Accident Analysis and Prevention 47 (2012) 128–139

Contents lists available at SciVerse ScienceDirect

Accident Analysis and Prevention

journal homepage: www.elsevier.com/locate/aap

Occupant kinematics in low-speed frontal sled tests: Human volunteers, Hybrid

III ATD, and PMHS
Stephanie M. Beeman a , Andrew R. Kemper a,∗ , Michael L. Madigan b , Christopher T. Franck c ,
Stephen C. Loftus c
a
Virginia Tech – Wake Forest University, School of Biomedical Engineering and Sciences, Center for Injury Biomechanics, 325 Stanger St. (MC 0194), Blacksburg, VA 24061, USA
b
Virginia Tech, Department of Engineering Science and Mechanics, Kevin P Granata Biomechanics Laboratory, Blacksburg, VA 24061, USA
c
Virginia Tech, Department of Statistics, Laboratory for Interdisciplinary Statistical Analysis, Blacksburg, VA 24061, USA

a r t i c l e i n f o a b s t r a c t

Article history: A total of 34 dynamic matched frontal sled tests were performed, 17 low (2.5 g, v = 4.8 kph) and 17
Received 31 August 2011 medium (5.0 g, v = 9.7 kph), with five male human volunteers of approximately 50th percentile height
Received in revised form and weight, a Hybrid III 50th percentile male ATD, and three male PMHS. Each volunteer was exposed
21 December 2011
to two impulses at each severity, one relaxed and one braced prior to the impulse. A total of four tests
Accepted 12 January 2012
were performed at each severity with the ATD and one trial was performed at each severity with each
PMHS. A Vicon motion analysis system, 12 MX-T20 2 megapixel cameras, was used to quantify subject 3D
Keywords:
kinematics (±1 mm) (1 kHz). Excursions of select anatomical regions were normalized to their respective
Biomechanics
Trajectories
initial positions and compared by test condition and between subject types. The forward excursions of
Human surrogate the select anatomical regions generally increased with increasing severity. The forward excursions of
Acceleration relaxed human volunteers were significantly larger than those of the ATD for nearly every region at both
Collision severities. The forward excursions of the upper body regions of the braced volunteers were generally
Impact significantly smaller than those of the ATD at both severities. Forward excursions of the relaxed human
volunteers and PMHSs were fairly similar except the head CG response at both severities and the right
knee and C7 at the medium severity. The forward excursions of the upper body of the PMHS were generally
significantly larger than those of the braced volunteers at both severities. Forward excursions of the
PMHSs exceeded those of the ATD for all regions at both severities with significant differences within
the upper body regions. Overall human volunteers, ATD, and PMHSs do not have identical biomechanical
responses in low-speed frontal sled tests but all contribute valuable data that can be used to refine and
validate computational models and ATDs used to assess injury risk in automotive collisions.
© 2012 Elsevier Ltd. All rights reserved.

1. Introduction that frontal automobile collisions present an important focus for

Of the nearly 27,000 total vehicle occupants killed annually in
the United States, 25,000 of these fatalities are accounted for in
passenger car and light truck occupants (Nhtsa, 2009). Over 50% of
passenger car and light truck occupant fatalities are due to frontal
collisions. These numbers are dwarfed by the 931,000 occupants
sustaining injuries in frontal collisions (Nhtsa, 2009). It has been
reported that the total economic cost incurred as a result of motor
vehicle collisions 2000 was $230.6 billion (Blincoe et al., 2002). Fur-
thermore, the lifetime economic cost to society for each fatality
or each critically injured survivor were both approximated to be
around $1 million (Blincoe et al., 2002). These numbers indicate
∗ Corresponding author at: 325 Stanger St., Room 449 ICTAS (MC 0194), Blacks-
burg, VA 24061, USA. Tel.: +1 540 231 1617; fax: +1 540 231 2953.
E-mail address: akemper@vt.edu (A.R. Kemper).
research given the severity and economic impact.
While it is necessary to analyze automotive collision data and
statistics to gain an understanding of the mechanisms leading
to mortality and morbidity, it is also critical to perform labora-
tory tests to improve occupant safety by examining limitations in
current automobile safety systems. Knowledge regarding human
response and tolerance is essential to improving injury counter-
measures. Human occupant responses in motor vehicle collisions
are commonly predicted and evaluated in a laboratory setting
using surrogates (Begeman et al., 1980; Pintar et al., 1990, 2007;
Yoganandan et al., 1991; Hardy et al., 2001a,b; Rouhana et al., 2003;
Prasad et al., 2008; Kemper et al., 2009, 2011; Hallman et al., 2010;
Crandall et al., 2011; Duma et al., 2011). Several of the surrogates
for occupants in motor vehicle collisions include human volunteers,
anthropomorphic test devices (ATDs), and post mortem human
surrogates (PMHSs). Each of these has inherent strengths and lim-
itations, but the primary goal is to demonstrate a similar response

0001-4575/$ – see front matter © 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aap.2012.01.016
 S.M. Beeman et al. / Accident Analysis and Prevention 47 (2012) 128–139
to humans in motor vehicle collisions that can be used to evaluate
human tolerance and enhance vehicle design and safety.
Human volunteers have been used in numerous research stud-
ies to quantify their response and tolerance levels when subjected
to a variety of acceleration severity levels (Armstrong et al., 1968;
Ewing et al., 1969; Mertz and Patrick, 1971; Hendler et al., 1974;
Cheng et al., 1979; Begeman et al., 1980; Wismans et al., 1987;
Thunnissen et al., 1995; Szabo and Welcher, 1996; Van Der Horst
et al., 1997; Brault et al., 1998, 2000; Deng et al., 1998; Siegmund
et al., 2001, 2003a,b; Blouin et al., 2003; Choi et al., 2005; Kumar
et al., 2005; Ejima et al., 2007, 2008, 2009; Sugiyama et al.,
2007; Arbogast et al., 2009; Siegmund and Blouin, 2009; Beeman
et al., 2011). The distinguishing attribute of human volunteers that
non-human surrogates do not currently possess is the combina-
tion of identical human anthropometry, anatomy, and physiologic
response of the target population (Crandall et al., 2011). Human
volunteers have functioning circulatory and respiratory systems,
resting muscle tone and active bracing capabilities, continuous
neural responses, and the ability to perform cognitive functions.
Existing high-risk populations can be ethically studied without pre-
senting additional risk of injury (Duma et al., 2005a; Guskiewicz
et al., 2007; Schnebel et al., 2007; Broglio et al., 2009; Rowson
et al., 2009, 2011a,b; Hanlon and Bir, 2010; Beyer et al., 2011;
Duma and Rowson, 2011; Funk et al., 2011; Rowson and Duma,
2011). However, no direct observation of injury can be performed in
studies that introduce appreciable risk of injury to subjects. Exper-
iments can be performed in a controlled, objective manner with
human volunteers in a laboratory. While it is essential to evaluate
the response of human volunteers, all human volunteer laboratory
testing must be performed at sub-injurious levels due to ethical
constraints.
The use of non-human surrogates allows experimentation to
exceed ethical safety thresholds for humans and examine injurious
or traumatic events. ATDs are mechanical surrogates designed to
represent a particular demographic according to gender, size, and
age. In addition, they are designed to exhibit a biofidelic response
for specific loading conditions (e.g. principal direction of force
and severity) (Mertz, 2002). The responses of these devices are
not validated for alternate loading conditions and thus may not
produce biofidelic responses beyond their intended design specifi-
cations. Examples of using an ATD outside of the intended design
specifications include when the Hybrid III, an ATD designed for
frontal loading, is utilized in a side-impact or rear-end collision
test (Siegmund et al., 2001; Mertz, 2002). These devices not only
have the requirement of producing a biofidelic response under the
intended design specifications, they must also be durable and pro-
vide repeatable results. ATDs are used in multiple tests with high
loading conditions throughout their functional lifetime (Crandall
et al., 2011). While the durability of ATDs is economically sound,
it may also act as a limitation of these surrogates. One example of
this is the thoracic spine of the Hybrid III ATD. The thoracic spine
of the Hybrid III ATD is a rigid steel component that is unable to
bend or elongate as opposed to the human spine which has mul-
tiple segments that allow flexion, extension, and stretch (Crandall
et al., 2011). Currently, the data from PMHS studies are primarily
used to validate the response of these research tools.
PMHSs have been widely used to evaluate the response of
humans and assess safety countermeasures. A major advantage of
using PMHSs as a surrogate for humans is that PMHSs have the exact
anthropometric and anatomical representation of the population of
interest. As with the ATDs, experimentation is allowed to exceed
ethical safety thresholds for humans and examine injurious or trau-
matic events. Unlike with ATDs, injury can be directly observed
in PMHSs. Although PMHSs lack physiologic functions, researchers
attempt to compensate for this by pressurizing the cardiovascu-
lar system, inflating the lungs and adding mechanical hardware to
129
100
90 Test Start
80 2.5g Representative
70 5.0g Representative
60
Acceleration (m/s2)
50
40
30
20
10
0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50
-10
-20
-30
Time (sec)
Fig. 1. Sled acceleration (x-direction) time histories.
simulate muscle tone (Crandall et al., 2011). PMHSs cannot be used
to quantify injuries that require a physiologic response such as con-
tusions or mild traumatic brain injuries. In addition to the physical
limitations of PMHSs as a research tool, obtaining PMHSs that meet
the desired criteria for age, gender, anthropometry, and pathology
can be difficult and costly.
Given the capabilities and shortcomings of each surrogate in
automobile safety research, performing matched tests with two
or more of these can aid in the understanding of the biomechan-
ical response of humans in an impact environment. These have
included the comparison of multiple human volunteer ages and
sizes (Arbogast et al., 2009), comparison of multiple ATD models
(Siegmund et al., 2001; Törnvall et al., 2005, 2007), comparison
of human volunteers with ATDs (Hendler et al., 1974; Siegmund
et al., 2001; Morris and Cross, 2005; Törnvall et al., 2007), compar-
ison of human volunteers with PMHSs (Mertz and Patrick, 1971;
Cheng et al., 1979), comparison of ATDs with PMHSs (Törnvall et al.,
2005; Lopez-Valdes et al., 2010), and studies that compare human
volunteers, ATDs, and PMHSs (Begeman et al., 1980). The relative
performance and biofidelity in each comparison are dependent
upon a multitude of factors including: the hypothesis being tested;
the loading environment; the ATD model; the anthropometry and
age of the PMHS; and the anthropometry and age of the volunteers.
Iterations of comparisons between pairings of these surrogates
have provided significant contributions to the literature, improve-
ments in ATD design, and increased efficacy of safety devices.
However, differences in forward excursions of human volunteers,
ATDs, and PMHSs have not been evaluated in matched frontal accel-
eration tests at lower severities. Therefore, the purpose of this study
was to investigate the response of volunteer, ATD, and PMHS occu-
pant kinematics in matched low-speed frontal sled tests.
2. Material and methods
A total of 34 dynamic matched frontal sled tests were conducted
with five male human volunteers, a Hybrid III 50th percentile
male ATD, and three PMHSs. Of the 34 total tests, 17 were con-
ducted at a low severity (2.5 g, v = 4.8 kph) and 17 were conducted
at a medium severity (5.0 g, v = 9.7 kph). All volunteer and ATD
tests were performed using a custom mini-sled and test buck
accelerated by a pneumatic piston (Beeman et al., 2011). PMHS
tests were performed on a 1.4MN ServoSledTM system (Seattle
Safety LLC, Kent, WA) with the same test buck. An example test
buck acceleration–time history for volunteer, ATD, and PMHS tri-
als at 2.5 g and 5.0 g are shown in Fig. 1. Sled pulse magnitudes
were selected to ensure the safety of enrolled human volunteers
(Arbogast et al., 2009). Several studies that subjected volunteers to
130 S.M. Beeman et al. / Accident Analysis and Prevention 47 (2012) 128–139

