400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A.

Tel: (724) 776-4841 Fax: (724) 776-5760

SAE TECHNICAL
PAPER SERIES
2001-01-3347
Integrating Monte Carlo Simulation, Momentum-
Based Impact Modeling, and Restitution
Data to Analyze Crash Severity
Nathan A. Rose, Stephen J. Fenton and Christopher M. Hughes
Knott Laboratory, Inc.
Reprinted From: ATTCE 2001 Proceedings
Volume 1: Safety
(P-367)
Automotive & Transportation Technology
Congress & Exhibition
October 1- 3, 2001
Barcelona, Spain
The appearance of this ISSN code at the bottom of this page indicates SAEs 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 $8.00 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.
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 SAE International and Messe Dsseldorf.
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
All SAE papers, standards, and selected
books are abstracted and indexed in the
Global Mobility Database
2001-01-3347
Integrating Monte Carlo Simulation, Momentum-Based Impact
Modeling, and Restitution Data to Analyze Crash Severity
Nathan A. Rose, Stephen J. Fenton and Christopher M. Hughes
Knott Laboratory, Inc.
ABSTRACT
Crash severity is quantified by the change in velocity
experienced by a vehicle during an impact along with the
time duration over which that change in velocity occurs.
Since the values of the input parameters for calculating
the change in velocity are not known exactly, there is
uncertainty associated with the calculated change in
velocity. Accurate evaluation of the crash severity will,
therefore, include analysis of the effect that uncertainties
in the values of the input parameters have on the
calculated change in velocity. Monte Carlo simulation, a
statistical technique, enables the reconstructionist to
evaluate the effect of uncertainty on the analysis of crash
severity.
Use of the Monte Carlo simulation technique is beneficial
since a reconstructionist can enter a range of values for
each input parameter. A probability distribution can be
assigned to the range of values, which indicates the
likelihood that any value in that range corresponds to the
actual value of the parameter. The simulation generates
thousands of possible combinations of the input
parameters selected from the specified ranges, monitors
the results of the combinations and analyzes them
statistically. Application of the Monte Carlo technique is
intended to improve the legitimacy of crash severity
analysis by helping the reconstructionist consider a wide
range of possible solutions within the bounds of the
imperfect data and report statistically meaningful ranges
for the change in velocity.
This paper demonstrates the application of the Monte
Carlo technique to impact severity analysis using a
derived two-dimensional, rigid body, momentum-based
impact model. Thorough guidance is given to aid the
reconstructionist in integrating the momentum model
with the Monte Carlo simulation technique and this
method is illustrated with a case study. Since the impact
model employs restitution constraints in the normal and
tangential directions, the effect of uncertainty in
formulating appropriate ranges for the values of the
restitution coefficients is discussed.
NOTATION
m mass
I moment of inertia
V velocity
angular velocity
V change in velocity
change in angular velocity
P impulse
a X-coordinate of C.G.
b Y-coordinate of C.G.
c tangential coordinate of C.G.
d normal coordinate of C.G.
e coefficient of restitution
orientation of the slip plane
Subscripts
1, 2 vehicle numbers
x X-direction
y Y-direction
t tangential direction
n normal direction
i pre-impact
f post-impact
INTRODUCTION
Crash severity, quantified by the change in velocity
experienced by each vehicle during the impact
1
, is a key
parameter for assessing occupant injuries in motor
vehicle accidents [10, 18, 25, 33]. The interest of
litigators in knowing the potential for occupant injuries
makes the accurate assessment of crash severity
essential. However, the data necessary to evaluate the
change in velocity experienced by each vehicle is not

1
Strictly speaking, crash severity is quantified by the change in
velocity experienced by the vehicle during impact along with the time
duration over which that change in velocity occurs [15]. In the
discussion that follows, crash severity is used interchangeably with
change in velocity. This is only to improve readability, and comes with
the recognition that the time duration of the change in velocity is also
an important parameter in assessing the crash severity.
Copyright 2001 SAE International and Messe Dsseldorf.
known with complete certainty [5, 19, 20, 21, 24, 28, 29,
30, 31, 32]. The location of the point of impact and the
points of rest, the orientation of the vehicle velocity
vectors before and after impact, and the location where
the resultant collision force is exchanged cannot be
estimated or measured perfectly. Estimation of
reasonable values for the coefficients of restitution at
varying levels of crash severity and varying impact
configurations is also fraught with uncertainty.
Coefficients of friction are typically unmeasured and
must be estimated. Even when they are measured, there
will remain questions about the fidelity of the
measurement in relation to the conditions at the time of
the crash. This uncertainty associated with the input
parameters results in subsequent uncertainty that is
associated with the calculated change in velocity.
Monte Carlo simulation provides a statistical analysis
technique to analyze the propagation of uncertainty from
the input parameters to the final result, leading to
statistically relevant conclusions regarding the probable
V experienced by a vehicle during an impact. The
Monte Carlo technique allows a range of values to be
specified for each input parameter of the impact model
reflective of the level of certainty associated with that
parameter. A probability distribution, that indicates the
likelihood that any value in that range corresponds to the
actual value, is attached to each of these ranges, and
the simulation then generates possible combinations of
these parameters based on the ranges and distributions.
The results of these combinations are tracked and
analyzed statistically. The final result is a probability
distribution that expresses the likelihood that any value of
the change in velocity corresponds to the actual value.
While application of Monte Carlo simulation to certain
aspects of crash reconstruction has been addressed in
the literature [19, 20, 30], previous papers have
addressed trivial applications. This paper presents
thorough and systematic guidance for the application of
Monte Carlo simulation to the non-trivial case of impact
analysis using a momentum-based rigid body impact
model. This impact model employs restitution constraints
in the normal and tangential directions, so the uncertainty
inherent in identifying the normal and tangential
directions is discussed. Guidance is given for the
formulation of reasonable ranges for the coefficients of
restitution.
ANALYTICAL UNCERTAINTY ANALYSIS
Analysis of crash severity should include analysis of the
effect that uncertainty has on the calculated change in
velocity. Several methods are available to analyze
uncertainty in a dependent variable based on
uncertainties in the independent variables. For the
analysis in this paper, the dependant variable is the
change in velocity and the independent variables consist
of the coefficients of friction and restitution, post-impact
travel distances, approach and departure angles, and
vehicle masses.
The most common method of uncertainty analysis is to
determine the upper and lower bounds of the dependent
variable based on the lowest and highest possible values
of each input parameter [5, 20]. This method has the
advantage of being straightforward, even with non-linear
equations. However, this upper and lower bound method
does not result in information regarding the likelihood
that any particular value within the obtained range for the
change in velocity corresponds to the actual value. No
basis is provided to conclude that crash severity values
near the mean of the output range are any more likely to
correspond to the actual value than those at the
extremes. Further, the probability that the values of all of
the independent variables fell at the extremes of these
ranges in the actual crash is small. Therefore, the range
generated for the dependent variable the change in
velocity will be unrealistic and wider than necessary.
Another method of uncertainty analysis uses differential
calculus to relate variations in the independent variables
to the resulting variation in the dependent variable. Brach
[5] and Tubergen [29] have expounded this method
within the context of crash reconstruction and that
discussion will not be repeated here. Suffice it to say that
in the impact model described below, the linear and
angular velocities are functions of ten independent
variables, the values of which are not known with
complete certainty. Analysis of the uncertainty using this
method will, therefore, include calculating partial
derivatives of the velocity equations with respect to each
of these ten independent variables. Further, analysis of
uncertainty by this method should be limited to cases
where the variations are small, since the method
approximates variations by linearizing the equations
around some nominal value. And finally, as in the first
method, this method gives no basis to draw statistical
conclusions regarding individual values within the range
of values obtained for the change in velocity.
An additional analytical method of uncertainty analysis
accounts for the statistical nature of the ranges
formulated for the independent parameters and allows
for statistical conclusions to be drawn regarding the
value of the dependent variable. A probability distribution
that indicates the likelihood that any value in a range is
likely to correspond to the actual value is specified for
each independent variable. This probability distribution
will be accounted for in analyzing the uncertainty
associated with the final result. Statistical judgements
can then be made regarding which values are most likely
to occur. Brach [5] has detailed this method for a limited
number of simple cases. However, this method becomes
impractical for more complex sets of equations, such as
the momentum equations considered below.
MONTE CARLO SIMULATION
Monte Carlo simulation provides a method of uncertainty
analysis that produces statistical data similar to that
produced by the analytical statistical method above,
while avoiding the limitations of that method for complex
equations. The Monte Carlo technique is an approximate
method of considering the effect of variations in the
independent variables on the uncertainty associated with
the dependent variable. The Monte Carlo technique
utilizes the power of a personal computer to accomplish
thousands of repeated calculations of the dependent
variable. These calculations are carried out with
randomly selected values for the independent variables
within the confines of ranges and probability distributions
assigned to those ranges. The results are tracked and
analyzed statistically. Conclusions relating to the likely
change in velocity experienced by vehicle during a crash
can then be drawn in a statistically legitimate manner.
The Monte Carlo analysis performed below uses a
commercially available Monte Carlo simulator called
Crystal Ball

