Traffic Injury Prevention

ISSN: 1538-9588 (Print) 1538-957X (Online) Journal homepage: https://www.tandfonline.com/loi/gcpi20

Analysis of rider and child pillion passenger

kinematics along with injury mechanisms during
motorcycle crash

Julaluk Carmai, Saiprasit Koetniyom & Wasif Hossain

To cite this article: Julaluk Carmai, Saiprasit Koetniyom & Wasif Hossain (2019) Analysis of rider
and child pillion passenger kinematics along with injury mechanisms during motorcycle crash,
Traffic Injury Prevention, 20:sup1, S13-S20, DOI: 10.1080/15389588.2019.1616180

To link to this article: https://doi.org/10.1080/15389588.2019.1616180

© 2019 The Author(s). Published with View supplementary material

license by Taylor & Francis Group, LLC

Published online: 05 Aug 2019. Submit your article to this journal

Article views: 2133 View related articles

View Crossmark data Citing articles: 3 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=gcpi20
TRAFFIC INJURY PREVENTION
2019, VOL. 20, NO. S1, S13–S20
https://doi.org/10.1080/15389588.2019.1616180

Analysis of rider and child pillion passenger kinematics along with injury
mechanisms during motorcycle crash
Julaluk Carmai , Saiprasit Koetniyom, and Wasif Hossain
Automotive Safety and Assessment Engineering Research Centre, The Sirindhorn International Thai–German Graduate School of Engineering,
King Mongkut’s University of Technology, North Bangkok, Thailand

ABSTRACT ARTICLE HISTORY

Objective: Traffic fatalities among motorcycle users are intolerably high in Thailand. They account Received 8 November 2018
for 73% of the total number of road fatalities. Children are also among these victims. To improve Accepted 4 May 2019
countermeasures and design of protection equipment, understanding the biomechanics of motor-
KEYWORDS
cycle users under impact conditions is necessary. The objective of this work is to analyze the over-
Motorcycle accident;
all kinematics and injuries sustained by riders and child pillion passengers in various accident multibody simulation; finite
configurations. element simulation; rider
Methods: Motorcycle accident data were analyzed. Common accident scenarios and impact kinematics; child pillion
parameters were identified. Two numerical approaches were employed. The multibody model was passenger;
validated with a motorcycle crash test and used to generate possible accident cases for various injury mechanisms
impact conditions specified to cover all common accident scenarios. Specific impact conditions
were selected for detailed finite element analysis. The finite element simulations of motorcycle-to-
car collisions were conducted to provide insight into kinematics and injury mechanisms.
Results: Global kinematics found when the motorcycle’s front wheel impacts a car (config-MC)
highlighted the translation motion of both the rider and passenger toward the impact position.
The rider’s trunk impacted the handlebar and the head either impacted the car or missed. The
hood constituted the highest head impact occurrence for this configuration. The child mostly
impacted the rider’s back. Different kinematics were found when car impacted the lateral side of
the motorcycle (config-CM). Upper bodies of both rider and child were laterally projected toward
the car front. The windshield constituted the highest proportion of head impacts. The hood and
A-pillar recorded a moderate proportion. The rider in finite element simulations with config-MC
experienced high rib stress, lung strain, and pressure beyond the injury limit. A high head injury
criterion was observed when the head hit the car. However, the simulation with config-CM exhib-
ited high lower extremities stress and lung pressure in both occupants. Hyperextension of the
rider’s neck was observed. The cumulative strain damage measure of the child’s brain was higher
than the threshold for diffuse axonal injury (DAI).
Conclusions: This study revealed 2 kinematics patterns and injury mechanisms. Simulations with
config-MC manifested a high risk of head and thorax injury to the rider but a low risk of severe
injury to the child. Thorax injury to the rider due to handlebar impact was only found in simula-
tions with config-MC. However, a high risk of skull, lower extremity, brain, and neck injuries were
more pronounced for cases with config-CM. A high risk of DAI was also noticed for the child. In
simulations with config-CM the child exhibited a higher risk of severe injury.

Introduction
In 2015, the World Health Organization (WHO 2015)
reported that 73% of road deaths in Thailand involved riders
and pillion passengers of 2- and 3-wheeled vehicles, which is
6 times that of other vehicles. Motorcycles are considered a
convenient and cheap mode of transportation. The
Department of Land Transport, Thailand (2017), reported
20.5 million registered motorcycles in 2017. Small motorcycles
with a capacity less than 125 cc accounted for 86% of motor-
cycle domestic sales in 2016 (Board of Investment Thailand
2016). They are often used as daily transportation to send
children to schools. Save the Children Thailand (2014)
reported that there are 1.3 million child motorcycle pillion
passengers. In 2009, it was reported that 62% of traffic injuries