Fig. 2. Subjects seated on test buck: relaxed volunteer, ATD, and PMHS.

accelerations exceeding these values reported subject discomfort severity. A single test was performed with each PMHS at each
and mild injuries (Ewing et al., 1969; Hendler et al., 1974; Cheng severity. The human volunteer data presented herein have been
et al., 1979). reproduced from a previously published manuscript, Beeman et al.
Approval to conduct the human subject testing presented in the (2011). This study expands upon the previous by adding a compar-
current study was granted by the Virginia Tech Institutional Review ison with matched ATD and PMHS experiments.
Board (IRB). In addition, all volunteers signed an informed consent An onboard data acquisition system (Diversified Technical Sys-
form prior to participating in the study. Selected volunteers were tems Inc., TDAS Pro, Seal Beach, CA) was used to record 148 channels
approximately 50th percentile male height (175 cm) and weight of data in the volunteer trials, 129 channels of data in the ATD,
(76.7 kg) (Table 1) (Schneider, 1983). and 170 channels of data in the PMHS trials, all at a sampling
All subjects were positioned in the center of the test seat (right rate of 20 kHz. Data collected for all subjects included subject head
to left), pelvis pushed backwards, with the feet centered on the foot accelerations, spine accelerations, chest contour, and forces at each
plates and hands on the simulated steering wheel (Fig. 2). The hands interface between the subject and test buck. In addition, surface
of the ATD and PMHSs were held in place with the use of masking electromyography (EMG) of 20 muscles (legs, arms, abdomen, back,
tape attached to the steering wheel. The legs of the ATD and PMHSs and neck) was measured in the volunteer trials and rib strain was
were positioned so that ankle, knee and hip joints were in line. For measured in PMHS trials. Belt tension was measured using load
each PMHS, the legs were held in place with masking tape attached sensors at the shoulder belt and lap belt were measured for all sub-
to the steering column. For PMHS 1, the head was held in place with ject types. Belt tension was also measured at the retractor for the
the use of masking tape attached to the seatback. For PMHS 2 and volunteers and ATD. Spool out of the belt was measured with a
3, the head was positioned so that cervical spine was in line with potentiometer attached to a custom seatbelt clamp. High-speed
the sagittal plane and held in place with the use of masking tape video was captured from the subjects’ aspect side at a sampling
attached to an overhead structure. Prior to testing, the masking tape rate of 1 kHz with the use of a high resolution, high light sensitiv-
on the legs, hands, and head was cut slightly to ensure that it would ity camera (Vision Research, Phantom V-9, Wayne, NJ). Analysis of
tear free during the test. All joints of the Hybrid III 50th percentile EMG, chest contour, strain, acceleration and force data were not the
male ATD were set to 1 g, barely restraining the weight of the limb focus of this manuscript.
when it was extended horizontally, prior to each trial (Nhtsa, 2005). A Vicon motion analysis system, consisting of 12 MX-T20
For all trials, a standard load limiting driver-side 3-point seatbelt 2 megapixel cameras, was used to quantify the 3D kinematics
was positioned around the test subject and the slack was removed. (±1 mm) of the subjects and the test buck at a sampling rate of
A total of 20 dynamic trials were performed with the five 1 kHz (Beeman et al., 2011). A total of 43 retro-reflective markers
human volunteers. Each volunteer was exposed to two impulses, were adhered to each human volunteer at key anatomical locations
one relaxed and one braced, at each of the two severities for a as described in Beeman et al. (2011). Markers were adhered to the
total of four trials per subject (Beeman et al., 2011). Prior to each ATD at the same general locations as the volunteer in the same
test, the volunteers were informed whether to remain relaxed or manner. Unlike with the volunteers, the markers on the right and
brace themselves for the sled impulse. The relaxed and braced left side of the Hybrid III head were aligned with the CG. Addition-
conditions were achieved as described in Beeman et al. (2011). ally, the greater trochanter markers were positioned at the H-Point
A total of eight tests were performed with the ATD, four at each socket hole of the ATD and the acromion markers were positioned
at the center of arm rotation. A total of 46 retro-reflective mark-
Table 1
ers were adhered to each PMHS. Bone screws were used to secure
Subject demographic and anthropometric information. markers at the bony landmarks on each PMHS at the same locations
as the volunteers. An additional marker was placed on the right iliac
Subject Gender Age (years) Height (cm) Weight (kg)
wing and the medial malleoli of each PMHS. The 3D reconstruction
Type Number of the Vicon data employed a minimum criterion of two cameras
Volunteer 1 Male 19 174.0 72.8 to triangulate the position of the markers. Marker trajectories were
2 Male 23 171.8 78.0 converted to the reference frame of the test buck, then to the SAE
3 Male 20 175.3 81.6 J211 sign convention (Sae, 1995).
4 Male 19 176.5 73.6
Determination of joint centers of volunteers during the dynamic
5 Male 20 176.4 76.8
ATD 1 Male – 177.8 77.7 trials has been previously explained in Beeman et al. (2011). Ankle
PMHS 1 Male 51 174.0 74.3 (AJC), knee (KJC), hip (HJC), wrist (WJC), elbow (EJC), and shoulder
2 Male 63 176.0 68.6 (SJC) joint centers were determined in the same manner for the
3 Male 79 184.0 86.4 PMHSs. As with the human volunteers, a static trial (500 Hz, 15,000
 S.M. Beeman et al. / Accident Analysis and Prevention 47 (2012) 128–139 131

Fig. 3. Schematic for joint angles in the sagittal plane.

