resistance, s is the
correspondence between µ-σ characteristics of tyre/road slip, ωsync is the synchronous speed, X1 is the stator
interactions Figure 2, and the torque-slip characteristics reactance, X2 is the referred rotor reactance.
of the induction machine when operated as a generator,
Figure 5.
The magnitude and form of the torque-slip
characteristics is dependent upon the supply voltage and
The brushless PM motor acts as the prime-mover, and impedance, and by adjusting these, a variety of tyre/road
drives the induction machine into the generator region of conditions can be readily simulated. It is noted that, due
operation, thereby inducing a negative slip with respect to the unsymmetrical nature of induction machine
to synchronous speed (1000rpm in this case). As the characteristics, different parameters for exciting the
torque demand to the PM machine is raised further, the induction machine are required to simulate ABS’s,
slip becomes increasingly negative, and the induction compared with TC systems.
machine develops an opposing torque, eventually
resulting in a stable steady-state condition.
Simulation model of traction control test facility

To demonstrate the attributes of the proposed test facility

for design and assessment of traction control algorithms
 Downloaded from SAE International by University of Liverpool, Monday, September 10, 2018

for electric vehicles, a simulation model of the test facility the pushover point and into the unstable torque-slip
as been employed, as illustrated in Figure 6. region, and hence, allowing the slip (and angular
velocity) to increase uncontrollably. The results from the
experimental test facility, Figure 8, confirm this dynamic
behaviour.
0.2

slip (p.u)
-0.2

-0.4

-0.6 Unstable
Figure 6. Simulation model of traction control test facility. -0.8
0 0.2 0.4 0.6 0.8 1 1.2 1.4
The load imposed on the brushless drive train is time(sec)
determined by the torque-slip characteristics of the
induction machine, the parameters of the induction Figure 8.a. Slip
machine being varied to simulate alternative µ-σ road
2500
conditions. For example, to simulate a dry road
condition, the induction motor has 60V excitation,
2000
providing a pushover torque capability of 9Nm. This
simulates the adhesion characteristics of a dry road on

speed (rpm)
1500
the primary drivetrain, when referred back through the
gearbox. The dynamic response of the system for a 1000
demanded drive train shaft torque of 12Nm is shown in
Figure 7. 500

0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
time(sec)

Figure 8.b. Speed of primary drive shaft

Figure 8. a & b. Experimental results from laboratory
test facility without traction control.

SIMULATION AND EXPERIMENTAL RESULTS

FROM TRACTION CONTROL TEST FACILITY
Figure 7.a. Slip
Initially, implementation of the "bang-bang" traction
control scheme to limit the slip to -0.1 p.u.is carried out,
resulting in a transient response which is shown in Figure
9a & b, where it is clear from the simulations that the
induction motor is now operating in the stable region of
the characteristic curve. Here, the controller is
accommodating ‘worst-case’ (icy road) conditions.

Figure 7.b.Torque on drivetrain

Figure 7.a.&b. Simulation results for road condition with
no traction control.
By consulting Figure 7, it can be seen that the demanded
torque is sufficient to force the induction machine past Figure 9.a.Slip
 Downloaded from SAE International by University of Liverpool, Monday, September 10, 2018