CONTACT Julaluk Carmai Julaluk.c@tggs.kmutnb.ac.th Automotive Safety and Assessment Engineering Research Centre, The Sirindhorn International
Thai–German Graduate School of Engineering, King Mongkut’s University of Technology, North Bangkok, 1518 Pracharat 1, Wongsawang, Bangsue, Bangkok
10800, Thailand.
Associate Editor Giovanni Savino oversaw the review of this article.
Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/gcpi.
Supplemental data for this article can be accessed on the publisher’s website.
ß 2019 The Author(s). Published with license by Taylor & Francis Group, LLC
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/4.
0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in
any way.
S14 J. CARMAI ET AL.
in Thailand were among children under 15 years old and
were due to motorcycle accidents (WHO, Regional Office for
South-East Asia 2015). In addition, 37% of motorcycle injuries
in Thailand occur to the head and neck (WHO, Regional
Office for South-East Asia 2015). In order to improve counter-
measures, understanding of kinematics and injury mechanisms
of rider and child passengers during a crash is necessary. This
article aims to analyze the overall kinematics and injuries sus-
tained by motorcycle riders and child pillion passengers in
various accident configurations.
Since 2002, there has been only one complete report on in-
depth motorcycle accident investigation in Thailand
(Kasantikul 2002). A total of 723 motorcycle accident cases in
Bangkok were investigated. The motorcycle speed was between
20 and 50 km/h and the speed of other vehicles was between
11 and 40 km/h. Recently, Karnjanapollert et al. (2018) recon-
structed motorcycle accidents in Thailand. The range of
speeds found from the reconstructions was 30–60 km/h.
However, up-to-date, in-depth motorcycle accident data are
rare in Thailand. Macro accident data are inadequate to pro-
duce effective vehicle-related safety solutions or countermeas-
ures. Multibody simulations have often been used to virtually
generate pedestrian and cyclist accidents, which subsequently
provides an understanding of the kinematics and injury mech-
anisms of pedestrians/cyclists (Bourdet et al. 2014; Fiest et al.
2009; Magriet et al. 2012; Untaroiu et al. 2010; Watson et al.
2006). Utaroiu et al. (2010) developed an advanced method-
ology that combined multibody simulations and an optimiza-
tion technique to identify the adult and child pedestrian pre-
impact postures and vehicle speeds. Fiest et al. (2009) also
used combined approaches of multibody simulation and finite
element simulation to investigate the relevance of rotation-
induced head injuries to pedestrians hit by a large flat-front
vehicle. In addition, Watson et al. (2006) and Magriet et al.
(2012) employed a multibody dynamics model as a tool to
study cyclists’ kinematics during crash. They found that the
head impact of the cyclist was further rearward than that of
the pedestrian. The head impact angles were also different,
although the head impact velocity was similar. Bourdet et al.
(2014) simulated a large number of pedestrian and bicyclist
accident scenarios involving small electric vehicles using multi-
body simulations. In relation to physical or virtual motorcycle
crash tests, ISO 13232 (International Organization for
Standardization [ISO] 2005) is usually employed to set up
crash configurations (Deguchi 2003; Mukherjee et al. 2001;
Namiki et al. 2003). Seven scenarios are recommended in ISO
13232 (ISO 2005) for assessment of protective devices installed
on a motorcycle. They include 143-9.8/0, 114-6.7/13.4, 413-0/
13.5, 413-6.7/13.4, 412-6.7/13.4, 414-6.7/13.4, 225-0/13.4.
In order to improve the understanding of the kinematics
of the rider and child pillion passenger, a number of acci-
dent cases with various impact conditions are required.
Multibody modeling is an inexpensive computerized tool
suitable for simulating a large number of cases. It was
employed to generate a number of possible accident cases
based on defined impact parameters so that the typical kine-
matic patterns were drawn from these generated data.
Nevertheless, the kinematics obtained were carefully consid-
ered because the vehicle deformation was not well
considered. However, finite element simulation can provide
more accurate results, including detailed injury analysis.
However, it is relatively computer-intensive and not optimal
to simulate many impact conditions. Only 3 simulation con-
ditions were selected for finite element simulations so that
further detailed investigation on occupants’ injuries could be
performed. In addition, whole-body collision simulations
were performed because they provide an understanding of
potential injury risk to various body segments.
Materials and methods
Motorcycle accident data classification formed the basis for
simulation work in this research. Selection of simulation
configurations was based on the analysis of these data.
Multibody simulations were employed to generate additional
possible motorcycle accidents cases based on identified real-
world accident scenarios. Three simulation conditions were
selected based on the results of multibody simulation for
detailed injury analyzes using finite element simulation.
Motorcycle accident data classification and common
accident scenarios
Motorcycle accident data were provided by a Thai insurance
company that insured 80% of registered motorcycles. They
were analyzed in terms of frequency and pattern. All data
were classified according to type of crash, type of other vehicle
involved, rider and passenger age range, injury severity level,
and injured body region. The classification of data for each
category is shown in Figures A1–A6 (Appendix A, see online
supplement). The total number of motorcycle accidents in
2014–2015 was 615,542, and 58.7% of motorcycle accidents
were due to the rider losing control. Sideswipe/angular colli-
sion was the second highest, accounting for 30.5%. The num-
ber of rear-end (2.3%) and head-on (0.003%) collisions was
low. Accidents with passenger cars (40.6%) were most com-
mon, followed by collisions with pickup trucks (14.9%).
Common age ranges for the rider were 14–20 years (22.9%)
and for the passenger, 10–19 years old (30.4%). Around 20%
of victims sustained serious and severe injuries. Fatalities
accounted for 1.12%. Head injury constituted 50.3% of all
fatalities. Common accident scenarios were identified from
sideswipe/angular collision because they accounted for the
highest portion when considering only fatalities and severe/
serious injuries. Common scenarios included motorcycle/other
vehicle changing lane or overtaking, motorcycle/other vehicle
making a U-turn, and turning at a junction or turning into an
alley across the path of the other. These are summarized in
Table A3 (Appendix A, see online supplement). The identified
scenarios were also classified into 2 impact configurations,
including (a) motorcycle impacting other vehicle (config-MC)
and (b) other vehicle impacting motorcycle (config-CM).
Multibody motorcycle–car collision model setup
A motorcycle–car collision model was developed using the
multibody simulation program PC-Crash. The child pillion
passenger was also included in the model because the

