Downloaded from SAE International by Eric Pierce, Wednesday, March 04, 2015

Performance Measurement of Vehicle Antilock Braking

Systems (ABS)

2015-01-0591
Published 04/14/2015

Karan R. Khanse and Eric Pierce

Virginia Tech

Michael Ng

University of Massachusetts

Saied Taheri
Virginia Tech

CITATION: Khanse, K., Pierce, E., Ng, M., and Taheri, S., "Performance Measurement of Vehicle Antilock Braking Systems (ABS),"
SAE Technical Paper 2015-01-0591, 2015, doi:10.4271/2015-01-0591.
Copyright 2015 SAE International

Abstract
Outdoor objective evaluations form an important part of both tire and
vehicle design process since they validate the design parameters
through actual tests and can provide insight into the functional
performances associated with the vehicle. Even with the industry
focused towards developing simulation models, their need cannot be
completely eliminated as they form the basis for approving the
performance predictions of any newly developed model. An objective
test was conducted to measure the ABS performance as part of
validation of a tire simulation design tool. A sample vehicle and a set
of tires were used to perform the tests- on a road with known profile.
These specific vehicle and tire sets were selected due to the
availability of the vehicle parameters, tire parameters and the ABS
control logic. A test matrix was generated based on the validation
requirements. The set of parameters to be measured were extracted
from this test matrix to understand the demand on instrumentation,
data acquisition and post processing of the acquired data. A cost
effective yet novel instrumentation circuit was designed to perform
the measurements. A brief description has been provided on the post
processing performed on the acquired data. The results of this test
have been discussed in details and the variations in results due to
changes in test conditions have been analyzed. These results were
finally used to validate the performance of the developed vehicle
simulation tool.

Introduction
Tractive forces that exist between the tires of a vehicle and the road,
determine to a large extent- the active safety of a vehicle. A number
of systems exist in vehicles today that aim at improving its active
safety. Antilock Braking System (ABS) is one such system. Reports
released by the National Highway Traffic Safety Administration
(NHTSA) show results with statistically significant decrease in
multi-vehicle crashes and fatal pedestrian crashes due to the inclusion

of ABS in vehicles [1]. ABS enhances safety by increasing the

traction potential of the vehicle. This traction is then utilized
effectively either to reduce the braking distance, or to facilitate
emergency steering maneuvers while braking. Designing an ABS to
fit a particular vehicle requires a considerable amount of
development. Slight changes in relevant vehicle and tire parameters
in the design phase require a corresponding tweak in the ABS control
logic. Currently, different manufacturers, who specialize in one of
these systems, develop the ABS, the tire or the vehicle. Due to this
reason, it would be mutually beneficial for all the industries involved
to work closely with each other in improving the system
performance. While the understanding and intent towards this exists,
what hinders the progress is the absence of a common evaluation
platform. The presence of such a common evaluation tool will not
only facilitate easier development, but will also reduce development
times. However, like for any given simulation tool, its success is
primarily determined by the accuracy of the predictions it makes as
compared to the real world performances. This requires considerable
effort in ensuring the accuracy of model predictions relative to the
effort spent on designing the tool itself.

Development of the ABS Simulation Tool

An ABS Simulation tool was developed to accurately predict the
straight line braking of a vehicle on a given road surface. As the
controllers for systems such as ABS keep advancing, there is an
increasing need to predict their behaviors with a high accuracy. Tires
play a critical role during Antilock Braking of a vehicle, as they
generate the necessary traction. Hence, from the point of view of a
tire company, such a tool would prove very useful as design factors
can be modified to help improve the ABS performance. Further, such
a simulation tool also proves useful from the point of view of an
automotive company. Tires are responsible for transferring forces
from the road to the vehicle during Anti-Lock Braking. With accurate
ABS models, and accurate tire models, the force interactions with the

Downloaded from SAE International by Eric Pierce, Wednesday, March 04, 2015

vehicle can be studied and design considerations for ABS can be

included. It is also important that the simulation tool not be
computationally intensive. Computationally intensive tools not only
increase the simulation times, but also increase the demand on the
machine requirements. Finite Element Model is one such example of
a computationally intensive tool. Discrete models are much less
computationally intense instead. The models that have been
developed and described here were designed keeping this in mind.
Further, individual simulation models should always be developed on
the platform, which is most suited to that particular type of model. In
this case, it was found that the models were best developed in
different interfaces. The models were then integrated by setting up an
effective communication path. Care was also taken that the
interactions between the different models in such a case does not
create compatibility issues. The following sections describe the tire,
ABS and vehicle models that were developed.