that works in tandem with Microsoft

Excel. To perform the analysis, the equations for the

impact model are configured in Excel. A range of values
is then assigned to each independent parameter called
assumptions in Crystal Ball

into which the actual

value of that parameter should fall. Each range is then
assigned a probability distribution that indicates the
likelihood that any value in that range corresponds to the
actual value of that parameter. Finally, the simulation
runs a set number of calculations of the dependent
variable, called the forecast in Crystal Ball

, with values
for each parameter chosen randomly within the confines
of the assigned ranges and probability distributions.
Results of these calculations are monitored and the final
result is a probability distribution that indicates the
likelihood that any value in the resulting range
corresponds to the actual value of the dependent
variable (the change in velocity). Figure 1 is an example
of the output probability distribution obtained for a
vehicles change in velocity from the case study below.
FIGURE 1
CHOOSING PROBABILITY DISTRIBUTIONS
When choosing probability distributions for each
independent parameter, Crystal Ball

allows the user to

select from a number of distributions including those
depicted in Figure 2 below. Those distributions that might
be applied within the context of crash reconstruction
include the uniform, normal, and custom distributions.
FIGURE 2
Selection of a uniform distribution implies that any value
in the range is equally likely to occur. This distribution
represents the most conservative assumption since the
value of the parameter is allowed to vary more freely
within the range than with any other distribution. The
result is the widest possible variation in the output
parameter [30]. The uniform distribution should be
selected for most independent parameters in crash
severity analysis since data that would justify choosing
any other distribution is typically not available.
For instance, when a range and probability distribution is
being assigned to the post-impact travel distance of a
vehicle, this range will often be based on a single
measurement of that distance. While we might expect
random errors in measurement of the post-impact travel
distance to vary normally, one measurement is
insufficient to tell us the mean and standard deviation
produced by those random errors. Further, uncertainty
associated with the measurements of post-impact travel
distances may go beyond random errors and include
errors associated with identifying physical evidence.
There may be uncertainty associated with the exact
location where tire marks terminate, for example. Using
a normal distribution with this independent variable would
be unjustifiable with the provided data.
An exception to the use of a uniform distribution is for
coefficients of friction. Goudie, et al presented data from
540 skid-to-stop tests under both wet and dry road
conditions and found that random variations in the values
of the coefficients of friction approximated a normal
distribution [13]. Still, the use of a normal distribution to
describe variations in the coefficient of friction for any
particular reconstruction may be problematic. While
Goudie tested three types of tires, his data is still for a
single surface and a single vehicle. Random variations in
the coefficient of friction might be expected to exhibit a
normal distribution for any vehicle-tire-surface
combination. However, each vehicle-tire-surface
combination would have a unique mean value and a
unique standard deviation. In any particular crash
reconstruction, the mean value for that particular vehicle-
tire-surface combination would not be known. Any
number of normal distributions would be possible (Figure
3). Even extensive post-accident testing at the site would
not necessarily yield the proper mean value and
standard deviation since that testing would likely involve
a different vehicle with different tires and suspensions
characteristics. Furthermore, the generation of a
statistically significant amount of data for a particular
reconstruction is typically not feasible.
FIGURE 3
We could perhaps place bounds on the possible mean
values and standard deviations and construct a custom
distribution for variations in the coefficient of friction. For
example, given some roadway surface, an experienced
reconstructionist may be able to conclude that the mean
value will likely occur between 0.7 and 0.85 (Figure 4).
Using the custom distribution feature in Crystal Ball