Table 1. Multibody simulation conditions.
Parameter Motorcycle impacting car (config-MC)
Impact position Lateral
P1 (half of car length)
P2 (one-sixth of car length)
Impact angle (anticlockwise from axial axis of car) 30
, 45
, 60
, 90
, 120
, 150

Motorcycle impact speed (km/h)
Car impact speed (km/h)
Total
accident data showed a significant number of child pillion
passengers (Figure A4) in motorcycle accidents. First the
impact configuration was set to be the same as that in the
experiment of Koetniyom et al. (2018). The motorcycle was
set to impact the car at 42.2 km/h. The height and weight of
the rider and child were set to be same as those of the 50th
percentile rider dummy (1.75 m, 78 kg) and the P6 child
dummy (1.17 m, 21 kg) respectively (Figure A8, see online
supplement). The simulation results were compared to the
experimental results for validation.
Multibody simulation configurations and conditions
The validated motorcycle–car collision model was employed
to generate more accident scenarios. This enabled a study of
the effects of the parameters that are insufficiently available
in real-world situations, such as impact speed, position, and
angle, which encouraged further investigation into rider and
passenger kinematics as well as head impact location during
a crash. Because the anthropometry of Thai population is
smaller than the 50th percentile dummy, the rider was
scaled to represent an average Thai rider in the age group
14–20 years. A rider with a height of 1.59 m and weight of
52 kg was used for subsequent studies. Two impact configu-
rations obtained from the accident analysis were considered
(Figure A7, see online supplement). Impact parameters
including speed, position, and angle were specified to gener-
ate accident simulations that represented accident scenarios
(Table A3) in the previous section. Impact speeds for both
simulation configurations were selected based on the studies
of Kasantikul (2002) and Karnjanapollert et al. (2018).
Simulation conditions are summarized in Table 1. These
simulation conditions also included some configurations rec-
ommended by ISO 13232 (ISO 2005) such as 413-0/13.4 for
config-MC and 143-9.8/0 for config-CM. However, the
motorcycle speed of 50 km/h was used instead of 48.24 km/h
and 40 km/h was used for the car speed instead of 35.28 km/
h. Furthermore, the impact angle and position recom-
mended in 412-6.7/13.4, 135-114/13.4 were also considered.
Motorcycle impacting car configuration (config-MC)
The front wheel of the motorcycle was set to impact the sta-
tionary car with 4 impact speeds as detailed in Table 1. Four
impact positions (Figure A7a) were selected to represent
accident scenarios when a car was changing lanes, overtak-
ing, turning, or making a U-turn. Furthermore, Kasantikul
(2002) reported that the points of first collision contact on
TRAFFIC INJURY PREVENTION S15
Impact configuration
Car impacting motorcycle (config-CM)
Front F (center of front wheel)
P3 (one-sixth of car width) M (center of motorcycle length)
P4 (half of car width) R (center of rear wheel)
120
, 150
30
, 45
, 60
, 90