Rigid Ring Tire Model

The In-Plane Rigid Ring Dynamic tire model as explained by
Zegelaar [2] has been developed for simulations of straight line
braking [4]. This includes the calculations for various properties of
the tire including effective loading radius, forces in the contact patch
and the related equations of motion. In this model, the tire belt is
modeled as a rigid ring connected to the axle or the rim through
springs and dampers which represent the pressurized sidewall. The
ring is connected to the road through a contact model which is
represented as a transient slip model, as illustrated in the Figure 3. In
this model, the translational sidewall stiffness and damping are
represented by cbx and kbx which is the same as vertical sidewall
stiffness and damping cbz and kbz. cb and kb denote the rotational
stiffness and damping. Using axle motions, velocities, inputs from the
enveloping model, and brake torques, this model can calculate the
reaction forces on the axle by the tire, wheel slip as well as the
contact patch length.

Tire Model
A dynamic tire model was developed that is suitable for the high
frequency events such as ABS and short wavelength disturbances
from the road. The following two sections describe the models
developed to simulate the enveloping property and the tire itself,
within the MATLAB/Simulink environment.

Enveloping Model
If the profile of any given road is considered, it can be seen that it is
never flat, no matter how smooth the road might seem. This
unevenness in the road profile excites the tire in a complex manner.
Due to the built-in flexibility in the geometry of tires, they act as
geometric filters to the unevenness, smoothening out the sharp edges.
As a result, the effective road disturbance acting on the tire is
different from the actual road profile over which the tire rolls. This is
known as the enveloping behavior of a tire. Figure 1, taken from [2],
shows this behavior. The Tandem Elliptical Cam Model as described
by Schmeitz [3] is used to model this enveloping property of the tire.
This model assumes two elliptical cams in tandem, connected by rods
and constrained to move only in the vertical direction, rolling over
the uneven road surface. Figure 2 shows this Tandem Elliptical Cam
Model rolling over a cleat.

Figure 1. Enveloping Property

Figure 2. Tandem Elliptical Cam Model

Figure 3. Rigid Ring Tire Model

ABS Model
The ABS model developed in Simulink consists of two main
components: 1. Control algorithm and 2. Pressure model. Both these
components are equally important to achieve realistic and favorable
results, and hence, a balance needs to be ensured. The ABS unit
requires longitudinal slip and wheel acceleration as inputs from the
tire model. Based on these inputs, the ABS, with the included
pressure model, calculates the braking torque, May, which again
serves as an input to the tire model.
The operating algorithm of an ABS is characterized by the number
and the type of brake states utilized by the ABS to achieve optimal
control of slip and wheel acceleration. In this ABS control logic
designed by Ding et al [5], the primary states for the braking system
include: Initialize, Hold Brake Pressure, Increase Brake Pressure,
Fast Increase Brake Pressure, Step Increase Brake Pressure,
Decrease Brake Pressure, Step Decrease Brake Pressure and Exit
Braking. An ABS system deployed in a vehicle is designed to
control four wheels with a single controller. In a general scenario, it
might be too difficult to calculate control signals on all four wheels
in a single cycle. This is accomplished by the use of controller flags
where one wheel is controlled during a cycle and this is rotated after
each cycle. This way, all four wheels can be controlled at the
desired rate. The selection of brake states are made using a
state-flow based approach where the decisions are made based on a
set of rules and thresholds. Further, the brake pressure which is
regulated by means of electromagnetic valves installed in the
Hydraulic Control Unit (HCU) in the actual system can be regulated
in the simulation through three states: Pressure Increase, Pressure
Hold and Pressure Release. Increase in pressure and decrease in

Downloaded from SAE International by Eric Pierce, Wednesday, March 04, 2015

pressure follows an exponential form. This change in pressure is

then converted to brake-torque at the wheels using the relationships
between pressure and brake torque gain.

