Downloaded from SAE International by University of British Columbia, Thursday, August 02, 2018

Development of Adaptive Traction Control System 2013-26-0085

Published on
9th-12th January
2013, SIAT, India
Chinmay Marathe
Indian Institute of Technology, Bombay, India
Ramanathan Annamalai
Tata Consultancy Services, India

U S Karle and K Prasanna Venkatesan

The Automotive Research Association of India (ARAI), India

Copyright © 2013 SAE International and Copyright © 2013 SIAT, India

ABSTRACT conditions like oil slippage, muddy patch on the roadside.

The basic governing factors are shown in Fig. 1.
Traction Control System (TRCS) has become a standard
feature for most of the vehicles due to safety considerations.
The system provides better drivability and acceleration
performance on low friction surfaces.

The TRCS typically tries to maintain slip value to an

optimum value to maximize traction force by modifying
engine torque and/or brake force intervention. However
the optimum slip value is not a constant value and varies
depending upon the road surface and tyre conditions. It is
challenging task to predict this value dynamically and adapt
TRCS under all driving situation.
Figure 1. Traction Governing Factors.
This paper presents an adaptive traction control system
which tries to operate at optimum slip point irrespective of
prior knowledge of road condition. The system continuously Problem when accelerating:
monitors vehicle dynamics parameters and corrects itself
 Spinning wheels prevent driver from accelerating.
to maximize available traction. The controller has been
developed using Matlab Simulink platform. The system  Spinning wheels cause loss of longitudinal and lateral
performance is validated in simulation (MIL technique) using grip.
a vehicle platform. The vehicle acceleration and slip control  Limiting factors.
results with TRCS and without TRCS for a typical passenger
car are presented.  μ (friction coefficient).
 Wheel load.
Keywords: Traction Control, Adaptive Traction, Optimum
Tire Slip, Acceleration Improvement.  Torque.
 Calculations:
INTRODUCTION  Traction Force Wheel (TFW) = Torque/r (1)
The TRCS is useful to avoid the excessive spinning of  Max Friction Force = μ * Wheel Load. (2)
driving wheels and loss of acceleration on low μ surfaces
during vehicle starting. It is also useful during driving when If Traction Force Wheel > Max Friction Force => Wheel
sudden negative change of μ is encountered on the road for Spin
 Downloaded from SAE International by University of British Columbia, Thursday, August 02, 2018

Symposium on International Automotive Technology 2013

Whenever driving force is much higher than the maximum of tyre and road surface combination (as shown in Fig. 4)
traction force, wheel starts spinning leading to excessive slip. which makes it more difficult.
This excessive slip causes further and further loss of traction
force. A typical longitudinal μ-slip relationship for a given
surface is given in Fig. 2.

Figure 4. μ Slip Curve for Various Surfaces.

Various techniques have been used to solve this issue.

There have been efforts to look at slip values in combination

μPeak : Peak coefficient related to static with wheel accelerations and adjust brake pressure
μSlip : Slip coefficient related to slip dynamically. [1]

Figure 2. Typical Longitudinal Slip Curve. Fuzzy Control technique using slip and acceleration based
rule map has been demonstrated by Pieter M. de. Koker [2].

Typically surface detection algorithms are also being tried

TRACTION CONTROL SYSTEM out where the slip is controlled with a built in surface-peak
Typically, TRCS (as shown in Fig. 3) monitor the wheel spin slip value tables.
and adjust the engine torque and/or brake torque to reduce The various surface detection methods as shown in Fig. 5
the driving force. are very well described and compared by Prof. Rajamani [3].

Figure 5. Road Friction Estimation Techniques.