frames) was performed with additional markers adhered after the
completion of the dynamic trials (Beeman et al., 2011). However,
the quiet standing in the anatomical position during the static trial
used for the volunteers was modified for the PMHSs. Each PMHS
was positioned in the supine position on a gurney for the static
trial. The additional markers adhered to the PMHSs were essential
for determining the HJCs throughout the dynamic trials would have
been obstructed by the lap belt during the dynamic trials. The pres-
ence of the markers in the static trial provided data regarding their
relative positions to markers present during the dynamic trials. The
head CG for each PMHS was defined as the mid-point between the
right and left markers at the anterior-superior point of the helix of
the outer ear in the coronal plane throughout the dynamic trials
(Naunheim et al., 2003). The position of the AJCs, KJCs, WJCs, and
EJCs of the ATD during the dynamic trials was determined using the
same methodology as with the volunteers. The HJCs were defined
as the marker located at the H-Point socket hole. The SJCs were
defined as the marker located on the shoulder joint fixture. As with
the PMHSs, the head CG was defined as the mid-point between the
left and right head markers aligned vertically with the CG of the
Hybrid III ATD head.
Joint center trajectory data were used to evaluate the initial posi-
tion of the volunteers prior to initiation of the sled pulse. Initial joint
angles were determined for each trial using pre-trigger data in the
sagittal plane (Fig. 3). The average initial position was determined
for the 10 relaxed volunteer trials and was used to position the ATD
on the test buck (Fig. 4). The position of the ATD was replicated for
each trial by marking reference points on the seat pan, seat back,
and each foot pedal to ensure repeatability for each trial. After posi-
tioning the ATD according to the reference points, the joint angles
Post hoc comparisons were made based on the statistical signifi-
cance of model terms. Specifically, differences in forward excursion
of each body region were determined between severities for each
subject type and between subject types for each severity. A sig-
nificance level of ˛ = 0.05 was adopted for all tests. To address
multiplicity, Fisher’s protected least significant difference (LSD)
was applied. Fisher’s protected LSD makes post hoc level compar-
isons for only those model terms which are significant according
to the ANOVA F tests. The standard error of each comparison is
computed using the estimated mean square error from the ANOVA
model and comparison-specific sample size. Residuals from the
model fit were examined via histogram and found to be suitably
normal as to justify this analysis.
3. Results
The test configurations and methodology used in the current
study provided the framework to compare the responses of dif-
ferent subject types in matched low-speed frontal sled tests. The
ANOVA table summarizing the fitted model is presented in Table 2.
The three-way interaction effect between body region, severity, and
-1200
Low - 1 Medium - 1
Low - 2 Medium - 2
-1000
Low - 3 Medium - 3
Low - 4 Medium - 4
-800
Z Position (mm)

were verified with an inclinometer.

-600
Excursions of select anatomical regions were normalized to their
respective initial positions to directly compare between test condi-
tions and subjects. Regions of interest included the EJCs, SJCs, KJCs, -400
HJCs, C7, and head CG. Peak forward (x) excursions of the select
regions of interest were determined for each trial. All trajectory
-200
data was cut at the respective peak forward excursions.
A full three-way ANOVA model was fit to explore the inter-
play between subject type (relaxed volunteer, braced volunteer, 0
0 -200 -400 -600 -800 -1000 -1200 -1400 -1600
ATD, PMHS), severity (low, medium), and body region as predic-
X Position (mm)
tors of forward excursion. This model included all two and three
way interaction effects as well as the main effect of each predictor. Fig. 4. Initial position for each ATD trial.
132 S.M. Beeman et al. / Accident Analysis and Prevention 47 (2012) 128–139

Table 2 -80
Statistical results of the three-way ANOVA model used to examine the interplay
between subject type, severity, and body region as predictors of forward excursion. Low Severity Braced
-60
Dependent variable ANOVA model term p-Value Medium Severity Braced
-40
Forward excursion Body region 0.000
Severity 0.000
Subject type 0.000 -20
360 320 280 240 200 160 120 80 40 0

Z (mm)
Body Region | Severity 0.009
Body Region | Subject Type 0.000 0
Severity | Subject Type 0.024
Body Region | Severity | Subject type 0.175 20
Bold values indicate statistical significance, p < 0.05.
40

60
subject type is not significant. Conversely, all two way interactions
are significant. This suggests that the effect of any of the predic- 80
tors on forward excursion depends on the level of both of the other X (mm)
factors (e.g. the effect of acceleration severity on forward excursion
Fig. 6. Representative comparison plot of low and medium severity braced volun-
depends on the body region and subject type). However, the nature
teer excursions. Note: Normalized head CG excursions shown.
of the two way interaction between any two variables does not dif-
fer by levels of the third predictor (e.g. the interaction between
-80
body region and subject type does not change at different levels
of acceleration severity). Based on these results, post hoc compar- Low Severity ATD
-60
isons of subject kinematics were then made between severities by Medium Severity ATD
subject type for all body regions, and between all subject types by -40
severity for all body regions.
In regard to increasing acceleration severity, all regional forward -20
excursions (x-axis) for the relaxed volunteers, braced volunteers, 360 320 280 240 200 160 120 80 40 0

Z (mm)
ATD, and PMHSs were generally larger for the medium severity 0
impulse than the low severity impulse This was statistically signif-
icant (p < 0.05) for all comparisons with the exception of the relaxed 20
volunteer left KJC and HIP, braced volunteer left EJC, right EJC, and
40
left SJC, and the PMHS right and left KJC and C7 (Tables 3–6). Rep-
resentative plots of the normalized forward excursions illustrating
60
the increased forward excursion with increasing severity are shown
in Figs. 5–8. 80
All subjects were exposed to the same sled acceleration pulses. X (mm)
Therefore, it was possible to discern differences in kinematic
Fig. 7. Representative comparison plot of low and medium severity ATD excursions.
response between the four subject types (Figs. 9 and 10). A rep-
Note: Normalized head CG excursions shown.
resentative plot of the normalized forward excursions illustrating
the kinematic response of the four subject types is shown in Fig. 11.
p-Values from the statistical analyses are tabulated in Tables 7–9. statistical analysis and the method used by Beeman et al. (2011) for
The comparisons between the relaxed and braced volunteers have these comparisons.
been omitted as they have been previously described in Beeman It was found that the average peak forward excursions of the
et al. (2011). However, it should be noted that there were no dif- relaxed volunteers were larger than those of the ATD at both
ferences in the statistical findings between the current method of severities (Table 7). This disparity was significant (p < 0.05) for
each region at both severities with the exception of both HJCs
at the medium severity. The average peak forward excursions of
-80 the relaxed volunteers and the PMHSs were, in general, com-
parable (Table 8). However, the average peak forward excursion
Low Severity Relaxed
-60 of head CG at the both severities were significantly (p < 0.05)
Medium Severity Relaxed larger for PMHSs than the relaxed volunteers. Forward excur-
-40 sions of the right knee and C7 at the medium severity were
significantly (p < 0.05) larger for the relaxed volunteers than the
-20 PMHSs.
360 320 280 240 200 160 120 80 40 0
Z (mm)

Differences were observed between braced volunteers and both

0
the ATD and PMHSs. The peak forward excursions of the ATD upper
body regions were found to generally be larger than those of the
20
braced volunteers, while the forward excursions of the lower body
40 regions of the braced volunteers were larger than those of the
ATD (Table 7). At the low severity, differences between the ATD
60 and braced volunteers were significant for both shoulders, C7, and
head CG. At the medium severity, differences between the ATD and
80 braced volunteers were significant for both elbows, both shoulders,
X (mm)
C7, and the head CG. The average peak forward excursions of the
Fig. 5. Representative comparison plot of low and medium severity relaxed volun-
braced volunteers were smaller than those of the PMHSs for each
teer excursions. Note: Normalized head CG excursions shown. region at both severities, with the exception of the right and left
 S.M. Beeman et al. / Accident Analysis and Prevention 47 (2012) 128–139 133

Table 3
Average peak forward excursions of the relaxed volunteers at the low and medium severity and the average percent increase in forward excursion as a result of increased
severity.

Region Forward excursion (mm) Increase in excursion (%) p-Value

Low Medium

Left Knee 38 ± 3 56 ± 8 47 ± 18 0.052

Hip 41 ± 9 56 ± 8 39 ± 21 0.103
Elbow 53 ± 10 83 ± 14 66 ± 52 0.001
Shoulder 81 ± 15 120 ± 25 49 ± 22 0.000
Right Knee 42 ± 3 65 ± 3 57 ± 17 0.011
Hip 42 ± 7 63 ± 7 53 ± 26 0.021
Elbow 53 ± 3 92 ± 18 75 ± 42 0.000
Shoulder 78 ± 7 135 ± 19 72 ± 23 0.000
Mid-Sagittal C7 89 ± 5 130 ± 15 48 ± 11 0.000
Head CG 123 ± 9 169 ± 14 38 ± 15 0.000

Bold values indicate statistical significance, p < 0.05.

Table 4
Average peak forward excursions of the braced volunteers at the low and medium severity and the average percent increase in forward excursion as a result of increased
severity.

Region Forward excursion (mm) Increase in excursion (%) p-Value

Low Medium

Left Knee 20 ± 9 45 ± 13 142 ± 50 0.006

Hip 16 ± 10 44 ± 17 203 ± 119 0.006
Elbow 20 ± 9 31 ± 14 60 ± 40 0.222
Shoulder 25 ± 14 37 ± 21 53 ± 50 0.208
Right Knee 21 ± 9 48 ± 14 153 ± 78 0.003
Hip 18 ± 11 47 ± 16 210 ± 145 0.001
Elbow 34 ± 9 33 ± 10 −7 ± 21 0.837
Shoulder 23 ± 10 44 ± 18 96 ± 41 0.020
Mid-Sagittal C7 35 ± 17 61 ± 25 81 ± 33 0.005
Head CG 64 ± 15 106 ± 20 68 ± 18 0.000

Bold values indicate statistical significance, p < 0.05.