,
the range between 0.7 and 0.85 could be assigned a
uniform distribution. The distribution could then tail off
linearly on either side to the extremes of the ranges
(Figure 4). This distribution displayed in Figure 4 by the
dashed line is wide enough to encompass the breadth of
values in the literature. Our intent is not to defend this
range of values, but only the general shape of the
distribution. A tighter range of friction coefficients could
be specified if there is justification for that tighter range,
such as consideration of the actual roadway surface
involved in the crash.
This distribution shape in Figure 4 recognizes that for a
given vehicle-tire-surface combination, variations in the
coefficient of friction are likely to exhibit a normal
distribution, but that in any particular reconstruction the
mean value of that normal distribution would be
unknown. At the same time, this distribution recognizes
that limits can be placed on possible mean values, and
values outside of these possible means need not be
considered as likely as those within the range of possible
means. This distribution also allows for a wide range of
standard deviations.
FIGURE 4
UNREALISTIC RESULTS
Some combinations of the independent variables
generated by the Monte Carlo simulator may produce
unrealistic results, and the output of the Monte Carlo
simulation should be filtered to eliminate these
combinations [30]. Mathematically, these unrealistic
results occur because the equations of the impact model
only constrain the output with respect to the independent
variables that are actually contained in the equations of
the impact model. For instance, there is no constraint in
the equations of the impact model placed on the lateral
acceleration that a vehicle can achieve. If a crash occurs
while one vehicle is making a left turn, the analyst may
be able to place limits on the speed of that vehicle based
on the radius of the turn and the maximum lateral
acceleration that the vehicle could achieve. The output
data of the Monte Carlo simulation would have to be
filtered based on this limit since the equations of the
impact model contain no inherent constraint that would
limit the speed of that vehicle based on maximum lateral
acceleration.
The coefficients of restitution provide another criterion by
which the result of each combination can be judged. The
impact model formulated below does not require the
restitution coefficients to be estimated in order to obtain
a solution, and therefore the results are not constrainted
with respect to reasonable restitution values. Output from
the Monte Carlo simulation should be filtered to include
only combinations that produce reasonable restitution
values. The formulation of reasonable restitution
coefficients is discussed below.
The most recent release of the Crystal Ball

software
allows for automatic filtering of the results of the Monte
Carlo simulation. The users manual should be consulted
for specific instruction for accomplishing this filtering.
WHAT DO THE RESULTS MEAN?
Application of the Monte Carlo technique is intended to
improve the legitimacy of crash severity analysis by
enabling the reconstructionist to consider thousands of
possible solutions within the bounds of the imperfect
data that is available. Legitimacy is not inherent in
application of the Monte Carlo technique, though, since
the Monte Carlo technique cannot establish the fidelity of
the impact model in any particular case. Meaningful
application of the Monte Carlo technique must come
after the assumptions of the impact model have been
adequately satisfied.
For example, a typical momentum-based impact model
assumes that tire forces can be neglected during the
impact. If a model with this assumption is applied to
analyze an impact of sufficient duration to make
momentum losses from tire forces significant, the error in
the solution induced by these momentum losses
undermines the fidelity of the model for that accident. No
amount of uncertainty analysis will solve this problem. A
more complex impact model is required that will account
for the tire forces that are neglected in the first model.
Monte Carlo simulation could then be applied with the
improved impact model.
Once the validity of the impact model has been
established, the validity of the ranges and probability
distributions assigned to the input parameters must
likewise be established. In the same way that an
improperly applied impact model undermines the
effectiveness of the Monte Carlo technique, so do poorly
chosen ranges for the input parameters. Test data from
the literature should be used, the reliability of the
available data should be assessed, and the range for
each parameter should be sufficiently wide to
encompass the full range of possible values. Once these
conditions are met, then the results obtained from Monte
Carlo simulation can be considered to give us
meaningful information about the likely value of the
change in velocity.
The results of the Monte Carlo analysis may be used in
various ways depending on the issues relevant to the
particular crash reconstruction. When the issue is crash
severity, a 51-percentile range, indicating the range of
values that are more probable than not, may be reported.
When the impact speeds of the vehicles are also at
issue, the 51-percentile range can likewise be reported
[18]. If the overall distribution obtained straddles the
speed limit, the reconstructionist may report the percent
probability that the vehicle was exceeding the speed
limit.
THE IMPACT MODEL
The two-dimensional impact model employed for
demonstration of the Monte Carlo simulation technique
makes use of the principle of impulse and momentum [6]
and is derived with the following assumptions typical of
momentum-based impact models [4, 7, 10, 16, 26, 27,
31, 32]:
1. Tire forces and other external forces are
assumed to be negligible compared to the
collision forces. This assumption allows for the
principle of conservation of linear momentum to be
applied and is generally considered valid for the
analysis of vehicular crashes. However, instances
exist when the duration of the impact is sufficiently
long to make external forces significant. Fonda [11]
has detailed a method for impact analysis that
includes the momentum losses that result from
external forces. Fondas method can be applied
when the impact duration is too long to ignore
external forces that cause significant momentum
losses.
2. The resultant impulse applied to each vehicle is
concentrated at the impact center. A number of
papers have discussed the identification of the
impact center [9, 14, 16, 31]. For the sake of brevity
this discussion will not be expanded here. Suffice it
to say that the impact center is the point in the crush
zone with a moment arm such that on the average
the cross product of that moment arm with the
impulse produces the correct rotational impulse
[31]. There is uncertainty associated with the
identification of the impact center that should be
accounted for in the analysis of crash severity.
Further research should be done to explore the
effect of variations in the location of the impact
center on the calculated change in velocity.
3. The interaction between the vehicles is assumed
to occur instantaneously. The collision forces are
transferred instantaneously and the pre-impact and
post-impact positions of the vehicles are assumed to
coincide. This instantaneous transfer of the collision
impulse is assumed to occur when the maximum
crush to each vehicle has been reached. The validity
of this assumption depends on the actual impact
duration. In a real crash the force is not transferred
instantaneously and movement of the vehicles does
occur. The longer the impact duration, the more
movement there will be and the less valid this
assumption is.
4. Each vehicle is treated as a rigid body with two
translational degrees-of-freedom and one
rotational degree-of-freedom. Yaw rotations of the
vehicles are considered, but roll and pitch rotations
are neglected. Vertical motion is also neglected.
5. Collision forces are treated as two-dimensional
and assumed to act in the same plane in which
the vehicles move.
6. The mass, center of gravity, and yaw moment of
inertia for each vehicle are assumed unchanged
by the impact.
The impact model employs a fixed X-Y coordinate
system located, for convenience, at the impact center, as
shown in Figure 5.
FIGURE 5
The location and orientation of this coordinate system
are arbitrary. After the orientation of the fixed coordinate
frame has been specified, the direction tangential to the
impact surface is identified in relationship to the X-axis.
The normal direction is by definition orthogonal to the
tangential direction, or normal to the impact surface. The
normal and tangential directions form a coordinate frame
useful for setting up constraint equations. The orientation
of the tangential direction () along with the approach and
departure angles of the vehicle velocities are measured
counterclockwise from the positive X-axis.
Defining an arbitrarily oriented fixed coordinate system
separate from the normal and tangential directions
allows the orientation of the tangential direction to be
included as an angular parameter in the impact analysis.
This allows the uncertainty inherent in the identification of
the normal and tangential directions to be considered in
the analysis of crash severity.
Formulation of the impact equations begins with
application of the principle of conservation of linear
momentum [6] in the X and Y directions. The following
equations are obtained:
The change in angular momentum of each vehicle is
calculated by application of the principle of impulse and
momentum. The following equations are obtained:
P
x
and P
y
are the components of the impulse applied to
Vehicle #1 in the X and Y directions, respectively. An
equal and opposite impulse is applied to Vehicle #2.
Since the impulse applied to each vehicle is equal to the
change in linear momentum experienced by that vehicle,
we obtain the following equations:
Substitution of equations (5) and (6) into (3) and (4)
yields four equations describing the impact. However, a
general two-dimensional impact model includes 3
degrees of freedom for each vehicle, therefore, six
equations are required to completely describe the
impact. Two constraint equations relating the
components of the closing speed to the components of
the separation speed at the impact center provide the
additional two equations necessary for a complete
impact model.
This impact model is similar to that derived by Ishikawa
[16] in that restitution constraints in the normal and
tangential directions provide the two additional equations
necessary for a complete model. The restitution
equations are as follows:
where c
1
and c
2
are the tangential coordinates of the
centers of gravity of vehicles #1 and #2, respectively,
and d
1
and d
2
are the normal coordinates of the centers
of gravity of vehicles #1 and #2, respectively. The
numerators of equations (7) and (8) are the normal and
tangential components of the separation speed. The
denominators are the normal and tangential components
of the closing speed. Transformation of restitution
equations (7) and (8) to the X-Y frame result in the
following equations:
where,
V m V m V m V m xf xf xi xi 2 2 1 1 2 2 1 1
+ = +
(1)
V m V m V m V m yf yf yi yi 2 2 1 1 2 2 1 1
+ = +
(2)
) (
b P a P ! ! I x y i f 1 1 1 1 1
+ =
(3)
) (
b P a P ! ! I x y i f 2 2 2 2 2
+ =
(4)
) ( ) (
V V m V V m P xf xi xi xf x 2 2 2 1 1 1
= =
(5)
) ( ) (
V V m V V m P yf yi yi yf y 2 2 2 1 1 1
= =
(6)
( ) ( )
( ) ( )