20, 30, 40, 50 Stationary
Stationary 20,30,40,50,60
64 conditions 60 conditions
the other vehicle frequently occurred at the front bumper
and the area between the front corner and the front door.
The motorcycle was set to impact the lateral side of the car
at position P1 (half the car length) and P2 (one-sixth of the
car length) with 6 impact angles (Table 1). However, only
120
and 150
impact angles were assigned to position P3
and P4 at the front bumper because an angle of 180
or
more did not represent the scenarios of interest.
Car impacting motorcycle configuration (config-CM)
The car front end was set to impact the lateral side of a sta-
tionary motorcycle (Figure A7b). Three impact positions
(Figures A7b, A8) including front wheel (F), center of motor-
cycle length (M), and rear wheel (R) were selected to represent
accident scenarios when a motorcycle was changing lanes,
overtaking, turning, or making a U-turn. The positions were
selected based on Kasantikul’s (2002) investigation, which clas-
sified the points of first collision on the motorcycle’s lateral
side into front, center, and rear (Figure A9, see online supple-
ment). Four impact angles between 30
and 90
were selected
to represent the scenarios of interest (Table 1).
Finite element motorcycle–car collision model setup
A finite element motorcycle–car collision model was devel-
oped (Figure B1, Appendix B, see online supplement). The
NCAC Toyota model was utilized for the simulations. Meshes
of some components were refined to 3–4 mm (Figure B2, see
online supplement). The revised model was validated against
the crash test data. The motorcycle’s finite element model was
developed using reverse engineering (Appendix B). The
dimensions and weight of the main components were meas-
ured (Figures B3, B4, see online supplement). The thickness of
the motorcycle frame, handlebar, tire, rim, and front fork tube
was also measured and was reflected in the shell elements.
Spokes were modeled using beam elements and the engine
was assumed rigid. The seat was simplified using rubber com-
posite behavior. Material properties were obtained from the
experiments. The dynamic behavior of the shock absorber was
simplified using spring elements and its stiffness was obtained
from the experiment. Details of the model are given in
Appendix B. A validation with the experiments was performed
for the front wheel impact because it is the first point of con-
tact with the other object in a frontal collision (Figures
B8–B10, see online supplement). The whole motorcycle kine-
matics was qualitatively compared with a crash test. Fifth per-
centile and 6-year-old occupant human body models
S16 J. CARMAI ET AL.

(THUMS) were employed. Their heights and weights were

close to those of the rider and passenger models used in mul-
tibody simulations. Both THUMS models were validated in
terms of component response and whole-body kinematics (Ito
et al. 2017; Iwamoto et al. 2007; Kimpara et al. 2005; Shigeta
et al. 2009; Watanabe et al. 2012). The postures of THUMS
were adjusted by forced displacement applied to nodes of
bones of upper and lower extremities. Presimulations were
performed to position each body part but prestrains were not
considered in the main simulations. Details of the THUMS
positioning procedure are illustrated in Appendix C (see
online supplement). Beam elements with 0.2 failure strain
were used to tie the rider’s palms to the handlebar and the
child’s palms to the rider’s hip bone. The motorcycle with the
occupant model was combined with the car model. Gravity
was applied to all models. A friction coefficient of 0.1 was
assumed for the contacts between THUMS and other parts. Figure 1. Comparison of kinematics of the rider and the child passenger during
The friction coefficient for the car–ground–motorcycle was a crash.
assumed 0.3. Two impact configurations, config-MC and con-
fig-CM, were set up (Figures A7, B1).

Finite element simulation conditions and injury indices

Three simulation conditions were selected to represent a
motorcycle impacting a car at the lateral position as well as
the front position and a car impacting a motorcycle. Based
on the multibody simulation results, the impact conditions
that led to severe head and chest injuries were considered
for detailed injury investigation. Illustration of each selected
condition is provided in Table D1 (see online supplement).
Justification of each selection was explained in the Results
section. Injury indices and criteria for each body part were
postulated as summarized in Table D2 (see online supple-
ment). A head injury criterion (HIC15) of 700 was utilized
to evaluate skull fracture (NHTSA 1999). It was calculated
based on the linear acceleration at the head center of grav-
ity. Brain cerebral contusion will occur when the principal
strain on the brain exceeds 0.3 (Margulies et al. 1985).
Furthermore, the risk of diffuse axonal injury (DAI) was
estimated using a cumulative strain damage measure
(CSDM). DAI occurred if the percentage volume of ele-
ments having a maximum principal strain over 0.25
exceeded 42.5% (Takhounts and Eppinger 2003). A 0.81
strain level was assumed as an indicator of risk of cervical
ligament damage (Bass et al. 2007). To postulate on lung
contusion, a pressure index with a ±10 kPa threshold was
utilized (Schaffer et at. 1958). A principal strain of 0.3 was
also used for predicting soft tissue injury (Yamada 1970).
Bone fracture was assumed to occur once the Von Mises
stress reached 130 MPa (Arnoux et al. 2008).
Results
Multibody simulations
Multibody simulation results were first compared with the
experiment. The child’s head impact velocity was 40 km/h in
the simulation and 40.5 km/h in the experiment. Occupant
Figure 2. Typical kinematics and head impact locations of the rider and the
child pillion passenger.
kinematics are shown in Figure 1 and Figures D1 and D2 (see
online supplement). The child’s head impacted the A-pillar
and the rider’s head impacted the front window. The rider’s
head rebounded rather than entering the compartment
because the car was modeled as a rigid body. This slightly
affected the child kinematics after impacting the A-pillar. The
child’s head peak linear acceleration was 789.4 g in the simula-
tion, which is slightly higher than the 777.8 g in the experi-
ment (Figure D3, see online supplement). The difference was
only 1.49%. Overall, the simulation results were comparable
with the experiment during the first impact. The model was
employed for kinematics study. In total, 124 simulations were
conducted for config-MC and config-CM.
Typical kinematics are shown in Figure 2 (see also
Figures D4–D5 and D7–D8, online supplement). Each
impact configuration revealed different global kinematics.
For config-MC, the front wheel of the motorcycle hit the
car, and both rider and child were translated toward the
impact position at the first instance. If the impact angle was
less than 60
, the motorcycle tended to slide forward along
the side of the car. The occupants were then translated fur-
ther away from the wheel’s first contact position (Figure
D4b). If the angle was greater than 60
, the front wheel did
not often slide away from the impact position and the rear
wheel raised up. The occupants translated toward the wheel
impact position (Figure 2, left) until the rider hit the handle-
bar. The lower extremities and abdomen were blocked by
the motorcycle and the upper body and the head hurled for-
ward. The rider’s trunk also impacted the handlebar. It was