Discrete-time traction control scheme

To obtain optimal traction between the tyre and road

surface, a requirement is to possess prior knowledge of
both the maximum value of adhesion coefficient, µmax,
and the slip at which it occurs, σµmax. A traction control
scheme would then prevent the slip between the road
surface and the tyre from exceeding σµmax. Clearly, both
µmax and σµmax are highly dependent on tyre and road
surface conditions, and obtaining real-time data is a
primary source of difficulty for practical controllers. With
regard to the proposed experimental test facility, this
problem is analogous to determining the pushover torque
of the induction machine and the slip at which it occurs,
without prior knowledge of the excitation voltage.
Figure 9.b. Speed of primary drive shaft
A state-observer technique to estimate the load/road-
Figure 9 a & b. Simulated results with slip control adhesion applied to the PM brushless-dc drive train, and,
scheme implemented. by knowledge of the slip, a straightforward method of
obtaining optimal traction without prior knowledge of µmax
Experimental results on the proposed test rig (Figure 10a and σµmax is employed. The desired wheel torque is
& b) show the induction motor operating in the stable allowed to be transmitted to the wheel shaft until the
region, which is analogous to limiting the operation of a traction controller detects that the slip has exceeded
traction control scheme to the stable region of the µ-σ σµmax, at which time, the demanded torque is set to zero.
curve for icy road conditions, and confirm the simulation This has the effect of allowing the slip to reduce and re-
predictions. enter the stable region of the µ-σ characteristic. Having
0.2 re-entered the stable region, the desired torque is again
allowed to be transmitted to the wheel shaft. This simple
0 ‘bang-bang’ type of operation is particularly suited to
electric traction control, since the bandwidth of
-0.2
slip (p.u)

electromagnetic torque production is relatively high

-0.4
compared to mechanical actuation, and the modulation
of torque is ‘smoothed’ at the vehicle wheel due to
-0.6 inertia. The controller determines whether operation is
taking place in the stable or unstable region by
-0.8 calculating the rate-of-change of load/adhesion with
0 0.2 0.4 0.6 0.8 1 1.2 1.4
respect to slip i.e.
time(sec)

Figure 10.a. Slip æ dµ ö

ç
sgn ç dt ÷ = sgn æ dµ ö = ì+ 1 operation in stable region
÷ ç ÷ í
ç dσ dt ÷ è dσ ø î− 1 operation in unstable region
2500 è ø

2000

Practically, the time derivatives of µ and σ are required to

speed (rpm)

1500
be low-pass filtered to remove high frequency noise,
1000 however, since the inertia of the wheel and the maximum
torque that can be applied to the wheel shaft is known a-
500
priori, appropriate filtering does not unduly affect
0
performance of the controller.
0 0.2 0.4 0.6 0.8 1 1.2 1.4
time(sec) Under the assumption that the electromagnetic time-
constant of torque production in the power electronic
Figure 10.b. Speed of drivetrain inverter is relatively small compared with the mechanical
Figure 10.a & b. Experimental results with traction time-constant, the dynamics of the drive train can be
control scheme implemented. modelled as a first order differential equation:

• kt 1
ω= Tdem − Tl (4)
J J
 Downloaded from SAE International by University of Liverpool, Monday, September 10, 2018

where ω is angular velocity, kt is the motor torque

constant, J is the inertia, Tdem is the demanded torque to
the drive train, and Tl is the load/adhesion imposed on
the drive motor. Discretizing, and defining xT=-Tl, the
resulting discrete-time dynamic equations of the power
train become:

é x (n + 1) ù é1 − 1 / J ù é x(n) ù ék t / J ù
ê ú=ê ê ú+ Tdem (n)
ë xT (n + 1)û ë0 1 úû ë xT (n)û êë 0 úû

é x ( n) ù
ω = [1 0]ê ú (5)
ë xT (n)û

Whilst this does not constitute a controllable system,

since there is no independent control of xT, it is
Figure 12.a. Simulation results for 40V.
observable, and is appropriate for state-observation.
Hence, real-time estimation of Tl or equivalently µ, is
possible. For the experimental test facility, Tl represents
the load imposed by the induction motor, whilst on the
electric racing vehicle, it represents the adhesion
between the road surface and the tyre, referred to the
drive motor. The resulting traction control scheme is
illustrated in Figure 11,

Figure 12.b. Simulation results at 60V.