cos sin
cos sin
, ,
, ,
V V
V V
e
y close x close
y sep x sep
n
+
+
=
(9)


i ti i ti
f tf f tf
t close
t sep
t
d V d V
d V d V
V
V
e
2 2 2 1 1 1
1 1 1 2 2 2
,
,
+
+
= = (8)
( ) ( )
( ) ( )

sin cos
sin cos
, ,
, ,
V V
V V
e
y close x close
y sep x sep
t
+
+
=
(10)
(7)


i ni i ni
f nf f nf
n close
n sep
n
c V c V
c V c V
V
V
e
2 2 2 1 1 1
1 1 1 2 2 2
,
,
+
+
= =
( )
f xf f xf x sep b V b V V 2 2 2 1 1 1 ,
+ =
(11)
In these equations a
1
and a
2
are the X-coordinates of the
centers of gravity of vehicles #1 and #2, respectively,
and b
1
and b
2
are the Y-coordinates of the centers of
gravity of vehicles #1 and #2, respectively. Equations (1)
through (4), in conjunction with equations (5) and (6),
and equations (9) and (10) now provide a complete
description of the impact.
The values of the normal and tangential restitution
coefficients are allowed to vary between values of -1 and
1. Ishikawa introduced the idea of negative restitution
within the context of accident reconstruction [16]. The
result of allowing restitution to fall below 0 is that the
impact model does not require a common velocity to be
achieved during the impact. This is advantageous since
the impact model can be used to analyze sideswipe or
break-through collisions where a common velocity is
never achieved. However, once the values of the
restitution coefficients drop below 0, the classical
definition of the restitution constraints as a ratio of
deformation and restitution impulses has been
abandoned.
If the orientation of the vehicle velocities before and after
impact can be estimated, the number of unknowns is
reduced to four and equations (1) through (4), with (5)
and (6), are sufficient to produce a solution. Writing
equations (1) through (4) with approach and departure
angles yields equations (15) through (18).
Application of the impact model demonstrated below
assumes that approach and departure angles can be
estimated and equations (15) through (18) are utilized in
the Excel spreadsheet for the Monte Carlo analysis. The
coefficients of restitution become unnecessary to
generate a solution, but they are used to monitor
possible solutions for reasonableness. The issue of
reasonable and unreasonable solutions is discussed
below.
Finally, equations for the change in velocity experienced
by each vehicle during the impact are obtained by vector
subtraction of the final velocities from the initial
velocities.
where,
The principle direction of force (PDOF) is calculated
geometrically by the following equation:
RESTITUTION THEORETICAL ASPECTS
Traditionally, the constraint equations for two-
dimensional rigid body impact models have been
formulated by defining a coefficient of restitution in the
normal direction and an equivalent friction coefficient in
the tangential direction [4]. In classical mechanics, the
normal coefficient of restitution is defined at the impact
center as the ratio of the normal impulses during the
restitution and deformation phases [6].
Since the impulses during these phases are equal to the
change in momentum of each vehicle during that phase,
substitution, algebraic manipulation, and consideration of
the two-body system renders the familiar form of the
normal coefficient of restitution [6].
Definition of the coefficient of restitution as the ratio of
the restitution and deformation impulses requires that the
) (
f yf f yf y sep a V a V V 1 1 1 2 2 2 ,
+ =
(12)
) (
i xi i xi x close b V b V V 1 1 1 2 2 2 ,
+ = (13)
) (
i yi i yi y close a V a V V 2 2 2 1 1 1 ,
+ =
(14)
( ) ( )
( )

2 1 1
2
2 2 2 2
1
1 1
1
sin
sin sin

=
m
V m V m
V
f f
i
(15)
( )

2
2 2
1
1 1 1 1 1
2 2
2
cos cos cos
cos
1
V m V m V m
m
V f f i i
+ + =
(16)
( ) ( )


i i f f a b V
I
m
a b V
I
m
1 1 1 1 1 1
1
1
1
1
1
1
1
1
1
1
sin cos sin cos + =
(17)
( ) ( )


i i f f a b V
I
m
a b V
I
m
2 2 2 2 2 2
2
2
2
2
2
2
2
2
2
2
sin cos sin cos + =
(18)
2 2
V V V jy jx j
+ = (19)
j
jx
jy
Vj
V
V
PDOF

+ =

1
tan 180 (23)

j ji
j
jf jx V V V
cos cos =

(20)

j ji
j
jf jy V V V
sin sin =

(21)
(24)
dt F
dt F
impulse n deformatio normal
impulse n restitutio normal
Dn
Rn
n e