All these systems involve extensive calculations. Also

Figure 3. Typical Traction Controller. methods like slip - slope detection depend upon fixed built-in
μ - slip curves. These curves shift depending on tyre type
and usage condition. Even if deterioration in acceleration
Typical Traction Controller is acceptable, the operation at higher slips will cause loss
However it is very difficult to decide the target slip and of lateral forces which can create problems in cornering. A
adjust the driving force and slip at optimum peak value simple adaptive traction control method to track the optimum
under all conditions. The peak slip value is also a function slip point is developed and tried out to avoid these issues.
2
 Downloaded from SAE International by University of British Columbia, Thursday, August 02, 2018

Symposium on International Automotive Technology 2013

CONCEPT Controller Development

If the vehicle is operating at slip less than the optimum slip The controller is developed using Matlab-Simulink platform.
the traction force will be less than the maximum possible The block schematic for one wheel is shown in Fig. 7.
force and also the vehicle acceleration will be correspondingly
less. At this point if the slip target is incremented, the vehicle
acceleration will increase. This process will continue till
the optimum slip point is reached. If slip target is further
incremented, there will be decrease in vehicle acceleration.

This phenomenon is used to achieve optimum slip point. As

shown in Fig. 6 the controller will produce a disturbance in
slip target by ∆ slip and look for change in longitudinal force
and vehicle acceleration. If change in vehicle acceleration
is positive, it will increment the slip target. As soon as the
vehicle acceleration stops increasing and start decreasing,
the controller will understand that it has reached peak point
Figure 7. Adaptive Traction Controller.
and will start decreasing the slip target. Thus controller will
continuously track the optimum slip point.
Traction control is achieved by a PID based slip controller.
Initially the slip target is set at somewhat higher value.
When traction control is enabled, and initial wheel flare is
under control, a predefined disturbance is generated in target
slip. The detection block then looks into delta slip change
versus delta acceleration change and decides to increment or
decrement the slip as per Table 1 . This loop is continuously
executed till traction control is enabled thereby always
tracking the optimum slip point.

Table 1. Delta Slip, Acceleration and Target Slip Change

Table.

Before peak slip

For +ve ΔS (S2-S1) ΔFx (Fx2-Fx1) is +ve
Change in vehicle acceleration is positive
After peak slip
For +ve ΔS (S2-S1) ΔFx (Fx2-Fx1) is -ve
Change in vehicle acceleration is negative
Figure 6. Slip Disturbance by ∆s Detects Current Slip
Position w.r.t. Peak Slip.

For detection of split μ surface, the disturbance generator is

IMPLEMENTATION phase shifted for left and right wheel and therefore response
Today model based development is an industry standard and of individual wheel is measured.
therefore the concept was tried out following standard V The cycle time of adaptation is small and vehicle speed
cycle development. change during this time is relatively small and does not
The following section describes controller development, affect the decision making.
simulation platform and testing/test results. The snapshot of Simulink diagram of the implemented logic
is shown in Fig. 8.

3
 Downloaded from SAE International by University of British Columbia, Thursday, August 02, 2018

Symposium on International Automotive Technology 2013

Figure 8. Adaptive Traction Control Algorithm

Implementation in Simulink.

Simulation Platform
The controller was tried out on a commercially available Figure 10. Plant Model Operation Window.
vehicle dynamics plant model. The plant model was
parameterized for a typical passenger car of 1460 kg using Testing
the available facility. The tires were simulated using magic
The different parameters of the controller were optimized
formula tires. The controller was interfaced to plant model
on Model in Loop (MIL) platform. The measure parameters
as shown in Fig. 9.
to be set are
1) Definition of delta slip and acceleration
2) Disturbance generator parameters.
3) Slip increment/decrement step size
4) PID parameters

Test Results
The controller performance was evaluated on different
maneuvers. The slip adaptation on low μ surface is
demonstrated in slip-time graph given in Fig. 11.

Figure 9. Adaptive Traction Control Interfaced with Plant

Model.

The vehicle dynamics plant model provides facility to

program the road surface, slope etc. It also displays basic
parameters like engine speed, vehicle speed, brake clutch,
accelerator pedal values, gear positions etc. Graphs of
selected parameters can also be seen. Typical layout of
vehicle dynamics plant model is shown in Fig. 10.
Figure 11. Slip-Time Graph 0.1 μ.