Also, by implementing the proposed discrete-time

observer based optimal traction control scheme, it is
possible to allow operation upto, but not beyond, the
stable region of the µ-slip curve (Figure 13). It can be
seen that µmax is –0.35 (p.u.) which represents a dry road
Figure 11. Discrete time traction control scheme
surface on a classical µ−σ characteristic.
where F defines the dynamics of the observer, and G
determines the poles (z=0.951,z=0.949) of the error
convergence dynamics between the actual and observed Measured
-4
states. The sampling period has been fixed at T=1×10
seconds.
Predicted
Initial simulation studies of the discrete time based
traction control scheme, show that the control of phase
voltage of the induction machine only affects the
magnitude of the torque-slip characteristic, with the value
of slip at pushover remaining unchanged. Hence, by
varying the phase voltage between 40V and 60V, it is
possible to obtain different observed load torque
characteristics, which are analogous to various adhesion
coefficients for dry and icy road conditions as shown on
Figure 12a and b. However, for both of these conditions, Figure 13. Discrete-time based ‘optimal’ bang-bang
the peak slip remains the same. traction control scheme
 Downloaded from SAE International by University of Liverpool, Monday, September 10, 2018

To simulate the dynamic µ-slip characteristic over a studies have shown that fuzzy systems are part of a
wider range of road conditions, again the variation of class of universal approximators of continuous functions.
phase voltage is employed. Hence, it is possible to The key advantage of fuzzy control is that there are
observe the discretised load torque (or µ) over a time many instances where TRUE and FALSE or ON and
period of 12sec, whilst the phase voltage is varied OFF fail to describe a given situation. These cases
between 55V and 48V (Figure 14). From the require a sliding scale where variables can be measured
measurement it can be seen that the incorporation of as PARTLY ON or MOSTLY TRUE and PARTLY FALSE
discrete-time observation technique allows the dynamic [8].
µ-slip characteristics to be obtained without a-priori
Traditional set theory is based on bivalent logic where an
knowledge of µmax.
object is either a member of a set or it is not. However,
with fuzzy logic, an object can be a member of multiple
599 sets with a different degree of membership in each set. A
577 degree of membership in a set is based on a scale from
555
Dry road 0 to 1; 1 being complete membership and 0 being no
Voltage (v)

533 membership. Hence, by applying fuzzy logic control to

51
this discrete time traction control system, an output (the
torque demand) is calculated based on the amount of
499 Wet road membership the input signals (observed load torque and
477 slip) has in the configured fuzzy sets. Therefore, it is
455 required that the system inputs go through three major
0 transformations before becoming a system output.
Observed Load Torque (Nm)

1 Figure 15 shows the overall fuzzy logic based traction

2 control scheme, which includes the transformation
3 process of fuzzification, fuzzy rule association and de-
4
fuzzification [8].

5
6
7
0 2 4 6 8 10 12
time(sec)

Figure 14. Measured results of observed load torque.

Fuzzy Logic based traction control scheme

In recent years fuzzy logic techniques have been applied

to a wide range of systems. Many electronic control
systems in the automotive industry such as automatic
transmissions, engine control and Antilock Braking
System (ABS) are currently being pursued in the United
States. These electronically controlled automotive
systems realise superior characteristics through the use
of fuzzy logic based control rather than traditional control
algorithms [7].

Fuzzy logic allows highly non-linear or mathematically Figure 15. Fuzzy Logic based traction control scheme
complex systems to be modelled in a reliable and
efficient manner. Moreover, fuzzy logic deals well with In order to establish the two appropriate input fuzzy
uncertain and noisy data. These characteristics offer membership functions and the rule base, a thorough
fuzzy control as a potentially important tool for the investigation of the ABS braking process has been
development of robust traction control systems. undertaken. In this case the rate of change of observed
load torque and the rate of change of slip were chosen.
One of the strengths of fuzzy systems is their ability to The two inputs are fuzzified into membership functions
express a confidence in reasoning results. Recent as shown on Figure 16a and 16b.
 Downloaded from SAE International by University of Liverpool, Monday, September 10, 2018

Although gaussian functions were initially used to For the testing, the Fuzzy Logic based Traction Control
investigate the ABS braking process, trials have shown scheme has two inputs (slip and load torque) to the
that functions based on polysigmoid curves provide controller and one output (torque demand), with 4
enhanced performance characteristics. possible rules.

