Content uploaded by Per Lovsund
Author content
All content in this area was uploaded by Per Lovsund on Jan 13, 2015
Content may be subject to copyright.
Yang et al., p1
SPEED LIMIT IN CITY AREA AND IMPROVEMENT OF VEHICLE FRONT DESIGN
FOR PEDESTRIAN IMPACT PROTECTION– A COMPUTER SIMULATION STUDY
Jikuang Yang
Xuejun Liu
Per Lövsund
Crash Safety Division
Chalmers University of Technology, Sweden
Claes Tingvall
Anders Lie
Peter Larsson
Thomas Lekander
Swedish National Road Administration, Sweden
Paper number: 232
ABSTRACT
This paper presented a part of results from an
ongoing project for pedestrian protection, which is
carried out at Chalmers University of Technology in
Sweden. A validated pedestrian mathematical model
was used in this study to simulate vehicle-pedestrian
impacts. A large number of simulations have been
carried out with various parameters. The injury-related
parameters concerning head, chest, pelvis and lower
extremities were calculated to evaluate the effect of
impact speed and vehicle front structure on the risk of
pedestrian injuries. The effect of following vehicle
parameters was studied: stiffness of bumper, hood
edge, hood top, windscreen frame, and shape of vehicle
front structures. A parameter study was conducted by
modeling vehicle-pedestrian impacts with various sizes
of cars, mini vans, and light trucks. This choice
represents the trends of new vehicle fleet and their
frequency of involvement in real world accidents. The
mechanical properties of the vehicle front were based
on the available data from EURO NCAP tests, and
from published literature.
Based on the results from the simulation study,
possible benefits from speed control in urban area can
be assessed. As the impact speed decreases from the 40
to 30 km/h, the probability of severe head injury will
decrease from 50% to lower than 25%.
The influences of the various compliance and
geometric parameters of vehicle front are analyzed.
The most significant parameters to pedestrian impact
protection are clarified, especially for head and lower
extremities. A procedure in new vehicle-front design is
presented, which can lead to a design guideline of safer
vehicles for pedestrians.
Furthermore, gaps in pedestrian protection are
identified, and the research priorities should be focused
on the adult head and lower extremities and child head
and thorax injuries.
INTRODUCTION
Among all road user categories, pedestrians are the
most vulnerable road users since they are unprotected
in a vehicle impact. Each year thousands of pedestrians
are killed or injured in road traffic accidents over the
world. Even though the incident of pedestrian fatalities
has dropped in most of the highly motorized countries
during the past two decades, the number of pedestrian
fatalities in road-vehicle accidents is still high. In the
European Union (EU) more than 7000 pedestrians are
killed each year (EEVC, 1998). The annual pedestrian
fatalities vary from about 3000 in Japan, 5500 in the
USA (NHTSA, 1998), and 19000 in China (TAPSM,
1997). The proportion of pedestrian fatalities in all
killed road users is 18.8% in the EU (ETSC, 1999).
Within the EU countries, the relative frequency of the
pedestrian fatalities varies remarkably from 14% in
Sweden to 32% in UK. Huge economic losses and
serious consequences result from these traffic
accidents. Pedestrian protection is therefore a priority
item in traffic safety strategies (EEVC, 1998; ETSC,
1999). Research into injury mechanisms of pedestrians
in vehicle accidents and counter-measures has been
widely performed, but little improvement of vehicle for
pedestrian safety has been made. There is a need to
develop effective safety countermeasures based on
knowledge of pedestrian responses and injury
mechanisms in vehicle accidents.
Since the 1970’s, extensive research has been
carried out in the area of pedestrian protection to
Yang et al., p2
determine the causes of accidents and how to avoid
them, as well as the causes of injuries and means of
reducing them. Many studies on injury mechanisms,
tolerance levels, influences of the vehicle design on
impact responses, protection assessment techniques,
and safety countermeasures have been carried out with
pedestrian substitutes such as biological specimens,
mechanical dummies and mathematical models
(Aldman et al., 1985; Cavallero et al., 1983; Cesari et
al., 1994). The impact speed and vehicle front
structures including geometry and stiffness have been
shown to be important injury-producing factors.
The main factor for measuring the severity of
vehicle-pedestrian impacts is the impact speed. In
approximately 70% of crashes, the driver braked before
the pedestrian was hit (Ashton, 1982; MacLaughlin et
al., 1987). Almost 95% of all pedestrian accidents
occurred at an impact speed lower than 50 km/h as
shown in Figure 1. Pedestrians struck at impact speeds
less than 25 km/h usually sustain minor injuries.