Vehicle Model
Vehicle Modeling was done using the commercially available
software package CarSim. CarSim is capable of predicting
vehicle performances in response to driver controls (e.g. braking) for
a given environment (e.g. road profile, coefficient of friction). The
models simulate physical tests to allow viewing the results much
more quickly and repeatedly as compared to physical testing. These
models are capable of replicating the system level behavior with high
fidelity. They contain the major effects that determine how the forces
are transferred through the suspension to the chassis. However, they
do not have the details of the linkage connections or structural
compliance. They do not include component details to determine
durability, fatigue or high frequency vibrations. CarSim can accept
Kinematics & Compliances data for a vehicle that can be obtained
from the Kinematics and Compliances (K&C) Rig.
K&C Rig is used to accurately establish the kinematic characteristics
of a vehicle's suspension and steering system geometries, and the
compliance characteristics of the suspension springs, anti-roll bars,
elastomeric bushes and component deformations. It provides the data
required for simulating vehicle performance in terms of ride, impact
isolation, steering and handling. This data combined with the
dimensional and inertia properties of the vehicle is sufficient to
predict the vehicle performances.

Determination of Test Details

As has been stated before, it is essential that the conditions for
the validation test be matched with those in the model. The
following describes details of how this was ensured for the
different aspects involved.

Tires
The test tire was selected based on the availability of the
experimental data and available tire parameters for modeling. The
specification of the tire used was 205/60 R15 91V. The tire data was
obtained from a trial version of TNO MF-Tool 6.1 [6] and was
recorded using standard test procedures and experiment measurement
conditions, as described in their documentation. Table 1 enlists some
of the tire parameters that were used as part of modeling.
Table 1. Tire Parameters

Vehicle
A sample vehicle was selected based on the availability of its data for
modeling. The Kinematics and Compliances data of the vehicle was
measured using a K&C Rig. The weight was measured using vehicle
load cells and the center-of-gravity (CG) was calculated using the
same load cells and lift-based measurement procedures. Other vehicle
dimensions were recorded using standard measuring instruments.
Table 2 describes the list of all parameters that were considered for
modeling, from the vehicle used for the test.
Table 2. Vehicle Parameters

Antilock Brakes
The ABS controller used for the model was that from the Antilock
Braking System installed on the sample vehicle. This controller was
previously developed at the same research lab and reference to its
development can be found in [5].

Road
The test was performed on the Smart Road which is a 2.2 mile
controlled-access testing facility at the Virginia Tech Transportation
Institute (VTTI). Broadly, there were two types of road surfaces
present on this road- concrete and asphalt. The concrete road was
seen to be smoother as compared to its rougher asphalt counterpart.
Profiles of these road surfaces were measured using vehicle mounted
road profiling system as part of a previous research project. The
resolution of the scanned profile was taken as 1 mm. A sample
2-dimensional section of this profiled road has been shown in Figure
4, which has a gradient. Further, the test was performed on the
straight section of the Smart Road, since the test involved straight
line braking trials. This straight section of the track is seen to have a
gradient of 2.3. This was seen as a deviation from what has been
considered in the models. In order to negate the effect of this gradient
on test results, it was decided that the test was to be performed uphill
as well as downhill for every test case. An assumption was made that
an average of the test results for braking distances for Uphill and
Downhill trials will be equal to the results obtained for the case if the
road were to be flat.

Downloaded from SAE International by Eric Pierce, Wednesday, March 04, 2015

GPS
A 5V active GPS antenna was used for distance and velocity
measurements. The GPS used was capable of measuring distance at
the rate of 100Hz, with an accuracy of 0.05% and a resolution of
1cm. Likewise, the velocity was also measured at the rate of 100Hz,
with an accuracy of 0.1km/h and a resolution of 0.01km/h. The GPS
antenna which has a magnetic base was mounted on the vehicle roof,
ensuring that atleast 7 satellites were locked during testing at all
times. Figure 5b shows the GPS antenna used.
Figure 4. Road Profile

Test Matrix
Before the test set-up could be designed, a test matrix was developed
to understand the expectations from the validation test. Table 3 shows
the developed test matrix. Using the individual parameters in the test
matrix, the instrumentation set-up was then designed which is
described in the next section.
Table 3. Test Matrix

VBox
The 3rd generation VBox was used for logging the test data from the
brake trigger and the GPS. As stated earlier, the VBox is capable of
recording the data at a rate of 100Hz. The logged data is directly
stored onto a compact flash card to easily transfer to a PC and analyze
using the VBox Tools software. While it simultaneously receives
signals on brake triggering, position as well as velocity, it is easy to
punch the brake trigger event to an exact location, at which point the
velocity is also known. Using this as one reference and the stop
velocity as another, the stopping distance can be calculated from the
position of the vehicle, as all are plotted against time simultaneously.
Figure 5c shows the VBox used.