Table 5
Average peak forward excursions of the ATD at the low and medium severity and the average percent increase in forward excursion as a result of increased severity.

Region Forward excursion (mm) Increase in excursion (%) p-Value

Low Medium

Left Knee 14 ± 1 34 ± 2 150 ± 8 0.044

Hip 15 ± 2 38 ± 2 150 ± 20 0.028
Elbow 32 ± 2 64 ± 4 104 ± 17 0.001
Shoulder 48 ± 3 83 ± 4 71 ± 9 0.001
Right Knee 15 ± 1 39 ± 2 162 ± 15 0.018
Hip 19 ± 2 52 ± 2 173 ± 23 0.001
Elbow 34 ± 2 73 ± 4 117 ± 12 0.000
Shoulder 53 ± 2 93 ± 4 76 ± 8 0.000
Mid-Sagittal C7 58 ± 3 89 ± 4 53 ± 9 0.003
Head CG 93 ± 4 144 ± 5 55 ± 6 0.000

Bold values indicate statistical significance, p < 0.05.

Table 6
Average peak forward excursions of the PMHSs at the low and medium severity and the average percent increase in forward excursion as a result of increased severity.

Region Forward excursion (mm) Increase in excursion (%) p-Value

Low Medium

Left Knee 20 ± 4 38 ± 8 91 ± 11 0.127

Hip 31 ± 8 66 ± 13 118 ± 25 0.003
Elbow 54 ± 2 90 ± 5 66 ± 6 0.003
Shoulder 84 ± 21 109 ± 20 31 ± 11 0.035
Right Knee 23 ± 7 44 ± 11 98 ± 32 0.072
Hip 30 ± 8 58 ± 12 101 ± 23 0.015
Elbow 52 ± 8 102 ± 16 97 ± 3 0.000
Shoulder 82 ± 3 120 ± 3 46 ± 3 0.002
Mid-Sagittal C7 86 ± 27 93 ± 15 12 ± 22 0.562
Head CG 173 ± 55 224 ± 81 29 ± 14 0.000

Bold values indicate statistical significance, p < 0.05.

134 S.M. Beeman et al. / Accident Analysis and Prevention 47 (2012) 128–139

Table 7
Between subject comparisons of average peak forward excursions by body region and severity: Relaxed vs. ATD and Braced vs. ATD.

Severity Region Relaxed vs. ATD Braced vs. ATD

Difference in excursion p-Value Difference in Excursion p-Value

(mm) (%) (mm) (%)

Low Left Knee 24 ± 1 64 0.013 6 ± 4 30 0.555

Hip 26 ± 4 63 0.008 1 ± 5 6 0.914
Elbow 25 ± 5 40 0.029 −12 ± 4 −60 0.219
Shoulder 33 ± 7 40 0.001 −23 ± 7 −91 0.017
Right Knee 27 ± 2 64 0.006 5 ± 4 23 0.590
Hip 23 ± 3 54 0.019 −2 ± 5 −10 0.862
Elbow 20 ± 2 37 0.041 1 ± 2 2 0.935
Shoulder 26 ± 3 33 0.008 −30 ± 4 −131 0.002
Mid-Sagittal C7 31 ± 3 35 0.002 −23 ± 8 −65 0.019
Head CG 30 ± 5 24 0.002 −28 ± 7 −44 0.003

Medium Left Knee 21 ± 3 38 0.027 10 ± 6 23 0.282

Hip 18 ± 4 33 0.061 6 ± 8 14 0.526
Elbow 19 ± 6 23 0.048 −33 ± 7 −109 0.001
Shoulder 37 ± 11 31 0.000 −46 ± 10 −125 0.000
Right Knee 26 ± 2 40 0.008 8 ± 6 18 0.381
Hip 11 ± 3 17 0.262 −5 ± 7 −11 0.597
Elbow 20 ± 8 21 0.042 −40 ± 5 −124 0.000
Shoulder 42 ± 9 31 0.000 −49 ± 8 −110 0.000
Mid-Sagittal C7 42 ± 7 32 0.000 −28 ± 11 −45 0.005
Head CG 25 ± 7 15 0.009 −37 ± 9 −34 0.000

Note: Differences reflect the ATD relative to relaxed volunteers and ATD relative to braced volunteers.
Bold values indicate statistical significance, p < 0.05.

Table 8
Between subject comparisons of average peak forward excursions by body region and severity: Relaxed vs. PMHS and Braced vs. PMHS.

Severity Region Relaxed vs. PMHS Braced vs. PMHS

Difference in excursion p-Value Difference in excursion p-Value

(mm) (%) (mm) (%)

Low Left Knee 18 ± 2 48 0.083 0 ± 5 −1 0.990

Hip 10 ± 6 25 0.335 −15 ± 7 −90 0.166
Elbow −1 ± 5 −2 0.903 −34 ± 4 −174 0.001
Shoulder −3 ± 14 −4 0.775 −59 ± 14 −232 0.000
Right Knee 19 ± 4 45 0.073 −3 ± 6 −17 0.796
Hip 13 ± 5 30 0.228 −12 ± 7 −67 0.262
Elbow 1 ± 5 3 0.887 −17 ± 5 −51 0.096
Shoulder −4 ± 4 −5 0.732 −59 ± 5 −259 0.000
Mid-Sagittal C7 3 ± 16 3 0.789 −51 ± 17 −144 0.000
Head CG −50 ± 32 −41 0.000 −109 ± 32 −169 0.000

Medium Left Knee 18 ± 6 32 0.087 7 ± 7 16 0.504

Hip −10 ± 8 −18 0.339 −22 ± 10 −51 0.036
Elbow −6 ± 7 −7 0.559 −59 ± 7 −191 0.000
Shoulder 11 ± 16 9 0.289 −72 ± 15 −196 0.000
Right Knee 21 ± 6 32 0.046 4 ± 9 8 0.728
Hip 5 ± 8 8 0.633 −11 ± 10 −23 0.299
Elbow −10 ± 12 −11 0.249 −70 ± 10 −214 0.000
Shoulder 15 ± 8 11 0.157 −76 ± 8 −172 0.000
Mid-Sagittal C7 38 ± 11 29 0.000 −31 ± 14 −51 0.003
Head CG −55 ± 47 −33 0.000 −117 ± 48 −110 0.000

Note: Differences reflect the PMHSs relative to relaxed volunteers and PMHSs relative to braced volunteers.
Bold values indicate statistical significance, p < 0.05.

knees at the low severity and the left knee at the medium severity
(Table 8). The average peak forward excursions of the left elbow,
both shoulders, C7, and head CG were significantly larger for the
PMHSs than the braced volunteers at low the severity. In addition
to those regions, the left hip and right elbow were also significantly
larger for the PMHSs than the braced volunteers at the medium
severity.
In comparing the peak forward excursions of the ATD and
PMHSs, it was found that the average peak forward excursions of
the PMHSs were larger than those of the ATD at both severities
(Table 9). At the low severity, differences between the ATD and
PMHSs were significant for the left elbow and shoulder, C7, and
head CG. At the medium severity, differences between the ATD and
PMHSs were significant for the left hip, both elbows, both shoulders,
and head CG.
4. Discussion
In the current study, the forward excursions of human volun-
teers, ATD, and PMHSs were evaluated in low-speed frontal sled
tests. Each subject was exposed to the same two low-speed frontal
sled impulses, one a low severity impulse and the other a medium
severity impulse. Each volunteer participated in two trials at each
severity. The first trial at each severity was in a relaxed muscle
 S.M. Beeman et al. / Accident Analysis and Prevention 47 (2012) 128–139 135