observed that the occupant’s head may not strike the car
directly depending on the impact position, angle, and speed.
Simulations with impact position P1 resulted in rider’s head
impact for all speeds and angles, whereas with positions P3
and P4, head impact occurred only at an angle of 150
for
all speeds. Simulations with impact position P2 at a 30

angle showed no head strikes against the car for all speeds,
similar to positions P3 and P4 at a 120
angle. The hood
was the location with the most occurrences (29.7%) of
rider’s head impact (Figure D6, see online supplement), and
the A-pillar, windshield, and front window constituted 12.5,
9.4, and 7.8%, respectively (Figure D6). No impact of the
rider’s head occurred 28% of the time. The head striking the
B-pillar, C-pillar, or rear window was found only at impact
position P1 with 120
and 150
angles. The child’s head
mostly made contact with the rider’s back (62.5%) or the
hood (10.9%). In addition, the child fell down without the
head impacting the car 23.4% of the time.
On the other hand, simulations with config-CM revealed
different kinematics. After the front of the car impacted the
occupants’ legs, their upper bodies were projected laterally
toward the car front (Figure 2, right; Figure D8) and then
rotated before their heads impacted the car. However, if the
impact position was at the front wheel, which is before the
seating position, the car did not impact the occupants dir-
ectly. The motorcycle then rotated (Figure D7b) and no
instances of head impact against the car were observed.
These constituted 40 and 48.3% of simulations with config-
CM for the rider and the child, respectively (Figure D9, see
online supplement). The windshield was the location on the
car most often struck by the rider’s head (43.6%) and the
child’s (30%) head. The hood accounted for 10.9 and 13.3%
for the rider and the child, respectively. The A-pillar ranked
third, with a significant number of head impact occurrences
(10.9% for the rider and 5% for the child).
Finite element simulations
Simulation conditions were selected based on the multibody
simulation results. The first 2 conditions were selected from
config-MC, which represented scenarios when the car was
changing lanes, making a U-turn, or turning right at a junc-
tion across the motorcycle’s path. Because it is possible that
the motorcycle either impacted the front or the side of the
car, one location was taken from the front and one from the
side. The first simulation condition was impact position P3
(front), with a 20 km/h speed and 150
angle. This condition
(MC.1) resulted in the rider’s head hitting the hood (29.7%
head impact occurrences) and the torso contacting the
handlebar (expecting serious thorax injury). The second
condition (MC.2) was impact position P1 (side), with a
40 km/h speed and 45
angle. This condition resulted in the
rider’s head striking the A-pillar, which could lead to serious
head injury. The third condition was selected from config-
CM, which represented the motorcycle changing lanes, over-
taking, turning, or making a U-turn across the car’s path.
The simulation with impact position at the rear wheel of the
motorcycle at 40 km/h and with a 45
angle (condition
CM.3) revealed the rider’s head impacting the windshield
TRAFFIC INJURY PREVENTION S17
Figure 3. Simulations of config-MC with condition MC.1 (position P3, angle
120
, speed 20 km/h).
(close to the A-pillar) and the child’s head striking the A-
pillar, for which severe head injury is expected. All selected
simulation conditions are summarized in Table D1
(Appendix D, see online supplement).
The overall kinematics and head impact locations
obtained from the multibody simulations and the finite
element simulations were compared for each corresponding
condition (Figures D10, D11, see online supplement).
Kinematics through time are illustrated in Figures D12–D14
(see online supplement). Global kinematics were similar but
slightly different head impact locations were observed. The
difference is due to the deformable motorcycle and the car
structure in the finite element simulation.
For condition MC.1, the finite element simulation exhibited
no head strike against the hood due to the direct impact of the
torso and handlebar (Figure 3; Figure D10a). The torso was
blocked and slowed down and the lower extremities did not
make contact with the car. In multibody simulation, the right
side of the handlebar contacted the torso and then rotated.
This allowed the rider to move forward and hit the hood.
Thorax injury was very likely to occur (Figure 3), due to
the direct impact of the torso against the handlebar. Three
ribs experienced stress that exceeded the 130 MPa threshold
for fracture (Table D2). Seventy-six percent of lung volume
had internal pressure exceeding the postulated injury criteria
(Schaffer et al. 1958). It was also found that 78% of heart
volume had a strain exceeding the 0.3 criteria for soft tissue
injury (Yamada 1970). The risk of rider’s head injury was
low (HIC15 ¼ 23.5, maximum brain strain ¼ 0.19) because
the head did not hit other objects. Lower extremity stress