Rate of change of slip

negative positive ps ns

Rate of change of
p z pb
n pb z

observ
Figure 17. Rule map for Fuzzy Traction controller.
Input variable: Load torque
All four rules are used for the controller and are
determined from the fuzzy rule map shown in Figure 17.
Figure 16.a. Membership functions for observed load The 4 rules cover each state of input combinations
torque. during the braking process. The output of the fuzzy
controller is the torque demand (or in the case of a
vehicle the tractive force). The fuzzy membership
functions for the output, torque demand, is shown on
negative small positive small Figure 18.

zero positive big

Input variable: Slip

Figure 16.b. Membership function for slip.

For the slip, the universe of discourse was chosen Output variable: Output Torque
between 0 and 100%, since during the operation of an
Figure 18. Membership function for Induction machine
induction machine at 0 slip the system is stable and at
torque demand.
100% the system is unstable. This is analogous to the
braking process of a vehicle, where a wheel can rotate
freely with slip at 0 or be locked at slip equal to 1. The membership functions, shown in Figure 16, were
created by experimentally testing the relationship
between slip and torque of the induction machine. A
Both inputs, slip and observed load torque, are fuzzified
fuzzy controller consists of fuzzification, rulebase and
into two membership functions. A smaller number of
defuzzification procedures. Fuzzification, that is, creating
membership functions causes less sensitive control, but
fuzzy inputs from crisp inputs, is accomplished using the
better computation speed, and vice versa. Therefore it is
membership functions in Figure 16a and 16b. The fuzzy
very important to understand and test the influence of the
output is generated using the rulebase, shown in the
number of membership functions, as there is always a
rulemap in Figure 17. A crisp output is generated from
compromise between execution speed and sensivity of
the fuzzy output by defuzzification, using the membership
the fuzzy controller.
functions shown on Figure 18. This is the basic fuzzy
control algorithm for a traction control system. By
"tuning" the membership functions the controller can be
optimised.
 Downloaded from SAE International by University of Liverpool, Monday, September 10, 2018

Implementing the fuzzy control scheme, with an Vehicle speed:

experimental setup analogous to a vehicle travelling on a
1 t
ò F − CV
2
dry road surface, Figure 19 shows the simulated V= t dt (7)
performance of the system. In comparison with limited M 0

‘bang-bang’ slip control, it can be seen that the fuzzy

scheme automatically adapts the slip control algorithm, Tyre model is described as:
identifies the unstable region of the torque-slip curve and F t = Nµ (σ ) (8)
reduces the slip. Eventually the slip stabilises at –0.25 Slip is defined as:
p.u. which is similar to the µ-slip characteristic of a dry
road surface(where the peak slip is between –0.25 and V − rω
σ= (9)
–0.35 p.u.). rω

where J is the inertia, Td is the machine torque, Ft is the

tractive force, r is the wheel radius, M is the mass, C is
the drag coefficient, V is the vehicle speed, µ is the
adhesion coefficient, ω is wheel speed and N is normal
mg
force ( m is mass, g is gravity).
4
From the simulation results (Figure 21), it can be seen
that by limiting the slip at –0.1 p.u., the vehicle system
model confirms the similarities in dynamic behaviour as
the traction control test facility results shown previously.

The vehicle speed for the simulation model was

measured using the tractive force acquired from the tyre
Figure 19. Simulation results for road conditions with model; the value of µ being obtained from a look-up
Fuzzy Logic Control. table for µ-slip characteristics. However, in the case of
measured data from the prototype electric vehicle, the
front wheels will provide the measurements for wheel
SIMULATION STUDIES OF ALL-ELECTRIC speed and the rotational speed of the rear axle will
RACING VEHICLE TRACTION CONTROL provide the vehicle speed; and any change between the
SYSTEM front and rear wheels will indicate slip.

To fully explore the operation of the proposed traction

control schemes, the all-electric racing vehicle was
chosen as a suitable platform, since it could be easily
reconfigured and safely tested.