Figure 5. (a) Brake Pedal Trigger, (b)GPS, (c)VBox

Part II
Inertial Measurement Unit (IMU)

Test Set-up
Primarily, two instrumentation circuits were designed for measuring
the parameters extracted from the test matrix. The first design, which
included the VBox, was designed from the point of view of
measuring the braking distance and velocity. The other, which
included the Inertial Measurement Unit, was designed from the
point of view of measuring the vehicle accelerations and rotational
rates. They are described in the following two parts:

The Polulu MinIMU-9 v3, a 3 axis- accelerometer, gyroscope and

magnetometer were used for the acceleration measurements during
the brake tests. The measurement range of accelerometer was selected
as 2g and the sensitivity as 0.061mg/LSB, considering vehicular
braking. Measurements for the roll, pitch, yaw, lateral acceleration
and vertical acceleration were also obtained from this IMU. However,
they were not considered for the current scenario of straight line
braking. Figure 6 shows the IMU used. It consists of 5 pins for its
connections, however only 4 of these are used for this application. (1)
SCL- clock line, (2)SDA- data line, (3)GND- grounding and (4)
VIN- power pin (to accept input voltage while using an external
power source, which in this case is 5V). During testing, this IMU was
mounted as close to the center-of-gravity as was practically possible
to get the correct measurements.

Part I
Brake Trigger
The brake pedal trigger was used to enable precise measurement of
brake pedal application during the brake test. The trigger events are
recorded at high frequencies of 100Hz. This high frequency provides
a maximum of 10ms gap between the pedal application and when it is
recorded. This time is then automatically included in the calculation
to add the extra distance covered in this short time span. Figure 5a
shows the brake pedal trigger that was used.

Figure 6. IMU [7]

Downloaded from SAE International by Eric Pierce, Wednesday, March 04, 2015

IMU Data Logger System

In the manner that the VBox was used to log data from the GPS and
brake trigger, a similar data logger system was necessary for logging
data from the IMU. An Arduino coupled with a Data Logger Shield
was used for this purpose. The following sections describe how this
system was designed and data logging was achieved:
Arduino
The Arduino that was used is a single board 8-bit micro-controller
that consists of an open source hardware board with multiple analog
and digital inputs along with the I2C serial bus. It is mainly used for
communication and is designed for receiving inputs from sensors.
This can then be coupled with another add-on module for recording
this data. The 4 pins for clock, data, ground and power from the IMU
are connected to the corresponding Arduino pins. The Arduino is
powered with the help of 6 AA batteries through the power jack to
generate the required power supply. The Arduino can be programmed
to accept the data from the IMU by defining directories along with its
time latencies i.e. the time delay between when the data is measured
to when it is recorded.
Data Logger Shield
The Data Logger Shield is essentially used to save the data on any
FAT16 or FAT32 formatted SD card. All of the recorded data are
time-stamped using the built-in Real Time Clock (RTC). The RTC
keeps the time going even when the Arduino is unplugged. It consists
of a chip, crystal and a lifetime battery back-up. This behaves as an
add-on attachment over the Arduino. A prototyping area is also
provided over this shield. It is in this area that the IMU is mounted
atop the stacking headers. Appropriate data writing can be ensured
with the help of the same Arduino program. Figure 7 shows the
IMU-Arduino-Data Logger Shield assembly that was used.
While portability of the IMU is an important factor, so is the
sampling rate. Replacing the 6 AA batteries with small DC power
supply could potentially improve the sampling rate from 100 Hz to
125 Hz. With the short duration tests that were planned for this case,
the voltage drop of the batteries was insufficient to slow the sampling
rate. Another potential improvement would be to create a 3D printed
case with variable external mounting options to allow faster IMU
placement with different test rigs. Further, a long-term improvement
would be to add an external switch to add a time-based trigger for
easier post-test data processing.
An important advantage of using this IMU- Arduino setup is that it
was developed for approximately a tenth of the cost of commercially
available units with comparable accuracy. Further, with the help of
this already developed circuit, the Arduino can be used for data
acquisition of any other parameter like steering angle measurements,
tire temperature measurements, etc. with slight modifications.