was low for both occupants. Because the child struck the
rider’s back, the HIC15 was only 59. It was observed that a
small portion of the child’s brain (0.5%) near the brain stem
experienced strain that exceeded 0.3. CSDM-0.25 was only
2%. Although the child’s rib stress did not exceed the postu-
lated fracture limit, almost 60% of the lung volume had
pressure beyond the injury threshold. The child’s maximum
lung strain was as high as 1.2.
For condition MC.2, the rider’s head hit the lower edge
of the A-pillar in finite element simulation and hit the upper
S18 J. CARMAI ET AL.
Figure 4. Simulations of config-MC with condition MC.2 (position P1, angle
45
, speed 40 km/h).
edge of the A-pillar in multibody simulation (Figure 4;
Figure D10b). The head impact speed was lower in finite
element simulation due to the deformation of both vehicles.
The motorcycle front wheel slightly deformed and slid along
the car and the bodies made contact with the side of the
car. The rider’s head hit the A-pillar, resulting in a high
HIC15 (1,344), but brain strain was lower than 0.3. The
child’s head hit the rider’s back, resulting in a low HIC15 of
173; 2.7% of brain volume had strain exceeded 0.3 limit.
Nevertheless, the CSDM level was insignificant. Lower
extremity stress was less than the postulated fracture limit
for both occupants. The occurrence of thorax injury for
both occupants was likely because the right part of the
upper body impacted against the door. Two right ribs of the
rider had a tendency to break because the stress exceeded
the limit. In addition, 72.5% of the lung volume had internal
pressure exceeding the injury criteria (Figure 4). Heart and
lung strain, which exceeded the 0.3 threshold, accounted for
almost 78% of volume. Similarly, in the child’s thorax, one
rib (no. 8) had stress just beyond 130 MPa; in addition, 60%
of lung volume had pressure exceeding the limit. The child’s
maximum lung strain was also high (2.2).
For condition CM.3, the head impact location of both
occupants in finite element simulation slightly shifted to the
left of the occupants (Figure D11). The rider’s head location
moved closer to the A-pillar and the child’s head impacted
the outer edge of the A-pillar. The head impact speed was
slightly lower in the finite element simulation.
The hood leading edge (HLE) firstly made contact with
the lower extremities, resulting in high stress above the frac-
ture limit on both the rider’s tibia and femur. However, the
child’s femur did not directly impact against the HLE, which
struck the right leg below the knee. Furthermore, because
the right leg was trapped between the car and the motor-
cycle, the child’s left tibia was extended outward by the
frame of the motorcycle. The bending in the left tibia led to
stress beyond the fracture limit (Figure 5). After the first
contact, the pelvis of both occupants ran onto the hood.
Both upper bodies twisted and moved toward the car front.
Figure 5. Simulations of config-CM with condition CM.3 (position R, angle 45
,
speed 40 km/h).
As a result, the child’s right back hit the fender, followed by
the head hitting the A-pillar. The child’s HIC15 was 2,127,
leading to high risk of skull fracture (NHTSA 1999).
Maximum brain strain (emax ¼ 0.86) was higher than 0.3.
CSDM-0.25 was 48%, which implied the possibility of DAI.
In addition, 72.5% of lung volume had pressure beyond the
injury criteria (red area in Figure 5). After the hip slid on
the hood, the rider’s elbow struck the windshield, followed
by the rider’s back and the head; 62% of the lung had pres-
sure exceeding the injury criteria. Only 18% of the lung had
strain beyond 0.3. The rider’s HIC15 was 1,518, and a high
risk of skull fracture was expected. Brain strain was insig-
nificant. Because the initial shoulder impact with the car
arrested the movement of the rider’s torso, the momentum
of the head swayed it backwards before contacting the wind-
shield. The swinging motion led to neck hyperextension
(Figure 5). High strain was found on cervical discs C2–C3
(emax ¼ 0.625). However, this value is still less than the
damage criterion of 0.81 (Bass et al. 2007).
Discussion
The simulation results revealed 2 different overall kinematics
and injury mechanisms for each impact configuration.
Typical global kinematics found in the simulations with con-
fig-MC involved initial forward translation of the rider and
the passenger as well as the motion of the motorcycle after
the first collision contact of the front wheel. It was observed
that when the front wheel impacted the car at a small
impact angle (less than 60
), the motorcycle rotated on a
vertical axis, causing the occupants to be thrown sideways
flat against the car. The rider’s head impact location
depended on the impact position. The impact position near
the front end tended to result in no head impact against the
car. The main injury mechanisms were due to body’s side
impact against the car. Lower limbs exhibited a low risk of
bending or fracture because the limbs interacted with a rela-
tively flat door panel, unlike the HLE. Risk of thorax injury
was noticed due to the torso side impact. The risk of the