= =
(25)
V V
V V
e
ni ni
nf nf
n
2 2
1 2

=
(22)
2 , 1 = j
vehicles reach a common velocity in the normal direction
at the impact center at some time during the collision.
The dividing line between the period of deformation and
the period of restitution is by definition the instant at
which this common velocity is achieved and the
restitution coefficient indicates the extent to which the
vehicles rebound from this common velocity. This
common velocity assumption constrains the value of the
coefficient of restitution to lie between 0 and 1.
The equivalent coefficient of friction has been defined by
Brach [5] as follows:
Thus, the equivalent friction coefficient is a ratio that
indicates the magnitude of the tangential impulse relative
to the normal impulse. As Brach has noted, this
coefficient accounts not only for sliding friction between
the vehicles, but also for structural engagement. This
parameter controls the orientation of the principle
direction of force relative to the crush surface.
The impact model derived above parts with the
equivalent coefficient of friction as the constraint
parameter in the tangential direction. Instead,
coefficients of restitution are defined in both the normal
and tangential directions. These are not coefficients of
restitution in the traditional sense, though, since they are
allowed to vary between 1 and 1. This has the effect of
relaxing the common velocity requirement since a
restitution value of less than 0 implies that a common
velocity was never achieved. If this allowance for
negative restitution values is made, the coefficient of
restitution can no longer be defined as the ratio of the
restitution and deformation impulses since there is no
longer a clear dividing line separating the impact into two
phases. Instead, the normal and tangential restitution
constraints are defined on the pattern of equation (24)
without direct reference to the impulses.
Parting with the traditional impulse ratio definition of the
coefficients of restitution and the common velocity
assumption that accompanies that definition is
advantageous within the context of crash reconstruction
since the impact model can then handle collisions where
a common velocity is not reached at any time during the
collision. Sideswipe collisions are the most prominent
example. Another is crashes where one vehicle engages
the corner of the other and simply breaks through the
structure of that vehicle [16].
RESTITUTION APPROPRIATE RANGES
The model derived above assumes that the approach
and departure angles of the vehicle velocities can be
estimated and so the coefficients of restitution are not
necessary to generate solutions. However, potential
solutions must be checked for reasonable restitution
values. Thus ranges for the coefficients of restitution
must be estimated even though the model does not
require them. Rejection of unreasonable solutions is
discussed below.
Formulation of ranges for the coefficients of restitution
can be problematic since restitution in vehicular
collisions is a complex structural phenomenon where the
non-homogeneity of the vehicle structure causes the
restitution response to vary depending on what portions
of the vehicle are engaged in the impact. Also, each
vehicle combination will exhibit a unique collision
response at any given impact configuration and severity
[1, 14]. Estimation of the coefficients of restitution for a
particular crash will involve applying staged collision data
that does not strictly represent the specific vehicle
structures, the impact configuration, or the severity of
that particular crash.
NORMAL RESTITUTION
Crash Severity and Normal Restitution - Researchers
have generally concluded that normal restitution values
decrease as crash severity increases [1, 14, 16, 17, 25].
Some impact models have exploited this trend by
assuming that restitution approaches zero for high
severity collisions and can therefore be neglected in
these collisions [21]. As the severity of a collision
decreases, however, the effect of restitution on the
change in velocity experienced by each vehicle becomes
more significant, and at some point restitution cannot be
neglected. Even in high severity collisions, consideration
of the restitution response can improve the fidelity of the
impact model and may be important with certain impact
configurations. Restitution values that approach zero in
high severity collisions are best established for frontal
barrier impacts, and more testing and discussion is
needed before concluding that this assumption is valid
for all impact geometries.
The non-homogeneity of the vehicle structure introduces
important exceptions to the trend of decreasing
restitution with increasing crash severity. Frontal barrier
impact (VTB) tests and front-to-rear vehicle-to-vehicle
(VTV) collisions are the most represented collision types
in the literature and, therefore, discontinuities in
restitution response are best documented for the front
structure of the vehicle. The front structure of a vehicle
can be considered as several generally homogeneous
regions, each with elastic and plastic deformation
regions. As crush progresses through these regions, a
new region of elastic deformation will follow a region of
plastic deformation. When the crush enters a new elastic
deformation region, the result is a localized discontinuity
in the relationship between crash severity and restitution
[14].
P
P
n
t
=
(26)
Siegmund, Bailey, and King [25] conducted 660 low-
speed impact tests and found a slight rise in the
coefficient of restitution for some vehicles in frontal
barrier and front-to-rear VTV impacts that corresponded
to the bottoming out of the isolators. Monson and
Germane [23] analyzed 181 NHTSA full width VTB
impacts and reported an average coefficient of restitution
that was higher at 35 mph than at 30 mph for vehicles
with transverse-mounted engines. This rise in restitution
was attributed to engagement between the engine and
the cowl panel. In both instances, rises in restitution with
increasing crash severity corresponded to the entry of
the deformation into a new elastic region.
Another fundamental problem with using crash severity
to estimate restitution is that the crash severity is the
very parameter being calculated. If the impact severity
for a particular crash has not been determined, then the
impact severity cannot be strictly correlated to the
restitution. Estimates of restitution based on crash
severity must inevitably be based on visual inspection of
the damage to the vehicles. Such estimates may be
legitimate based on the experience of the
reconstructionist, but they remain estimates with some
level of uncertainty involved.
Impact Configuration and Restitution - Ishikawa [17] has
presented analysis of a series of 45 staged vehicle-to-
vehicle collisions at a number of impact configurations
and severities. Thirty-two of these collisions were side
impacts to the target vehicle, and the remaining 13 were
frontal impacts to the target vehicle. For the side
impacts, impact angles included 75 degrees, 90 degrees,
120 degrees, 135 degrees, and 150 degrees. For the
frontal impacts, impact angles included 120 degrees, 135
degrees, 150 degrees, 165 degrees, and 180 degrees.
The cars used in the collisions were Japanese
passenger cars weighing approximately 1000 kg.
FIGURE 6
Ishikawas data seems to indicate that any relationship
between crash severity and normal restitution breaks
down unless collisions are first grouped by their impact
configuration. Figure 6 depicts normal restitution values
reported by Ishikawa for 32 side impact tests plotted
against the closing speed in the normal direction. The
impact configurations and severities vary widely in the
data, and the results do not show a clear correlation
between crash severity and normal restitution for this
eclectic mix of collisions.
Figure 7 depicts normal restitution values reported by
Ishikawa for 13 frontal impact tests plotted against the
closing speed in the normal direction. Again, the impact
configurations and severities vary widely and again, there