-320
Low Severity PMHS
-240
Medium Severity PMHS
-160

-80
360 320 280 240 200 160 120 80 40 0

Z (mm)
0

80

160

240

320
X (mm)

Fig. 8. Representative comparison plot of low and medium severity PMHS excur-
sions. Note: Normalized head CG excursions shown. Fig. 10. Average normalized peak forward excursions of the volunteers, ATD, and
PMHSs at the medium (5.0 g) severity.
state whereas for the second trial the subject tensed their muscles
to brace themselves prior to the impulse. Overall it was determined
volunteer right elbow. However, in the current manuscript the
that the severity of the impulse significantly affected the kinematic
increase in forward excursion as a result of increased acceleration
response of each subject type. Additionally, significant differences
severity was not statistically significant for the relaxed volunteer
were observed between subject types at both severities.
right knee and hip as well as the braced volunteer elbows and left
The experimental design allowed for the change in subject for-
shoulder. Paired t-tests were used to assess significant differences
ward excursion to be quantified with increasing acceleration. For
between the forward excursions in Beeman et al. (2011). For that
each subject type, increasing the severity of the sled impulse gen-
method, each individual test has its own estimate of variability.
erally resulted in an increase in average peak forward excursions
However, the statistical analysis model used in the current study to
of all regions of interest. The majority of increases in peak forward
adjust for multiplicity (n = 160 comparisons) utilizes a pooled esti-
excursions were determined to be statistically significant. How-
mate of variability for all comparisons. Therefore, the differences in
ever, those regions that were not found to increase significantly
statistical analysis between Beeman et al. (2011) and the one pre-
with acceleration severity still exhibited a considerable increase in
sented within this manuscript, namely the definition of variance
forward excursion. For example, average increases of 47% and 39%
for the post hoc t-tests, account for these differences. However, the
were observed for the left knee and hip of the relaxed volunteer,
overall conclusions were not affected by the method of statistical
respectively, yet these were not statistically significant. Increases
analysis.
of 60% and 53% were observed for the left elbow and shoulder of the
One of the major differences between the PMHSs and both the
braced volunteer, respectively, yet these were not statistically sig-
ATD and the volunteers is the effect of gravity on the head and
nificant. The left and right knee of the PMHSs exhibited increases
neck. The displacement and flexion of the PMHS C7 and head CG
of 91% and 98% in average peak forward excursion, respectively,
were influenced by the effects of gravity because functional mus-
yet were not statistically significant changes. Variability between
cles in the back and neck are not present to keep the cervical spine
subjects was a likely contributor to the statistical findings.
and head upright. The effect of gravity for the PMHSs was unlike
In Beeman et al. (2011), the forward excursions of all body
that of the human volunteers and the ATD. The human volunteers
regions of both volunteers were presented as having signifi-
have active muscles and the ATD has a stiff cervical spine which
cantly increased with acceleration severity except for the braced
supports the head, all of which acted to maintain the initial upright

-320
Relaxed Volunteers
Braced Volunteers -240
ATD
PMHS -160

-80
280 240 200 160 120 80 40 0
Z (mm)

80

160

240

320
X (mm)

Fig. 9. Average normalized peak forward excursions of the volunteers, ATD, and Fig. 11. Representative comparison plot of volunteer, ATD, and PMHS excursions.
PMHSs at the low (2.5 g) severity. Note: Normalized head CG excursions for low severity shown.
136 S.M. Beeman et al. / Accident Analysis and Prevention 47 (2012) 128–139