rider’s skull fracture depended on the location of head
impact. Striking against a stiff part like the A-pillar would
lead to a high HIC and hence possible skull fracture. For an
impact angle greater than 60
, the motorcycle tended to
rotate on a horizontal axis, the rear wheel then raised up,
and both occupants were ejected upwards while translating.
Initial contact was between the rider’s torso and the handle-
bar followed by the rider’s head striking the car if the speed
was high. Thorax and head injuries were highlighted. A high
risk of thorax injury due to impact with the handlebar was
expected. The child passenger translated together with the
rider and frequently experienced a low risk of severe injury
regardless of ground impact. However, high pressure within
the lung was observed due to compressing against the rider’s
back. The location of head impact against the car differed
depending on the impact position, angle, and speed.
Simulations with config-CM exhibited different kinematics.
Both the rider and passenger were laterally projected toward
the car front while the upper body rotated during collision.
Both occupants sustained similar injury patterns. The main
injury mechanisms were similar to those of pedestrians. The
occurrence of lower limb fracture was likely due to the first
collision contact with the HLE. Relatively low rib stress was
found in comparison with config-MC. A bent-knee posture
resulted in relatively less severe injury than the pedestrian’s
stretched leg posture. Hyperextension of the neck was
observed, which is in line with the study of Kasantikul et al.
(2003). A high risk of skull fracture and thorax injury was
likely in config-CM due to direct head and upper body
impact with the car front. Component stiffness of the car is
a main influence on head acceleration due to inertia effect.
Wearing a helmet would definitely help mitigate the head
injury. However, the rate of helmet use is less than 50% in
Thailand. Law enforcement regarding helmet use would help
reduce the number of fatalities. The findings from the cur-
rent study will be useful for improvements in the design of
protective equipment such as helmets and airbag jackets.
The rider’s head impact speed found in this study at a car
collision speed at 40 km/h is higher than the helmet stand-
ard drop test speed (20.6 km/g). Furthermore, a risk of DAI
due to the brain rotation was observed. The rotational effect
has not yet been considered in helmet drop tests. It is chal-
lenging to design a helmet that can mitigate this type of
injury. In addition, the thorax injury due to handlebar
impact implies the need for protective equipment to mitigate
this type of injury.
Limitations
One limitation of this work is that the whole motorcycle
impact response was only qualitatively validated with crash
test data from the literature. However, motorcycle key com-
ponents were separately validated with experiments. The
other limitation is that THUMS organ impact responses
were validated against postmortem human subjects (Shigeta
et al. 2009) but the validity was only confirmed for the
selected force conditions. In addition, this study utilized
CSDM to predict the occurrence of DAI (Takhounts and
TRAFFIC INJURY PREVENTION S19
Eppinger 2003). The validity of CSDM has not yet been fully
proved.
Funding
The authors gratefully acknowledge Proactive Co. Ltd for
support of PC-Crash Software and Toyota Motor
Corporation for research funding of THUMS simulations.
ORCID
Julaluk Carmai http://orcid.org/0000-0002-5530-9625
References
Arnoux PJ, Behr M, Llari M, Thollon L, Brunet C. Injury criteria
implementation and evaluation in FE models applications to lower
limb segments. Int J Crashworthiness. 2008;13:653–665.
Bass CR, Lucas SR, Salzar RS. Failure properties of cervical spinal liga-
ments under fast strain rate deformations. Spine. 2007;32:7–13.
Bourdet N, Luttenberger P, Teibinger A, Mayer C, Winllinger R.
Pedestrian and bicyclists head impact conditions against small elec-
tric vehicle. Paper presented at: International Research Council on
Biomechanics of Injury (IRCOBI); September 10–14, 2014; Berlin,
Germany.
Board of Investment Thailand. Thailand automotive industry: the next
generation. 2016. Available at: https://www.boi.go.th/upload/content/
BOI-brochure%202015-automotive-20150325_70298.pdf. Accessed
October 17, 2018.
Deguchi M. Modeling of a motorcycle for collision simulation. Paper
presented at: 18th International Conference on the Enhanced Safety
of Vehicles (ESV); May 19–22, 2003; Nagoya, Japan.
Department of Land Transport, Thailand. Government open statistical
data. 2017. Available at: http://www.data.go.th. Accessed October 17,
2018.
Fiest F, Gugler J, Arregui-Dalmases C, et al. Pedestrian collisions with
flat-fronted vehicles: injury patterns and importance of rotational
accelerations as a predictor for traumatic brain injury (TBI). Paper
presented at: 21st International Technical Conference on the