Figure 7. IMU mounted on the DataLogger Shield and Arduino

Testing
To evaluate the straight line antilock braking of the vehicle, the initial
velocity of the vehicle was considered to be 65km/h (40.3mi/h).
Starting at this velocity, the vehicle was brought to a complete stop
using ABS, simulating panic braking situations. As has been
mentioned earlier, the gradient of the road was taken into
consideration by performing the runs uphill and downhill, and then
averaging the two test runs. Every run was performed 3 times to
check for repeatability in the results. The vehicle was maintained at
the test weight throughout, with considerations of driver, co-driver
and gas in the vehicle. This was confirmed after the test when the
vehicle was weighed again. Standard testing protocols were followed
while testing. Quick checks were made after each run if the data was
being recorded appropriately. A test log was maintained to make
critical notes such as what test was performed on which run- along
with a time stamp, so as to make data processing easier. With two
different circuits recording data simultaneously, it was made sure that
data miss-match does not occur. Syncing the times of these two
clocks precisely with each other ensured this. It was made sure that
an experienced driver was performing the test. The most critical
aspects of driving were maintaining the initial speed up to the
required mark and to simulate the panic braking condition with a
completely depressed brake pedal. A post-test analysis showed that
the initial test speed error was maintained at lower than 1.1% of the
originally required speed for all test runs. This confirmed that the
variations due to human factors were kept to a minimum, so as not to
affect the test results in any manner. The tests were performed in
favorable environmental conditions. Wind speeds were ensured to be
lower than 10-15mph so as not to have a considerable effect on the
vehicle to affect its braking performance. Apart from the above
mentioned details, all standard safety procedures were followed while
performing these tests.

Results and Analysis

Using the procedure mentioned in the Testing Section, data was
collected for the test conditions outlined in the Test Matrix. This data
was processed using data processing algorithms that were developed
for this program. The following discusses the results obtained. The
simulation results presented here for comparison have been published
previously in [8].

Downloaded from SAE International by Eric Pierce, Wednesday, March 04, 2015

Braking Distance
Figure 8 shows a comparison of the braking distances for the
validation results vs simulation results:

Figure 9. Velocity Plot- comparison to simulation

Figure 8. Braking Distance- comparison to simulation

1.

Simulation comparisons have been presented for the two road

cases. As stated earlier, the uphill and downhill results have
been averaged out.

2.

For the flatter road, a braking distance of 19.47m was observed.

The simulation result of 19.59m lies within 0.6% of the
validation result.

3.

For the rougher road, a braking distance of 20.89m was

observed. The simulation result for this road profile was 21.35m.
This was within 2.2% of the validation result.

4.

For the simulation tool that was developed, a variation of 3-5%

was considered to be within tolerable limits considering current
similar commercially available tools as benchmark. This was
mainly due to the fact that the simulation tool in its current
form, considers a constant coefficient of friction throughout
the braking simulation. Further, temperature changes in the
tire contact patch have not been considered in the simulations.
These two factors are assumed to have significant effects on
the braking results predicted. With that, 0.6% and 2.2% were
considered to be close correlations. Model development for
the above mentioned factors is currently under progress, and a
closer correlation with the test results should be observed once
they are factored in.

5.

Upon closer observation, it is seen that the braking distance

for the rougher road is higher than that for the flatter road, for
both- simulation and test results. This shows that the rougher
road tends to set vibrations in the tires, which reduces the tire's
capability to perform well under ABS conditions.

Figure 10. Velocity Plot- Different test cases

1.

Figure 9 shows a comparison of the distribution of Velocity

vs Distance between the test and simulation. To avoid the
effect of roll-back (backward jerk given by the vehicle when
it stops completely) and initial speed variations from affecting
the braking distance results, measurements were considered
between two fixed speeds which would lie between the end
limits. In line with this, the stop speed for calculations was
considered as 7.2km/h. Braking distance was calculated as the
difference between the distance values at these two fixed speeds
from the plot.

2.

The plot shows the simulation result traces the test result
closely. A correlation coefficient of 0.9923 is observed between
the two plots. This is considered as a very highly correlated
data.

Velocity

Further, a second derivative with respect to time of this plotted

data gives a more accurate estimate of the correlation. This
correlation has been verified with the help of acceleration plots
in the next section.

Figures 9 and 10 show a comparison of the velocity distribution with

respect to distance for the different cases involved:
3.

Figure 10 shows a plot of vehicle velocities with respect to the

distance covered. This is the velocity distribution of the four

Downloaded from SAE International by Eric Pierce, Wednesday, March 04, 2015

cases of road profile and uphill/downhill combinations. As seen,

the four plots show a gradual reduction in velocities and they
occur in the increasing order of their braking distances. Nonintersection and moreover uniform increase in separation of
these plots (post initial period-due to slight difference in initial
speed) indicate consistency in the rate of speed reduction.