is no clear correlation between crash severity and normal
restitution.
More research should be done to break Ishikawas data
down into classes by impact configuration to see if the
trend of decreasing normal restitution with increasing
crash severity holds within these groups. There may not
be sufficient data at this time to demonstrate this
conclusively, but a conceptual framework might be laid to
direct the expansion of the available staged collision
data.
FIGURE 7
While we may expect to find decreasing restitution
values with increasing crash severity for collisions
grouped by their impact configuration, there may be
certain impact configurations that do not reach the low
restitution values observed in high severity frontal
collisions. Side impacts at the axle may be one of these
classes. Impact configuration is, therefore, an important
consideration when estimating restitution. More
experimental and theoretical work needs to be
completed to clarify restitutions dependence on collision
geometry.
Variations in the Data - Normal restitution values show
considerable variation from test to test. For example,
barrier tests at 30 mph analyzed by Monson and
Germane exhibited coefficients of restitution varying
between approximately 0.04 and 0.24 with one test
falling outside this range at about 0.35. Tests at 35 mph
exhibited coefficients of restitution varying between 0.08
and 0.23. Scatter in coefficients of restitution observed
during their low-speed VTB and VTV impact tests led
King, Siegmund, and Bailey to conclude that accurate
estimates of the coefficients of restitution for a given
collision can only be determined by performing tests with
the actual vehicles [18]. Their tests exhibited restitution
values between 0.25 and 0.75 and they found noticeable
differences between the different bumper types.
TANGENTIAL RESTITUTION
Estimating values for the coefficient of restitution in the
tangential direction proves more problematic than
estimating corresponding values in the normal directions.
Data from staged collisions is limited. The primary data
available for estimating coefficients of restitution in the
tangential direction (and equivalent friction coefficients)
again comes from Ishikawas staged collision data [16,
17].
Side Impact Data - Figure 8 depicts tangential restitution
values reported by Ishikawa for 32 side impact tests
plotted against tangential closing speed.
This group of data shows the greatest potential for
displaying a general relationship between crash severity
and restitution. There is a trend of decreasing tangential
restitution with increasing tangential closing speed. This
trend makes intuitive sense in light of the physical
interpretation of the tangential coefficient of restitution. A
negative restitution coefficient implies that a common
velocity was never reached in the tangential direction.
We would expect that the higher the closing speed in the
tangential direction, the harder it would become for
structural engagement to cause a common velocity to be
reached. Thus, in general, at lower closing speeds the
engagement of the vehicle structures seem more
capable of bringing the vehicles to a common velocity in
the tangential direction.
FIGURE 8
Still, there is considerable variance in the relationship
between tangential closing speed and tangential
restitution, indicating that there are other important
factors. Intuitively, we would expect the normal closing
speed to affect tangential restitution values, as well,
since normal closing speed effects the extent of
structural engagement. A higher closing speed in the
normal direction would generally yield more significant
structural engagement and therefore a greater ability by
the vehicle structure to produce a common velocity in the
tangential direction.
This intuition is confirmed by Ishikawas staged collision
data. Figure 9 isolates the side impact tests with
tangential closing speeds between 30 and 40 mph and
plots their tangential restitution values against the normal
closing speed.
The structural engagement between the vehicles
appears to remain superficial up to around 15 mph, with
the normal closing speed playing little role in the
tangential restitution value. Above a 15 mph normal
closing speed there appears to be a general trend of
increasing tangential restitution values with increasing
normal closing speed.
FIGURE 9
Frontal Impact Data - Figure 10 depicts tangential
restitution values reported by Ishikawa for 13 frontal
impact tests plotted against the closing speed in the
tangential direction. Again, there is no clear correlation
between the tangential restitution values and the crash
severity.
FIGURE 10
To estimate a reasonable range of tangential restitution
values for a particular crash, the reconstructionist should
start by establishing a correlation between that particular
crash and the physical interpretation of the tangential
restitution coefficient. For instance, for many sideswipe
collisions a common velocity is never reached and
therefore the reconstructionist should expect a negative
restitution value. Likewise, the reconstructionist may be
able to conclude that structural engagement between the
vehicles was such that a common velocity in the
tangential direction was reached and that the tangential
coefficient of restitution should be zero or greater.
Beyond that general characterization, the
reconstructionist should consider the tangential
restitution values from collisions that most resemble the
crash in question. This data is limited and needs
expanding. Ishikawas data is helpful since he reports
values for both e
t
and . Other researchers report
neither.
GENERAL CONSIDERATIONS
Selection of Normal and Tangential Directions -
Restitution values reported for staged VTV tests, such as
those reported by Ishikawa, depend on the
reconstructionists selection of the normal and tangential
directions. Since there is uncertainty associated with the
selection of these directions, analysis of the same data
by another reconstructionist would likely lead to different
restitution values. It would be worthwhile to investigate
the effect of the uncertainty in the normal and tangential
directions. At any rate, reconstructionists should
recognize that restitution values published in the
literature have uncertainty associated with them.
Summary of Available Data - Reconstructionists should
be familiar with the staged collision data that is available
and use this data for estimating coefficients of restitution.
References 2, 3, 8, 15, 16, 17, 21, and 23 are excellent
sources of data. In the analysis of any particular crash
the reconstructionist should use the data from stage
collisions that best mimics that crash. The available data
will rarely be strictly analogous and the reconstructionist
should formulate ranges for the restitution values that
reflect the level of uncertainty associated with those
values. The ranges should be wide enough to
encompass the full range of possible restitution values.
For the foreseeable future, frontal barrier impact data will
continue to be the most readily available. While the
number of staged VTV collisions in the literature at
varying impact configurations and severities continues to
grow, there remain large gaps in the data for many
impact configurations and for most vehicle combinations.
Even if these gaps were filled for impact configuration,
testing of every vehicle combination is not feasible and
this gap is permanent. Analysis of crash severity should
give an accurate picture of the uncertainty inherent in
estimates of the coefficients of restitution.
CASE STUDY
BACKGROUND - The following case study is an attempt
to tie together the ideas that have been discussed,
including (1) the application of the impact model, (2) the
formulation of ranges and probability distributions for the
independent parameters, (3) the filtering of the data for
realistic restitution values, and (4) the formulation of
conclusions based on the filtered data. Detailed
instructions for the use of Crystal Ball