Enhanced Safety of Vehicles Conference (ESV); June 15–18, 2009;
Stuttgart, Germany.
International Organization for Standardization. Motorcycles—Test and
Analysis Procedures for Research Evaluation of Rider Crash Protective
Devices Fitted to Motorcycles. Geneva, Switzerland: ISO Secretariat;
2005. ISO 13232:2005. Available at: https://www.iso.org/standard/
37426.html. Accessed February 2, 2019.
Ito K, Tokuyama M, Miyazaki H, Hayashi S, Kitagawa Y, Yasuki T.
Development of child finite element (FE) models and vehicle-to-
pedestrian collision simulations. Paper presented at: 25th Technical
Conference on the Enhanced Safety of Vehicles (ESV); June 5–8,
2017; Detroit, MI.
Iwamoto M, Nakahira Y, Tamura A, Kimpara H, Watanabe I, Miki K.
Development of advanced human models in THUMS. Paper pre-
sented at: 6th European LS-DYNA User’s Conference; May 29–30,
2007; Gothenburg, Sweden.
Karnjanapollert P, Koetniyom S, Carmai J. The use of multibody
dynamics simulations to investigate motorcyclist kinematics and
injuries in accidents. Paper presented at: 14th International
Conference of Automotive Engineering (ICAE); April 2–3, 2018;
Bangkok, Thailand.
Kasantikul V. Motorcycle Accident Causation and Identification of
Countermeasures in Thailand: Volume I. Bangkok. 2002. Available at:
http://www.mosac.eu/public/file/KasantikulMotorcycleAccident
ResearchinThailand-Bangkok2002.pdf. Accessed October 17, 2018.
Kasantikul V, Ouellet JV, Smith TA. Head and neck injuries in fatal
motorcycle collisions as determined by detailed autopsy. Traffic Inj
Prev. 2003;4:255–262.
S20 J. CARMAI ET AL.
Kimpara H, Lee JB, Yang KH, et al. Development of a three-dimen-
sional finite element chest model for the 5th percentile female.
Stapp Car Crash J. 2005;49:251–269.
Koetniyom S, Carmai J, Kassim KAA, Ahmad Y. Kinematics and injury
analysis of front and rear child pillion passenger in motorcycle
crash. International Journal of Automotive and Mechanical
Engineering. 2018;15:5522–5534.
Magriet VS, Stefanie H, Radarius C, Fredriksson R. Cyclist kinematics
in car impacts reconstructed in simulation and full scale testing with
Polar dummy. Paper presented at: International Research Council
on Biomechanics of Injury (IRCOBI); September 12–14, 2012;
Dublin, Ireland.
Margulies SS, Thibault LE, Gennarelli TA. A study of scaling and head
injury criteria using physical model experiments. Paper presented at:
International Research Council on Biomechanics of Injury
(IRCOBI); June 24–26, 1985; Gothenburg, Sweden.
Mukherjee M, Chawla A, Mohan D, Singh M, Sakurai M, Tamura Y.
Motorcycle–car side impact simulation. Paper presented at:
International Research Council on Biomechanics of Injury
(IRCOBI); October 10–12, 2001; Isle of man, UK.
Namiki H, Nakamura T, Lijima S. A Computer Simulation for Motorcycle
Rider-Motion in Collision. Pennsylvania: SAE International; 2003.
SAE Technical Paper 2003-32-0044.
NHTSA. Development of improved injury criteria for the assessment
of advanced automotive restraint systems–II. 1999. Available at:
https://www.nhtsa.gov/sites/nhtsa.dot.gov/files/rev_criteria.pdf. Accessed
September 10, 2018.
Save the Children Thailand. The 7% Project. 2014. Available at: http://
www.7pecent.org/about/. Accessed March 19, 2018.
Schaffer KE, McNulty WP, Carey C, Liebow AA. Mechanisms in devel-
opment of interstitial emphysema and air embolism on decompres-
sion from depth. J Appl Physiol. 1958;13:15–29.
Shigeta K, Kitagawa Y, Yasuki T. Development of next generation
human FE model capable of organ injury prediction. Paper pre-
sented at: 21st International Technical Conference on the Enhanced
Safety of Vehicles (ESV); June 15–18, 2009; Stuttgart, Germany.
Takhounts EG, Eppinger RH. On the development of the SIMon finite
element head model. Stapp Car Crash J. 2003;47:107–133.
Untaroiu C, Cradall JR, Takahashi Y, Okamoto M, Ito O, Fredriksson
R. Analysis of running child pedestrians impacted by vehicle using
rigid body models and optimization techniques. Saf Sci. 2010;48:
259–267.
Watanabe R, Katsuhara T, Miyazaki H, Kitagawa Y, Yasuki T.
Research of the relationship of pedestrian injury to collision speed,
car-type, impact location and pedestrian sizes using human FE
model (THUMS version 4). Stapp Car Crash J. 2012;56:269–321.
Watson J, Hardy R, Kayvantash K. Understanding the nature of
cyclist’s head impacts. Paper presented at: International Research
Council on Biomechanics of Injury (IRCOBI); September 20–22,
2006; Madrid, Spain.
World Health Organization. Global Status Report on Road Safety 2015.
2015. Available at: https://www.who.int/violence_injury_prevention/
road_safety_status/2015/en/. Accessed September 26, 2017.
World Health Organization, Regional Office for South-East Asia. Child
development and motorcycle safety. 2015. Available at: http://www.
who.int/iris/handle/10665/173782. Accessed March 16, 2017.
Yamada H. Strength of biological materials. In: Evean FG, ed.
Baltimore, MD: Williams & Wilkins; 1970:106–204.