4
 Downloaded from SAE International by University of British Columbia, Thursday, August 02, 2018

Symposium on International Automotive Technology 2013

The vehicle speed–time graph on a low μ without traction Fig. 13 demonstrates the wheel spin control on μ transition
control, with traction control and with adaptive traction surface with adaptive traction as compared to without traction
control are compared in Fig. 12. control.

The controller has been evaluated on μ transition surface to

check the performance. It has been found that it is able to
detect the μ transition and adjust the target slip accordingly.
The results of slip adaptation on μ transition surface are
given in Fig. 14.

Figure 12. Comparison of Vehicle Speeds. Figure 14. Slip Time During μ Transition from 0.1μ to
0.5μ

CONCLUSIONS
An adaptive traction controller has been developed which
tracks the optimum slip point irrespective of prior information
about road surface, tire condition etc. The system has been
evaluated on different road surfaces in MIL platform. The
results demonstrate that the vehicle performance can be
improved with the adaptive traction controller.

The system uses relatively simple logic and avoids large

equations and can be implemented on standard embedded
platform.

The same concept can also be used for Antilock Braking

System (ABS).

REFERENCES
1. P i e t e r M d e K o k e r, J G o u w s , L . P r e t o r i u s . ,
Electrotechnical Conference, MELECON ‘96. 8th
Mediterranean, ISBN 0-7803-3109-5, 13-16 May 1996

2. Weida Wang, Lijuan Yuan, Wei Zhang, Hongming

Zou.,Computer, Mechatronics, Control and Electronic
Engineering (CMCE), 2010 International Conference,
ISBN: 978-1-4244-7957-3, 24-26 Aug. 2010
Figure 13. Vehicle Speeds With and Without Adaptive 3. R a j a m a n i . R , P i y a b o n g k a r n . N , L e w. J , Yi . K ,
Traction Control for μ Transition from 0.1μ to 0.5μ. Phanomchoeng.G, Control Systems, IEEE, ISSN: 1066-
033X, Aug. 2010

5
 Downloaded from SAE International by University of British Columbia, Thursday, August 02, 2018

Symposium on International Automotive Technology 2013

4. Thomas D. Gillespie, “Fundamentals of Vehicle DEFINITIONS/ABBREVIATIONS

Dynamics”, SAE, 1992
μ Surface Friction
5. Bosch, Driving Safety Systems, Robert Bosch GmbH,
1999 ABS Antilock Braking System

6. Bosch, Automotive Safety Handbook, Robert Bosch TRCS Traction Control System
GmbH, 2007 VS Vehicle Speed
7. CARMAKER, Reference Manual Version 3.0, IPG
Automotive GmbH.

CONTACT INFORMATION
Chinmay Marathe
B/4, Indradhanu Society,
Kothrud, Pune 411038
India
Email: [email protected]

ACKNOWLEDGEMENTS
This work was initiated as a BE project and continued
thereafter. The project was sponsored by Tata Consultancy
Services (TCS). Authors thank TCS management for
providing all the necessary support and training. All the
simulation testing was carried out at the Automotive
Research Association of India ARAI. Authors thank, Mr. A.
B. Komawar (Senior Deputy Director, ARAI) for providing
simulation facilities. Authors thank Mr. Ravindra G. Shah for
his extensive support in preparing the manuscript.

The Technical Paper Review Committee (TPRC) SIAT 2013 has approved this paper for publication. Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of
This paper is reviewed by a minimum of three (3) subject experts and follows SAE guidelines. SIAT 2013. The author is solely responsible for the content of the paper.
The Automotive Research Association of India
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or Survey No. 102, Vetal Hill, Off Paud Road, Kothrud,
transmitted, in any form by any means, electronic, mechanical photocopying, recording, or otherwise, PUNE 411038. Maharashtra, (INDIA)
without the prior written permission of SIAT 2013. Tel.: +91-20-3023 1111  Fax: +91-20-2543 4190
Email: [email protected]
www.araiindia.com
Printed in India