Initially the vehicle performance was analysed using a

Matlab system model (Figure 20).
Torque Td
dem and Drive, m otors, ω
W heel
G ears
Slip
V Calculation

Vehicle
Figure 21. Matlab electric vehicle simulation result with
slip control at –0.1 p.u.

Ft
Tyre m odel CONCLUSION

Figure 20.Matlab system model of electric vehicle. An analysis of different traction control schemes is
presented in this paper. Simulation studies and
Where wheel speed is defined as: experimental trials on the laboratory test facility and
simulation studies based on the prototype electric vehicle
1 t
ω=
J ò0
Td − Ft r dt (6) were used to evaluate the response of the proposed
techniques in real driving situations. The paper
confirmed that application of observer techniques are
 Downloaded from SAE International by University of Liverpool, Monday, September 10, 2018

appropriate for on-line determination of peak adhesion • Drag coefficient, c 0.5

coefficient for different tyre-road surface conditions, and
• Gear ratio, 5:1
fuzzy logic based control strategies offer substantial
potential for optimal traction control of driven wheels.
REFERENCE

FUTURE WORK
1. Harned, J.L., Johnston, L.E., Scharpf, G.,”Measurement of
The following areas have been highlighted for future tire brake force characteristic as related to wheel
investigation: slip(antilock)control system”,SAE Trans,Vol78,
• Implementation of novel intelligent control schemes No.69214,1969.
to provide automatic adaption of slip control 2. Ray,L,R.,”Non linear system and tire force estimation for
algorithms on the test facility. advanced vehicle control”,IEEE Trans on Control Systems
Technology,Vol3, No1, 1995.
• Implementation of control schemes on the prototype
high-performance racing vehicle. 3. Klien,R., “Antilock Braking System and vehicle speed
estimation using Fuzzy Logic”,FuzzyTECH Application
The results of these investigations will be published in paper,http//www.fuzzytech.com/,1999.
due course. 4. Constantin Von Altrock., “Fuzzy Logic in Automotive
Engineering”, Embedded Systems Conference, 1996.
5. Robinson,B.J,Riley,B.S,”A study of various car antilock
braking system”,Research report 340, Department of
ACKOWLEGEMENT Transport and Road Research Laboratory, U.K.,1991.
6. Mellor,P.H, Schofield,N., Hor,P.J.,”Brushless permanent
th
The authors would like to acknowledge the UK magnet drives for all-electric racing karts”,Proc.29 Int.
Engineering and Physical Sciences Research Council for Symposium on Automotive Technology Automation,1996.
providing a research studentship to Ms.Khatun. 7. Elting,D.,Fennich,M.,Kowalczyk,R.,Hellenthal,B.,”Fuzzy
Antilock Brake System Solution”, Application note,
http://developer.intel.com/,1999.
CONTACT 8. De Koker,P.M., Gouwes,J., Pretorius,L., “Fuzzy Control
Algorithm for Automotive Traction Control Systems”, IEEE
Trans on Industrial Applications in Power Systems,
Ms.Parvin Khatun can be reached through the Computer Science and Telecommunication, vol.1, 1996.
Department of Electronic & Electrical Engineering,
University of Sheffield, Mappin Street, Sheffield S1 3JD,
U.K., at +44 114 2225195, or email:
elp97pnk@sheffield.ac.uk.
Dr.C.M.Bingham can be reached at the University of
Sheffield on +44 114 2225849, or e-mail:
c.bingham@sheffield.ac.uk.
Prof.Phil Mellor can be reached through the Department
of Electronic & Electrical Engineering, University of
Bristol, Merchant Venturers Building, Woodland Road
Bristol BS8 1UB, U.K., at +44 117 954 5259, or e-mail:
P.H.Mellor@bristol.ac.uk.

APPENDIX

Vehicle parameters:

• Mass, m 160 kg
2
• Gravity, g 9.81m/s
• Machine Torque, Td 12Nm
• Wheel radius, r 0.139 m
2
• Referred wheel Inertia, Jw 0.8 kgm
2
• Machine Inertia, Jm 0.0.156 kgm