software are
contained in the user manual and are not repeated here.
FIGURE 11
The crash considered is an intersection collision
involving a Chevrolet pickup and a Nissan sedan. The
crash occurred when the Nissan attempted to make a
left turn in front of the oncoming Chevrolet. Figure 11
depicts the collision geometry, the point of impact, the
rest positions of the vehicles, and the coordinate system
used for application of the momentum model. Both
vehicles traveled approximately 22 feet after impact.
Parameter Low End High End
Weight of Chevy 6661 6861
Chevy Moment of Inertia 4712 5208
Chevy Approach Angle 181 191
Chevy Departure Angle 187 197
Chevy Post-Impact Travel Distance 20 22
Post-Impact Deceleration 0.60 0.95
Weight of Nissan 3410 3510
Nissan Moment of Inertia 1849 2043
Nissan Approach Angle 272 292
Nissan Departure Angle 192 202
Nissan Post-Impact Travel
Distance
20 24
Post-Impact Deceleration 0.6 0.95
Chevy X-Coordinate of Impact
Center
97.5 103.5
Chevy Y-Coordinate of Impact
Center
-16 -10
Nissan X-Coordinate of Impact
Center
-3.2 3.2
Nissan Y-Coordinate of Impact
Center
-28.6 -22.6
Orientation of Tangential Direction 96 106
TABLE 1
MONTE CARLO SIMULATION - The equations of the
impact model were entered in an Excel spreadsheet.
Each of the independent parameters was assigned a
range of values and a uniform probability distribution.
The range used for each variable is listed in Table 1.
A few of these ranges deserve comment. The Chevy
truck was carrying a load, but details regarding the
weight of the load were not available. Thus, the weight of
the Chevrolet was allowed to vary more widely than the
weight of the Nissan. Occupant weights were reasonably
well known. Vehicle weights were taken from published
data.
The point of impact and the points of rest were
documented by the police and were considered well
established. The motion of both vehicles between the
point of impact and the points of rest was relatively
straight. Therefore, the departure angles of both vehicles
were well established and were only allowed to vary
within a 10-degree range.
The approach angle of the Chevy was known reasonably
well. There were no pre-impact skid marks by the Chevy
before impact and the driver indicated that he had very
little time to react before the accident. This was
consistent with time-space analysis performed by these
engineers. The driver of the Chevy may have been able
to steer before impact, so the approach angle was
allowed to vary between 181 and 191 to allow for the
possibility of an evasive steer to the left before impact.
The ranges on the post-impact deceleration rates of the
vehicles are wide enough to encompass the breadth of
values in the literature for dry asphalt. We do not intend
to prescribe the use of this range. The reconstructionist
should use data from the literature and knowledge of the
specific surface and formulate a range on a case by
case basis.
Finally, the impact center was located using the crush
energy analysis portion of EDCRASH. Here, again, we
are not prescribing this as the preferred method for
locating the impact center.
Having assigned these ranges and a uniform distribution
to each of the independent variables, the Monte Carlo
simulation was run with 200,000 trials. The data
generated by the simulation was filtered to eliminate
unrealistic results based on conceptual considerations of
the restitution. Since there was significant engagement
between the vehicles, the authors concluded that the
vehicles reached a common velocity in both the normal
and tangential directions. The restitution values in both
the normal and tangential directions, therefore, had to be
greater than or equal to zero. Restitution values were
also required to be less than 1, since these values would
imply that energy was added to the collision.
Tangential restitution values were further restricted to a
range of 0.0 to 0.25 since it was clear from physical
evidence that the vehicles did not depart far from the
common velocity reached in the tangential direction.
Finally, impact speeds calculated for the Nissan were
required to be positive. Of the original 200,000
combinations of the parameters, 3,481 or approximately
2% of these combinations survived these rejection
criteria. This is similar to the percentage of accepted
values reported by other authors [30]. However, in the
experience of the authors, this percentage varies widely
from case to case.
Histograms were generated based on these remaining
values and are displayed in Figures 12 through 15. Table
3 lists the mean, standard deviation, and 51% range for
these histograms.
FIGURE 12
FIGURE 13
FIGURE 14
FIGURE 15
Mean Standard
Deviation
51%
Range
Chevy Change in
Velocity
11.8 .7 11.3 to 12.4
Nissan Change in
Velocity
23.1 1.4 22.0 to 24.1
Chevy Impact
Speed
33.1 1.8 31.8 to 34.5
Nissan Impact
Speed
4.8 2.0 3.3 to 6.2
TABLE 2
CONCLUSIONS AND RECOMMENDATIONS
MONTE CARLO SIMULATION
1. Monte Carlo simulation provides a statistical analysis
technique to analyze the propagation of uncertainty
in crash severity analysis. The results of a Monte
Carlo simulation allow the reconstructionist to draw
conclusions about the probable crash severity with
an understanding of which values are truly most
likely.
2. The Monte Carlo technique cannot establish the
fidelity of the impact model in any particular case.
Meaningful application of the Monte Carlo technique
must come after there has already been a strong link
established between the impact model and the
accident that the model mimics. In other words, the
assumptions of the model must fit well with the
actual accident.
3. Poorly chosen ranges for the input parameters
undermine the effectiveness and usefulness of the
Monte Carlo simulation technique. The range for
each parameter should be wide enough to
encompass the full range of possible values.
4. Selecting the uniform distribution to describe
variations in the independent variables represents
the most conservative assumption since the value of
the parameter is allowed to vary more widely than
with any other distribution and will result in the widest
variation in the dependent variable. The uniform
distribution should be preferred for most independent
parameters in crash severity analysis since data that
would justify choosing any other distribution is
typically unavailable.
5. A custom distribution can be constructed to
represent variations in the coefficients of friction that
recognizes that random variations in the friction
values vary normally, but that also gives adequate
consideration to uncertainty in the mean and
standard deviation of that normal distribution.
6. Certain combinations of values of the independent
parameters may produce unrealistic results, and the
output of the Monte Carlo simulation should be
filtered. Ranges formulated for the coefficients of
restitution provide one criterion by which to judge the
reasonableness of the result of each combination
and should be used to rule out certain combinations.
However, the range set on normal and tangential
restitution values should not be overly restrictive,
giving adequate consideration to the uncertainty
associated with restitution values.
THE IMPACT MODEL
1. The goal of this discussion has not been to defend
the fidelity and superiority of any particular impact
model, or even to defend the use of certain
constraint parameters within the context of
momentum impact models. The model utilized in this
paper was developed to give the reader insight into
issues surrounding the analysis of uncertainty in