Table 9 Braced volunteers were found to have a more similar response to

Between subject comparisons of average peak forward excursions by body region
the ATD than the relaxed volunteers. Specifically, significant differ-
and severity: ATD vs. PMHS.
ences in peak forward excursion were only observed for upper body
Severity Region ATD vs. PMHS regions. At the low severity, the forward excursions of both shoul-
Difference in excursion p-Value ders, C7, and head CG were significantly larger for the ATD than the
braced volunteers. At the medium severity, the forward excursions
(mm) (%)
of both elbows, both shoulders, C7, and head CG were significantly
Low Left Knee −6 ± 2 −44 0.586 larger for the ATD. All joints of the Hybrid III 50th percentile male
Hip −16 ± 5 −103 0.156
ATD were set to 1 g, barely restraining the weight of the limb when
Elbow −22 ± 2 −71 0.042
Shoulder −35 ± 12 −73 0.001 it was extended horizontally, prior to each trial (Nhtsa, 2005). This
Right Knee −8 ± 4 −53 0.471 provided stability in the initial position as well as some resistance to
Hip −10 ± 4 −52 0.357 motion. This resistance resulted in similar responses of the lower
Elbow −18 ± 5 −54 0.097 extremities. However, the resistance of the upper body was not
Shoulder −29 ± 2 −56 0.008
Mid-Sagittal C7 −28 ± 16 −48 0.012
able to suppress the forward motion of the upper body to the same
Head CG −80 ± 32 −86 0.000 extent as muscle bracing in the braced volunteers. The magnitude
of the difference in excursion was observed to be directly related
Medium Left Knee −3 ± 5 −10 0.760
Hip −28 ± 7 −75 0.011 to acceleration severity. In other words, the overall average differ-
Elbow −25 ± 3 −39 0.022 ence (mm) in forward excursion for each body region increased
Shoulder −26 ± 12 −31 0.019 with acceleration severity.
Right Knee −5 ± 6 −12 0.662 Overall, the average peak forward excursions of the ATD were
Hip −6 ± 7 −11 0.596
between those of the relaxed volunteers and braced volunteers.
Elbow −29 ± 9 −41 0.008
Shoulder −27 ± 2 −29 0.014 While not specifically citing forward excursions, Begeman et al.
Mid-Sagittal C7 −4 ± 9 −4 0.723 (1980) also found the dynamic responses of ATDs to fall between
Head CG −81 ± 47 −56 0.000 relaxed and braced volunteers. The ATD may serve as an interme-
Note: Differences reflect the PMHSs relative to ATD. diate between these two extreme muscle conditions at the low
Bold values indicate statistical significance, p < 0.05. level accelerations evaluated in the current study. As acceleration
severity increased, the ATD became statistically more similar to the
position and influenced the overall kinematic responses. Supple- relaxed volunteers and statistically more dissimilar to the braced
mental fixtures were required to maintain the initial head position volunteers, while still proving to be more similar to the braced
of the PMHSs prior to the test. Masking tape was used to temporar- volunteers overall. The resistance to forward motion achieved by
ily secure the head upright in each trial. Although the setup kept the the level at which the Hybrid III 50th percentile male ATD joints
head upright prior to the initiation of the sled pulse, the subsequent are set may decrease with increasing levels of acceleration. This
interaction of the chest with the belt caused the head and cervical is consistent with the conclusion drawn in Beeman et al. (2011)
spine to rotate about the base of the neck with limited mechanical regarding braced volunteers. The results of that study implied that
resistance. each subject and body region may possess a unique acceleration
When comparing the average relaxed volunteer response to the threshold at which muscle bracing is insufficient to significantly
average ATD response, significant differences were observed for reduce forward excursions as inertial forces become larger. While
each region at both severities with the exception of the hips at the the effects of muscle bracing on the forward excursions of vari-
medium severity. For every region at both severities, the average ous body regions may diminish with increasing acceleration, the
forward excursion was larger for the volunteers than the ATD. These change in initial position between a relaxed and braced subject as
differences cannot be attributed to the positioning of the ATD on well as the increased internal forces acting about long bones and
the test buck as the positioning was based off of the initial position joints may alter the resulting injury patterns in a collision (Beeman
of the relaxed volunteers. However, they are likely a result of the et al., 2011). These effects are important to consider when using
design of the ATD itself. The Hybrid III ATD is commonly utilized in the response of an ATD to predict the risk of injuries for a human
motor vehicle safety standard testing (e.g. FMVSS 208) which are at any acceleration severity.
performed at a higher severity than in the current study. The ATD The forward excursions of the relaxed volunteers were, in gen-
is designed to be durable and repeatable under these high loading eral, similar to those of the PMHSs. The general similarity in
conditions. These design specifications may compromise the biofi- response between the two subject types is not illogical given that
delity at lower severities. One example of this is the thoracic spine the PMHS and the relaxed volunteer are anatomically and com-
of the Hybrid III ATD, a steel component that does not allow bend- positionally the same. At the low severity, only the head CG was
ing (Crandall et al., 2011). Only the cervical and lumbar regions are significantly different as the average relaxed volunteer excursion
able to exhibit flexion. The lack of thoracic spine bending affects was superseded by that of the average PMHS. At the medium sever-
the overall motion of the torso, including the interaction with the ity, the average right knee and C7 excursions were significantly
restraint system. Additionally, the shoulder joint of the Hybrid III larger for the relaxed volunteers than PMHSs while the average
has several shortcomings in its biofidelity (Törnvall et al., 2007). head CG excursion was significantly larger for the PMHS than the
These include the shape, components, stiffness, rotational degrees relaxed volunteers. Wismans et al. (1987) compared the head CG
of freedom, and ability to translate relative to the thorax. Specifi- response of volunteers and PMHSs at a 15 g acceleration level and
cally, the shoulder joint lacks the general shape and structure of a concluded that PMHS head excursions and rotations appeared to
human shoulder that affect the belt interaction with the shoulder be larger (Wismans et al., 1987). This is consistent with the current
and chest. A movable clavicle and scapula do not exist in the ATD study performed at a lower severity, and is likely due to the afore-
which affect the human shoulder range of motion. In testing per- mentioned effect of gravity and lack of resting muscle tone in the
formed by Törnvall et al. (2007), it was shown that the shoulder back and neck to support the head of the PMHSs.
design of the Hybrid III was much stiffer than human volunteers The average peak forward excursions of the PMHSs were, in
and allowed for extremely limited anterior, superior, and medial general, larger than those of the braced volunteers at both severi-
motions. The range of motion of the volunteers in that study proved ties. The only exceptions were the left knee at the low severity and
to be at least three times larger than the Hybrid III. the left and right knee at the medium severity. The differences in
 S.M. Beeman et al. / Accident Analysis and Prevention 47 (2012) 128–139
peak forward excursion between the PMHSs and braced volunteers
were found to be significant for the left elbow, both shoulders, C7,
and head CG at the low severity. In addition to these regions, dif-
ferences in the response of the right elbow and the left hip were
significant at the medium severity. The extended upper extremities
with active bracing were shown to greatly contribute to the atten-
uated forward excursions of the upper bodies of the volunteers
(Beeman et al., 2011). The lack of resting muscle tone and active
muscles in the PMHSs resulted in the upper extremities providing
less resistance to forward motion. The forward excursions of the
PMHS upper body, specifically the shoulders, are dictated largely
by the interaction of the chest with the belt. The arms and shoul-
ders of the PMHSs move medially and continue to extend forward
as the chest interacts with the belt. The same shoulder extensions
were not observed in the braced volunteer trials because bracing
with the upper extremities decreases the interaction between the
subject and the shoulder belt. Again, the lack of resting muscle tone
in the back and neck to support the head of the PMHSs as well as
the effect of gravity are likely key contributors to the observed dif-
ferences in C7 and head CG excursions between PMHSs and braced
volunteers.
The average peak forward excursions of the PMHSs were larger
than those of the ATD at both severities for every region. This gen-
eral trend is consistent with the findings of Lopez-Valdes et al.
(2010) who found that even at the low-speed level of v = 9 kph
(current study medium severity: v = 9.7 kph) the head and spine
forward displacements of the ATD were less than those of the
PMHSs (Lopez-Valdes et al., 2010). At the low severity, this was
significant for the left elbow, both shoulders, C7, and head CG.
At the medium severity, this was significant for the left hip, both
elbows, both shoulders, and head CG. As previously noted, it is pos-
sible that the general rigidity of the ATD and the limited range of
motion of the shoulder joints affect the biofidelity, particularly at
the low accelerations used in the current study. As the chest and
shoulder of the subjects load the belt, the torso is able to pivot
about the belt because the restraint is not symmetric. As with the
human volunteers, the range of motion of the shoulder joint for the
PMHSs allows the right arm and shoulder to move medially and
continue to extend forward as the chest interacts with the driver-
side shoulder belt. Comparatively, this motion is not possible for
the ATD as a result of shoulder joint construction. These attributes
cause the excursions of the both shoulder to be significantly dif-
ferent between the ATD and PMHSs at both severities. With regard
to the forward excursions of the subjects’ elbows at the low sever-
ity, the PMHS forward excursions were significantly larger than
those of the ATD for the left elbow. At the medium severity, the
forward excursions of the elbows were significantly larger for the
PMHSs than the ATD bilaterally. The lack of significance in the dif-
ference between the forward excursions of the PMHS and ATD right
elbow may be partially attributable to the hands of the subjects.
The right hand of PMHS 2 and PMHS 3 both become detached from
the steering column as the wrist rotates around the handle. With-
out contact with the sled, the force from the steering column is no