Acceleration

used to measure the longitudinal acceleration of the vehicle at

the CG with the help of the newly developed IMU-Arduino
circuit. This circuit was developed at a tenth of the cost of other
commercially available units that measure vehicle accelerations
with comparable accuracy. Further improvements have been
suggested in the circuit for improving the data recording
techniques. This will be implemented in the next set of
validation tests.
5.

Simulation results for the braking distance are observed to lie

within 2.2% of the validation results. This is seen to be well
below the expected result of 3-5% for the model with its current
capabilities.
Further, it is also expected that there will be a closer correlation
with the test results once the coefficient of friction and
temperature modules are added to the model.

6.

To better evaluate the correlation between the braking results,

its derivatives with respect to time were studied. Velocity and
acceleration plots were obtained. Both these plots showed close
correlation between simulation and test results. MFDD values
obtained from simulation were within 0.8% of that obtained
from the test. The acceleration plots also helped capture the
dynamics of the ABS.

7.

Thus, with the help of the designed setup, results obtained

for braking distance, velocity and acceleration with the given
correlations were able to validate the simulation tool.

Figure 11. Acceleration plot- comparison to simulation

1.

Figure 11 shows the comparison between the acceleration plots

obtained from the test and simulation results. This plot has been
truncated to the region that does not involve the development
and reduction of this deceleration in order to have detailed
views. Here, positive acceleration is considered for braking.

References
1.

Garrott, W. and Mazzae, E., An Overview of the National

Highway Traffic Safety Administration's Light Vehicle Antilock
Brake Systems Research Program, SAE Technical Paper 199901-1286, 1999, doi:10.4271/1999-01-1286.

2.

It is seen that the simulation result generally has the same trend
as the validation test result.

2.

3.

The Mean Fully Developed Deceleration (MFDD) value for

the test was seen to be 8.945m/s2, and for the simulation result
was seen to be 8.873m/s2. They lie within 0.8% of each other.

Zegelaar P., The dynamic response of tires to brake torque

variations and road unevenness. PhD. Thesis, TU Delft,1998

3.

4.

Fluctuations in the acceleration plots indicate jerks in the

longitudinal direction. These jerks are a result of the ABS
controller in operation, as it increases and decreases brakepressure depending on the slip at the wheels.

Schmeitz A., A semi-emperical three-dimensional model of the

pneumatic tire rolling over arbitrarily uneven road surfaces. PhD
Thesis, Delft University of Technology

4.

Sivaramakrishnan, S, Siramdasu, Y, and Taheri, S. A New

Design Tool for Tire Braking Performance Evaluations, ASME
International Design Engineering Technical Conferences &
Computers and Information in Engineering Conference, Buffalo,
New York, Aug., 2014

5.

Ding N., Wang W., Yu G., Zhang W., Xu X., Nenggen D.,
Weida W., Guizhen Y., and Wei Z., Research and Validation
of the adaptive control strategy for ABS based on experimental
knowledge, automotive engineering, vol 31, no. 1, 2009.

6.

TNO Automotive, MF-Tool 6.1 User Manual 2010

7.

Polulu Online Product Images- http://www.pololu.com/

product/2468/pictures

8.

Khanse Karan R., Siramdasu Yaswanth, Dr. Taheri Saied,

Development of a Simulink-CarSim Interaction Package for
ABS Simulation of Discrete Tire Models 2014 Tire Science
and Technology Journal

Conclusions
1.

Prior development of the simulation tool for predicting ABS

performances with a focused tire model gave rise to the need
for performing measurements on the actual vehicle as part of
validation on the model.

2.

A test was designed in such a way as to have maximum

correlation with the data available for simulations.

3.

The developed test matrix helped identify parameters of

interest for the objective test performed on the vehicle. These
parameters were then used to develop instrumentation circuits
for data acquisition.

4.

One circuit was used to measure the braking distance and

velocity distribution over distance travelled. The second was

Downloaded from SAE International by Eric Pierce, Wednesday, March 04, 2015

The Engineering Meetings Board has approved this paper for publication. It has successfully completed SAEs peer review process under the supervision of the session organizer. The process
requires a minimum of three (3) reviews by industry experts.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or
otherwise, without the prior written permission of SAE International.
Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE International. The author is solely responsible for the content of the paper.
ISSN 0148-7191
http://papers.sae.org/2015-01-0591