impact modeling and crash severity analysis.
2. The impact model developed makes use of
coefficients of restitution in both the normal and
tangential directions. These coefficients of restitution
are allowed to vary between -1 and 1, with the
implication that there is no common velocity
requirement inherent in the model.
3. Parting with a common velocity requirement is
advantageous since the impact model can handle
sideswipe and break-through collisions.
UNCERTAINTY AND RESTITUTION
1. Restitution in vehicular collisions is a complex
structural phenomenon where the non-homogeneity
of the vehicle structure causes the restitution
response to vary depending on what portions of the
vehicle are engaged in the impact. Also, each
vehicle combination will exhibit a unique collision
response at any given impact configuration and
severity.
2. Researchers have generally concluded that normal
restitution values decrease as crash severity
increases. However, the non-homogeneity of the
vehicle structure introduces important exceptions to
the trend.
3. Ishikawa has presented a series of staged collisions
that seem to indicate that any relationship between
crash severity and restitution breaks down unless
collisions are first grouped by their impact
configuration.
4. Restitution values show considerable variation from
test to test. Ranges formulated for the coefficient of
restitution should be wide enough to encompass this
variation.
ACKNOWLEDGMENTS
The authors would like to thank Bruce Barnes, M.S., P.E,
Jeff Glinter, M.S., P.E. and Richard Ziernicki, Ph.D., P.E.
of Knott Laboratory, and Dr. John Trapp of the University
of Colorado for their helpful reviews of this paper. Any
errors, omissions and shortcomings remain the authors.
REFERENCES
1. Antonetti, Vincent W., Estimating the Coefficient of
Restitution of Vehicle-to-Vehicle Bumper Impacts,
980552, Society of Automotive Engineers,
Warrendale, PA, 1998.
2. Bailey, Mark N., Wong, Bing C., Lawrence, Jonathan
M., Data and Methods for Estimating the Severity of
Minor Impacts, 950352, Society of Automotive
Engineers, Warrendale, PA, 1995.
3. Bailey, Mark N., Lawrence, Jonathan M., Fowler,
Stuart J., Williamson, Peter B., Cliff, William E.,
Nickei, Jeff S., Data from Five Staged Car to Car
Collisions and Comparison with Simulations, 2000-
01-0849, Society of Automotive Engineers,
Warrendale, PA, 2000.
4. Brach, Raymond M., Brach, R. Matthew, A Review of
Impact Models for Vehicle Collision, 870048, Society
of Automotive Engineers, Warrendale, PA, 1987.
5. Brach, Raymond M., Uncertainty in Accident
Reconstruction Calculations, 940722, Society of
Automotive Engineers, Warrendale, PA, 1994.
6. Beer, Ferdinand P., Johnston, E. Russell, Vector
Mechanics for Engineers: Dynamics, 5
th
Edition, New
York: McGraw-Hill Book Company, 1988.
7. Bhushan, Bharat, Analysis of Automobile Collisions,
750895, Society of Automotive Engineers,
Warrendale, PA, 1975.
8. Cliff, William E., Moser, Andreas, Reconstruction of
Twenty Staged Collisions with PC-Crashs Optimizer,
2001-01-0507, Society of Automotive Engineers,
Warrendale, PA, 2001.
9. Crystal Ball 2000, Decisioneering, Denver, CO,
1988-2000.
10. Day, Terry D., Hargens, Randall L., An Overview of
the Way EDCRASH Computes Delta-V, 870045,
Society of Automotive Engineers, Warrendale, PA,
1987.
11. Fonda, Albert G., Nonconservation of Momentum
During Impact, 950355, Society of Automotive
Engineers, Warrendale, PA, 1995.
12. Fonda, Albert G., Principles of Crush Energy
Determination, 199-01-0106, Society of Automotive
Engineers, Warrendale, PA, 1999.
13. Goudie, D.W., Bowler, J.J., Brown, C.A., Heinrichs,
B.E., Siegmund, G.P., Tire Friction During Locked
Wheel Braking, 2000-01-1314, Society of
Automotive Engineers, Warrendale, PA, 2000.
14. Howard, Richard P., Bomar, John, Bare, Cleve,
Vehicle Restitution Response in Low Velocity
Collisions, 931842, Society of Automotive Engineers,
Warrendale, PA, 1993.
15. Husted, Daniel C., Biss, David J., Heverly, David E.,
The Appropriate Use of Delta-V in Describing
Accident Severity, 1999-01-1295, Society of
Automotive Engineers, Warrendale, PA, 1999.
16. Ishikawa, Hirotoshi, Impact Model for Accident
Reconstruction Normal and Tangential Restitution
Coefficients, 930654, Society of Automotive
Engineers, Warrendale, PA, 1993.
17. Ishikawa, Hirotoshi, Impact Center and Restitution
Coefficients for Accident Reconstruction, 940564,
Society of Automotive Engineers, Warrendale, PA,
1994.
18. King, David J., Siegmund, Gunter P., Bailey, Mark
N., Automobile Bumper Behavior in Low-Speed
Impacts, 930211, Society of Automotive Engineers,
Warrendale, PA, 1993.
19. Knott, Albert W., Probable Speed Analysis:
Statistical Modeling of Auto Accidents vs.
Deterministic Modeling, Journal of the National
Academy of Forensic Engineers, Vol. X No. 2,
December 1993.
20. Kost, Garrison, Werner, Stephen M., Use of Monte
Carlo Simulation Techniques in Accident
Reconstruction, 940719, Society of Automotive
Engineers, Warrendale, PA, 1994.
21. McHenry, Raymond R., McHenry, Brian G., Effects
of Restitution in the Application of Crush
Coefficients, 970960, Society of Automotive
Engineers, Warrendale, PA, 1997.
22. Metz, L.D., Metz, L.G., Sensitivity of Accident
Reconstruction Calculations, 980375, Society of
Automotive Engineers, Warrendale, PA, 1998.
23. Monson, Kenneth L., Germane, Geoff J.,
Determination and Mechanisms of Motor Vehicle
Structural Restitution from Crash Test Data, 1999-
01-0097, Society of Automotive Engineers,
Warrendale, PA, 1999.
24. Schockenhoff, Gerhard, Appel, Hermann, Rau,
Hartmut, Representation of Actual Reconstruction
Methods for Car-to-Car Accidents; as Confirmed by
Crash Tests, 850066, Society of Automotive
Engineers, Warrendale, PA, 1985.
25. Siegmund, Gunter P., Bailey, Mark N., King, David
J., Characteristics of Specific Automobile Bumpers in
Low-Velocity Impacts, 940916, Society of Automotive
Engineers, Warrendale, PA, 1994.
26. Smith, Gregory C., Conservation of Momentum
Analysis of Two-Dimensional Colliding Bodies, With
or Without Trailers, 940566, Society of Automotive
Engineers, Warrendale, PA, 1994.
27. Steffan, Hermann, Moser, Andreas, The Collision
and Trajectory Models of PC-CRASH, 960886,
Society of Automotive Engineers, Warrendale, PA,
1996.
28. Taylor, John R., An Introduction To Error Analysis:
The Study of Uncertainties in Physical Measurement,
2
nd
Edition, Sausalito: University Science Books,
1997.
29. Tubergen, Renard G., The Technique of Uncertainty
Analysis as Applied to the Momentum Equation for
Accident Reconstruction, 950135, Society of
Automotive Engineers, Warrendale, PA, 1995.
30. Wood, Denis P., ORiordain, Sean, Monte Carlo
Simulation Methods Applied to Accident
Reconstruction and Avoidance Analysis, 940720,
Society of Automotive Engineers, Warrendale, PA,
1994.
31. Woolley, Ronald L., The IMPAC Computer Program
for Accident Reconstruction, 850254, Society of
Automotive Engineers, Warrendale, PA, 1985.
32. Woolley, Ronald L., The IMPAC Program for
Collision Analysis, 870046, Society of Automotive
Engineers, Warrendale, PA, 1987.
33. Ziernicki, Richard M., Jacobson, Olof, Head Injury in
Automobile Accidents, 942377, Society of
Automotive Engineers, Warrendale, PA, 1994.
CONTACT
The authors welcome comments, questions, or
suggestions and can be contacted at the following
address:
Nathan A. Rose
Knott Laboratory, Inc.
7185 S. Tucson Way
Englewood, CO USA 80112
(303) 925-1900
[email protected]
A copy of the Excel file used for the analysis in this paper
will be provided upon request.