longer directly acting through the hand and forearm to increase
elbow flexion.
Direct comparisons of the results of this study to previous stud-
ies are limited. Very few studies expose human volunteers, ATDs,
and PMHSs to matched test conditions. For studies that do involve
these subject types in matched frontal sled tests, the responses of
the subjects are often focused primarily on reaction forces, neck
forces, and head acceleration. However, the focus of this manuscript
is on the displacement of various body regions. In addition, few
studies examine the response of ATDs or PMHSs, in general, at the
acceleration severities used in the current study. Studies involv-
ing human volunteers are subject to ethical constraints. Therefore,
a benefit of using non-living surrogates is the ability to examine
137
potentially injurious events. Relevant comparisons of the relaxed
and braced human volunteer responses between the current study
and previously published studies have been discussed in Beeman
et al. (2011). Displacement data for ATDs and PMHSs in frontal tests
at acceleration severities similar to those used in the current study
are not available for comparison.
Overall, the trends in peak forward excursions between subject
types within this study highlighted the strengths and weaknesses of
surrogates. Specifically, it demonstrated the importance of employ-
ing the Hybrid III ATD in the loading environment for which it was
designed. This ATD was validated with PMHS data from frontal
acceleration tests at higher severities than those used in the present
study. It has been designed to be used in frontal tests with high load-
ing environments (e.g. front impact regulation FMVSS 208). The
results of this study confirm that when this particular ATD is used
outside the realm of its design and validation, specifically at lower
severities, the surrogate does not exactly predict the response of
a relaxed or braced human. Rather, the response of the ATD is
between that of a relaxed and braced human volunteer. Although
the ATD is validated with PMHS data, the responses of the two sur-
rogates were not in agreement at the low severities of this study.
Additionally, it is important to note that even though the PMHS has
identical anatomical components as a live human, it is not a perfect
predictor of the response of a live human. Aspects of the surrogate
such as a lack of resting muscle tone influence the accuracy and
precision of the results. This effect is particularly pronounced at the
low severities used in the current study; muscles were shown to
exert enough force to, in part, control the kinematics in low-speed
frontal sled test (Beeman et al., 2011). Regardless of the limitations
of surrogates, ATDs and PMHSs are extremely useful in estimating
the kinematic response of humans and evaluating safety devices at
injurious levels.
Biofidelic ATDs can be used to optimize the design of injury
reducing systems based on their performance in a variety of loading
environments. ATDs have been extremely beneficial for predicting
human response for research and design in focal areas such as auto-
mobile safety as well as sports biomechanics and military impacts.
The frontal and side airbag designs require accurate dummy kine-
matics in order to maximize their protective abilities (Duma et al.,
1999, 2000, 2002, 2003, 2005c; Berliner et al., 2000; Hardy et al.,
2001b,a; Banglmaier et al., 2005; Jernigan et al., 2005; Prasad et al.,
2008). Biofidelic ATDs also contribute to reduced injury risk in
sporting environments by maximizing the protective capabilities of
sporting equipment (Vinger et al., 1999; Rowson et al., 2008, 2010;
Shain et al., 2010). Moreover, in military crash environments, it is
critical that the ATD kinematics match the human response so that
the restraint systems can be designed to reduce the risk of injuries
(Power et al., 2002; Duma et al., 2005b; Stitzel et al., 2005; Crowley
et al., 2009). The results the current study and those similar will
ultimately contribute to the continuous evaluation and potential
improvements to the design and implementation of ATDs.
Kinematics is only a part of the dataset required to fully evalu-
ate the response of human volunteers and human surrogates. The
occupant kinematic data regarding forward excursions presented
in the current article is a subset of the multitude of data collected
during the low-speed frontal sled tests, and is one of several pub-
lications that will stem from the overall study. Future publications
will present and discuss the results and implications of the EMG,
subject acceleration, load cell, and chest contour data.
5. Conclusions
This study illustrates that human volunteers, ATD, and PMHSs
do not have identical biomechanical responses in low-speed frontal
sled tests but all contribute valuable data as surrogates to those
138 S.M. Beeman et al. / Accident Analysis and Prevention 47 (2012) 128–139
individuals involved in motor vehicle collisions. In general, the
forward excursions of select anatomical regions increased with
increasing acceleration severity for all subject types. The forward
excursions of relaxed human volunteers proved to be significantly
larger than those of the ATD for nearly every region at both
severities. In general, only the upper body regions of the braced
volunteers were significantly different from the ATD at both sever-
ities, with the ATD excursions being larger. The forward excursions
of the relaxed human volunteers and PMHSs were fairly similar
with the exception of the head CG response at both severities and
the right knee and C7 at the medium severity. In general, the for-
ward excursions of the upper body of the PMHSs were significantly
larger than those of the braced volunteers at both severities. The
average forward excursions of the PMHSs exceeded those of the
ATD for all regions at both severities with significant differences
within the upper body regions.
Differences in an occupant’s kinematic response are important
to consider as they reflect the potential for different interactions
with safety devices and the vehicle interior. Modified interac-
tions give rise to different injury risks and patterns associated
with a prescribed vehicle collision mode. ATDs and PMHSs are
extremely useful in estimating the kinematic response of humans
and evaluating safety devices at injurious levels. However, this
study demonstrated the overall importance of human volunteers.
The knowledge gained from the combined biomechanical data of all
three types of surrogates can be used to refine and validate compu-
tational models and ATDs used to assess injury risk in automotive
collisions.
Acknowledgements
The authors would like to thank Toyota Motor Engineering &
Manufacturing (TEMA) and Toyota Central Research and Develop-
ment Laboratories (TCRDL) for sponsoring this research.
References
Arbogast, K.B., Balasubramanian, S., Seacrist, T., Maltese, M.R., García-España, J.F.,
Hopely, T., Constans, E., Lopez-Valdes, F.J., Kent, R.W., Tanji, H., Higuchi, K., 2009.
Comparison of kinematic responses of the head and spine for children and adults
in low-speed frontal sled tests. Stapp Car Crash J. 53, 329–372.
Armstrong, R.W., Waters, H.P., Stapp, J.P., 1968. Human muscular restraint during
sled deceleration. Stapp Car Crash J. 12, 440–462.
Banglmaier, R.F., Pecoraro, K.M., Feustel, J.R., Scherer, R.D., Rouhana, S.W., 2005.
Development and evaluation of a proposed neck shield for the 5 percentile
hybrid iii female dummy. Stapp Car Crash J. 49, 509–531.
Beeman, S.M., Kemper, A.R., Madigan, M.L., Duma, S.M., 2011. Effects of bracing on
human kinematics in low-speed frontal sled tests. Ann. Biomed. Eng. 39 (12),
2998–3010.
Begeman, P.C., King, A.I., Levine, R.S., Viano, D.C., 1980. Biodynamic response of the
musculoskeletal system to impact acceleration. SAE Technical Paper Series SAE
801312.
Berliner, J., Athey, J., Baayoun, E., Byrnes, K., Elhagediab, A., Hultman, R., Jensen,
J., Kim, A., Kostyniuk, G., Mertz, H., Prest, J., Rouhana, S., Scherer, R., Xu, L.,
2000. Comparative evaluation of the q3 and hybrid iii 3-year-old dummies in
biofidelity and static out-of-position airbag tests. Stapp Car Crash J. 44, 25–50.
Beyer, J.A., Rowson, S., Duma, S.M., 2011. Concussions experienced by major league
baseball catchers and umpires: field data and experimental baseball impacts.
Ann. Biomed. Eng.
Blincoe, L., Seay, A., Zaloshnja, E., Miller, T., Romano, E., Luchter, S., Spicer, R., 2002.
The Economic Impact of Motor Vehicle Crashes 2000. U.S. Department of Trans-
portation, Washington, DC.
Blouin, J.S., Descarreaux, M., Belanger-Gravel, A., Simoneau, M., Teasdale, N., 2003.
Attenuation of human neck muscle activity following repeated imposed trunk-
forward linear acceleration. Exp. Brain Res. 150 (4), 458–464.
Brault, J.R., Siegmund, G.P., Wheeler, J.B., 2000. Cervical muscle response during
whiplash: evidence of a lengthening muscle contraction. Clin. Biomech. (Bristol,
Avon) 15 (6), 426–435.
Brault, J.R., Wheeler, J.B., Siegmund, G.P., Brault, E.J., 1998. Clinical response of human
subjects to rear-end automobile collisions. Arch. Phys. Med. Rehabil. 79 (1),
72–80.
Broglio, S.P., Sosnoff, J.J., Shin, S., He, X., Alcaraz, C., Zimmerman, J., 2009. Head
impacts during high school football: a biomechanical assessment. J. Athl. Train
44 (4), 342–349.
Cheng, R., Mital, N.K., Levine, R.S., King, A.I., 1979. Biodynamics of the living human
spine during -gx impact acceleration. SAE Technical Paper Series SAE 791027,
723–763.
Choi, H.Y., Sah, S.J., Bumsoo, L., Cho, H.S., Kang, S.J., Mun, M.S., Lee, I., Lee, J., 2005.
Experimental and numerical studies of muscular activations of bracing occu-
pant. In: 19th Enhanced Safety of Vehicles Conference, Washington, DC, Paper
No. 05-0139.
Crandall, J.R., Bose, D., Forman, J., Untaroiu, C.D., Arregui-Dalmases, C., Shaw, C.G.,
Kerrigan, J.R., 2011. Human surrogates for injury biomechanics research. Clin.
Anat. 24 (3), 362–371.
Crowley, J.S., Brozoski, F.T., Duma, S.M., Kennedy, E.A., 2009. Development of the
facial and ocular countermeasures safety (focus) headform. Aviat. Space Environ.
Med. 80 (9), 831.
Deng, B., Melvin, J.W., Rouhana, S.W., 1998. Head-neck kinematics in dynamic for-
ward flexion. SAE Technical Paper Series SAE 983156.
Duma, S.M., Crandall, J.R., Pilkey, W.D., Seki, K., Aoki, T., 1999. Dynamic response of
the hybrid iii 3 year old dummy head and neck during side air bag loading. Proc.
Inst. Mech. Eng. Part D: J. Automobile Eng. 213 (D5), 471–480.
Duma, S.M., Crandall, J.R., Rudd, R.W., Kent, R.W., 2003. Small female head and neck
interaction with a deploying side airbag. Accid. Anal. Prev. 35 (5), 811–816.
Duma, S.M., Crandall, J.R., Seki, K., Aoki, T., 2000. Comparison of the q3 and hybrid
iii 3 year old dummy head and neck response during side air bag loading. Proc.
Inst. Mech. Eng. Part D: J. Automobile Eng. 214 (D7), 675–684.
Duma, S.M., Jernigan, M.V., Stitzel, J.D., Herring, I.P., Crowley, J.S., Brozoski, F.T., Bass,
C.R., 2002. The effect of frontal air bags on eye injury patterns in automobile
crashes. Arch. Ophthalmol. 120 (11), 1517–1522.
Duma, S.M., Kemper, A.R., Stitzel, J.D., Mcnally, C., Kennedy, E.A., Matsuoka, F., 2011.
Rib fracture timing in dynamic belt tests with human cadavers. Clin. Anat. 24
(3), 327–338.
Duma, S.M., Manoogian, S.J., Bussone, W.R., Brolinson, P.G., Goforth, M.W., Donnen-
werth, J.J., Greenwald, R.M., Chu, J.J., Crisco, J.J., 2005a. Analysis of real-time head
accelerations in collegiate football players. Clin. J. Sport Med. 15 (1), 3–8.
Duma, S.M., Ng, T.P., Kennedy, E.A., Stitzel, J.D., Herring, I.P., Kuhn, F., 2005b. Deter-
mination of significant parameters for eye injury risk from projectiles. J. Trauma
59 (4), 960–964.
Duma, S.M., Rath, A.L., Jernigan, M.V., Stitzel, J.D., Herring, I.P., 2005c. The effects of
depowered airbags on eye injuries in frontal automobile crashes. Am. J. Emerg.
Med. 23 (1), 13–19.
Duma, S.M., Rowson, S., 2011. Past, present, and future of head injury research. Exerc.
Sport Sci. Rev. 39 (1), 2–3.
Ejima, S., Ono, K., Holcombe, S., Kaneoka, K., Fukushima, M., 2008. Prediction of the
physical motion of the human body based on muscle activity during pre-impact
bracing. In: IRCOBI Conference, pp. 163–175.
Ejima, S., Ono, K., Holcombe, S., Koji, K., Fukushima, M., 2007. A study on occupant
kinematics behaviour and muscle activities during pre-impact braking based on
volunteer tests. In: IRCOBI Conference, pp. 31–45.
Ejima, S., Zama, Y., Ono, K., Kaneoka, K., Shiina, I., Asada, H., 2009. Prediction of pre-
impact occupant kinematic behavior based on the muscle activity during frontal
collision. In: 21st Enhanced Safety of Vehicles Conference, Stuttgart, Germany,
Paper No. 09-0913.
Ewing, C.L., Thomas, D.J., Patrick, L.M., Beeler Jr., Smith, G.W.M.J., 1969. Living human
dynamic response to – gx impact acceleration ii – accelerations measured on the
head and neck. Stapp Car Crash J. 13, 400–415.
Funk, J.R., Rowson, S., Daniel, R.W., Duma, S.M., 2011. Validation of concussion risk
curves for collegiate football players derived from hits data. Ann. Biomed. Eng.
Guskiewicz, K.M., Mihalik, J.P., Shankar, V., Marshall, S.W., Crowell, D.H., Oliaro, S.M.,
Ciocca, M.F., Hooker, D.N., 2007. Measurement of head impacts in collegiate
football players: relationship between head impact biomechanics and acute
clinical outcome after concussion. Neurosurgery 61 (6), 1244–1252, discussion
1252-3.
Hallman, J.J., Yoganandan, N., Pintar, F.A., 2010. Biomechanical and injury response
to posterolateral loading from torso side airbags. Stapp Car Crash J. 54, 227–257.
Hanlon, E., Bir, C., 2010. Validation of a wireless head acceleration measurement
system for use in soccer play. J. Appl. Biomech. 26 (4), 424–431.
Hardy, W.N., Schneider, L.W., Rouhana, S.W., 2001a. Abdominal impact response to
rigid bar, seatbelt, and airbag loading. Stapp Car Crash J. 45, 1–32.
Hardy, W.N., Schneider, L.W., Rouhana, S.W., 2001b. Prediction of airbag-induced
forearm fractures and airbag aggressivity. Stapp Car Crash J. 45, 511–534.
Hendler, E., O’rourke, J., Schulman, M., Katzeff, M., Domzalski, L., Rodgers, S., 1974.
Effect of head and body position and muscular tensing on response to impact.
SAE Technical Paper Series SAE 741184.
Jernigan, M.V., Rath, A.L., Duma, S.M., 2005. Severe upper extremity injuries in frontal
automobile crashes: the effects of depowered airbags. Am. J. Emerg. Med. 23 (2),
99–105.
Kemper, A.R., Kennedy, E.A., Mcnally, C., Manoogian, S.J., Stitzel, J.D., Duma, S.M.,
2011. Reducing chest injuries in automobile collisions: Rib fracture timing and
implications for thoracic injury criteria. Ann. Biomed. Eng. 39 (8), 2141–2151.
Kemper, A.R., Stitzel, J.D., Mcnally, C., Gabler, H.C., Duma, S.M., 2009. Biomechani-
cal response of the human clavicle: the effects of loading direction on bending
properties. J. Appl. Biomech. 25 (2), 165–174.
Kumar, S., Ferrari, R., Narayan, Y., 2005. Kinematic and electromyographic response
to whiplash loading in low-velocity whiplash impacts—a review. Clin. Biomech.
(Bristol, Avon) 20 (4), 343–356.
Lopez-Valdes, F.J., Lau, A., Lamp, J., Riley, P., Lessley, D.J., Damon, A., Kindig, M.,
Kent, R., Balasubramanian, S., Seacrist, T., Maltese, M.R., Arbogast, K.B., Higuchi,
 S.M. Beeman et al. / Accident Analysis and Prevention 47 (2012) 128–139
K., Tanji, H., 2010. Analysis of spinal motion and loads during frontal impacts.
Comparison between pmhs and atd. Ann. Adv. Automot. Med. 54, 61–78.
Mertz, H.J., 2002. Anthropomorphic test devices. In: Nahum, A.M., Melvin, J.W.
(Eds.), Accidental Injury: Biomechanics and Prevention. , 2nd ed. Springer Sci-
ence + Business Media, Inc., New York, NY, pp. 72–88.
Mertz, H.J., Patrick, L.M., 1971. Strength and response of the human neck. SAE Tech-
nical Paper Series SAE 710855.
Morris, R., Cross, G., 2005. Improved understanding of passenger behaviour during
pre-impact events to aid smart restraint development. In: 19th Enhanced Safety
of Vehicles Conference, Washington, DC, Paper No. 05-0320.
Naunheim, R.S., Bayly, P.V., Standeven, J., Neubauer, J.S., Lewis, L.M., Genin, G.M.,
2003. Linear and angular head accelerations during heading of a soccer ball.
Med. Sci. Sports Exerc. 35 (8), 1406–1412.
Nhtsa, 2005. Laboratory test procedure for fmvss 208, occupant crash protection. In:
Transportation, U.S.D.O. ed. Washington, pp. A-10.
Nhtsa, 2009. Traffic Safety Facts 2008: A Compilation of Motor Vehicle Crash Data
from the Fatality Analysis Reporting System and the General Estimates System.
U.S. Department of Transportation, Washington, DC.
Pintar, F.A., Sances Jr., A., Yoganandan, N., Reinartz, J., Maiman, D.J., Suh, J.K., Unger, G.,
Cusick, J.F., Larson, S.J., 1990. Biodynamics of the total human cadaveric cervical
spine. SAE Technical Paper Series SAE 902309.
Pintar, F.A., Yoganandan, N., Stemper, B.D., Bostrom, O., Rouhana, S.W., Digges, K.H.,
Fildes, B.N., 2007. Comparison of pmhs, worldsid, and thor-nt responses in sim-
ulated far side impact. Stapp Car Crash J. 51, 313–360.
Power, E.D., Duma, S.M., Stitzel, J.D., Herring, I.P., West, R.L., Bass, C.R., Crowley,
J.S., Brozoski, F.T., 2002. Computer modeling of airbag-induced ocular injury in
pilots wearing night vision goggles. Aviat. Space Environ. Med. 73 (10), 1000–
1006.
Prasad, P., Schneider, L.W., Hardy, W.N., 2008. Interactions of out-of-position small-
female surrogates with a depowered driver airbag. Stapp Car Crash J. 52,
187–203.
Rouhana, S.W., Bedewi, P.G., Kankanala, S.V., Prasad, P., Zwolinski, J.J., Meduvsky,
A.G., Rupp, J.D., Jeffreys, T.A., Schneider, L.W., 2003. Biomechanics of 4-point
seat belt systems in frontal impacts. Stapp Car Crash J. 47, 367–399.
Rowson, S., Beckwith, J.G., Chu, J.J., Leonard, D.S., Greenwald, R.M., Duma, S.M.,
2011a. A six degree of freedom head acceleration measurement device for use
in football. J. Appl. Biomech. 27 (1), 8–14.
Rowson, S., Brolinson, G., Goforth, M., Dietter, D., Duma, S., 2009. Linear and angular
head acceleration measurements in collegiate football. J. Biomech. Eng. 131 (6),
061016.
Rowson, S., Duma, S.M., 2011. Development of the star evaluation system for football
helmets: integrating player head impact exposure and risk of concussion. Ann.
Biomed. Eng. 39 (8), 2130–2140.
Rowson, S., Duma, S.M., Beckwith, J.G., Chu, J.J., Greenwald, R.M., Crisco, J.J., Brolin-
son, P.G., Duhaime, A.C., Mcallister, T.W., Maerlender, A.C., 2011b. Rotational
head kinematics in football impacts: an injury risk function for concussion. Ann.
Biomed. Eng.
Rowson, S., Mcnally, C., Duma, S.M., 2010. Can footwear affect achilles tendon load-
ing? Clin. J. Sport Med. 20 (5), 344–349.
Rowson, S., Mcneely, D.E., Brolinson, P.G., Duma, S.M., 2008. Biomechanical analysis
of football neck collars. Clin. J. Sport Med. 18 (4), 316–321.
139
Sae, 1995. Instrumentation for Impact Test, sae j211/1 mar 1995. Society of Auto-
motive Engineers, Warrendale, PA.
Schnebel, B., Gwin, J.T., Anderson, S., Gatlin, R., 2007. In vivo study of head
impacts in football: a comparison of national collegiate athletic association divi-
sion i versus high school impacts. Neurosurgery 60 (3), 490–495, discussion
495-6.
Schneider, L.W., 1983. Development of Anthropometrically based Design
Specifications for an Advanced Adult Anthropomorphic Dummy Family.
NHTSA.
Shain, K.S., Madigan, M.L., Rowson, S., Bisplinghoff, J., Duma, S.M., 2010. Analysis of
the ability of catcher’s masks to attenuate head accelerations on impact with a
baseball. Clin. J. Sport Med. 20 (6), 422–427.
Siegmund, G.P., Blouin, J.S., 2009. Head and neck control varies with perturbation
acceleration but not jerk: implications for whiplash injuries. J. Physiol. 587 (Pt
8), 1829–1842.
Siegmund, G.P., Heinrichs, B.E., Lawrence, J.M., Philippens, M.M., 2001. Kinetic and
kinematic responses of the rid2a, hybrid iii and human volunteers in low-speed
rear-end collisions. Stapp Car Crash J. 45, 239–256.
Siegmund, G.P., Sanderson, D.J., Myers, B.S., Inglis, J.T., 2003a. Awareness affects
the response of human subjects exposed to a single whiplash-like perturbation.
Spine (Phila Pa 1976) 28 (7), 671–679.
Siegmund, G.P., Sanderson, D.J., Myers, B.S., Inglis, J.T., 2003b. Rapid neck muscle
adaptation alters the head kinematics of aware and unaware subjects undergo-
ing multiple whiplash-like perturbations. J. Biomech. 36 (4), 473–482.
Stitzel, J.D., Hansen, G.A., Herring, I.P., Duma, S.M., 2005. Blunt trauma of the aging
eye: injury mechanisms and increasing lens stiffness. Arch. Ophthalmol. 123 (6),
789–794.
Sugiyama, T., Kimpara, H., Iwamoto, M., Yamada, D., 2007. Effects of muscle tense
on impact responses of lower extremity. In: IRCOBI Conference, Maastricht,
Netherlands, pp. 127–139.
Szabo, T.J., Welcher, J.B., 1996. Human subject kinematics and electromyographic
activity during low speed rear impacts. Stapp Car Crash J. 40, 295–315.
Thunnissen, J., Wismans, J., Ewing, C.L., Thomas, D.J., 1995. Human volunteer head-
neck response in frontal flexion. A new analysis. SAE Technical Paper Series SAE
952721, 439–460.
Törnvall, F.V., Holmqvist, K., Davidsson, J., Svensson, M.Y., Håland, Y., Öhrn, H., 2007.
A new thor shoulder design: a comparison with volunteers, the hybrid iii, and
thor nt. Traffic Inj. Prev. 8 (2), 205–215.
Törnvall, F.V., Svensson, M.Y., Davidsson, J., Flogard, A., Kallieris, D., Haland, Y.,
2005. Frontal impact dummy kinematics in oblique frontal collisions: evaluation
against post mortem human subject test data. Traffic Inj. Prev. 6 (4), 340–350.
Van Der Horst, M.J., Thunnissen, J.G.M., Happee, R., Van Haaster, R.M.H.P., Wismans,
J.S.H.M. 1997. The influence of muscle activity on head–neck response during
impact. SAE Technical Paper Series SAE 973346, 487–507.
Vinger, P.F., Duma, S.M., Crandall, J.R., 1999. Baseball hardness as a risk factor for eye
injuries. Arch. Opthalmol. 117, 354–358.
Wismans, J., Philippens, M., Van Oorschot, E., Kallieris, D., Matten, R., 1987. Com-
parison of human volunteer and cadaver head–neck response in frontal flexion.
Stapp Car Crash J. 31, 1–13.
Yoganandan, N., Skradem, D., Pintar, F.A., Reinartz, J., Sances Jr., A., 1991. Thoracic
deformation contours in frontal impact. SAE Technical Paper Series SAE 912891.