Serious injuries occur frequently at speeds of 25 to 55
km/h whilst at speeds greater than 55 km/h, pedestrians
are most likely to be killed (Ashton, 1982).
0
1000
2000
3000
4000
5000
6000
7000
8000
Number of cases
51015202530354045505560657080
Impact speed (km/h)
Slight injuries
Serious injuries
Fatalities
All injuries
and fatalities
Figure 1. The injury severity distribution as a function of
impact speed (based on Ashton, 1982).
Pedestrians were primarily impacted by the car
front with a high frequency in car-pedestrian accidents.
The European Experimental Vehicle Committee
(EEVC) has therefore proposed test procedures to
evaluate the fronts of passenger cars for pedestrian
safety (EEVC, 1998). The EEVC proposal recommend
impact tests to the car fronts with subsystem impactor
representing segments of the human body. The
subsystem test procedures can be implemented to
detect the vehicle front local stiffness and impact
energy that are main factors to cause pedestrian
injuries.
In order to develop a new vehicle with pedestrian
friendly front that can meet the requirements of the
subsystem tests it is necessary to have an effective
approach for the new vehicle front design to minimize
the risk of pedestrian injury in an unavoidable accident.
This paper described an approach to investigate the
influences of impact speed and vehicle front structures
on pedestrian injuries.
METHOD AND MATERIAL
The approach described in this paper by using
mathematical model can be one of phases in a general
strategy to find an effective counter-measure for
pedestrian protection. The general strategy consists of
three phases.
Phase I
The first phase is to develop a global mathematical
model that includes system definition, model
development and validation.
The system in an accident of pedestrian impacted
by a car front consists of exterior parts of car front,
road, and pedestrian victims. The elements involved in
vehicle-pedestrian impacts can be identified through an
in depth study of accident data.
Figure 2. Distribution of injuries to an adult pedestrian in
frontal car-pedestrian collisions, trajectories of
the head with respect to small and big cars,
changes of the locations of the head impact at
varying impact speeds.
Figure 2 shows the whole system in a car-
pedestrian accident and the distribution of the
pedestrian injuries that are influenced by impact speed,
car front shape and stiffness, and age, length, size of
the pedestrian, as well as standing position of the
pedestrian relative to the car.
The Configuration of the Pedestrian Model - A
human-body mathematical model was developed at
Chalmers University of Technology in Sweden by
using MADYMO program (Yang, 1997; Yang et al.,
2000). It was validated against impact tests with
postmortem human subject (PMHS) and used as a
pedestrian substitute for simulation of car-pedestrian
impacts (Figure 3). Important injury-related parameters
can be calculated by means of the model, including
impact forces, accelerations for different body
Yang et al., p3
segments, HIC, transverse dislocation and contact
forces between articular surfaces, knee-ligament strain,
and knee-bending angle. The leg fracture can be
predicted by using a frangible joint defined in the
breakable leg segments. It is therefore considered a
valuable tool to predict the risk of pedestrian injuries in
accidents. The model is also to be used for a parameter
study on improvement of vehicle-front design for
pedestrian safety.
Figure 3. The baseline model set-up for car-pedestrian
simulations, BCH=Bumper Central Height,
BL=Bumper Lead Length, HEH=Hood-Edge
Height, HL=Hood Length, HA=Hood Slope
Angle, WA=Windscreen Angle.
The car front model consists of bumper, hood
edge, hood top and windscreen as ellipsoids to
approximate the exterior profile of vehicle. The four
wheels are represented by four identical ellipsoids to
produce the braking deceleration of the vehicle by the
friction force between wheels and ground. The vehicle
model is pitched to provide a geometric attitude
equivalent to 0.6g braking. Force-deformation
characteristics of car front components are obtained
from published data and EURO NCAP tests with
headform and legform impactors. The friction
coefficient is 0.6 for foot/ground and wheels/ground,
0.5 for contact between body segments and car front
structures.
Phase II
The second phase is an early stage design that can
begin with evaluation of vehicle-front structures using
mathematical model and subsystem tests. The problem
with existing vehicle can be identified. This is then
followed by optimizing the design and comparing with
the original solution to minimize the risk of pedestrian
injuries. The mathematical model developed in the first
phase is used for assessment and optimization of
vehicle front structure. The important vehicle elements
responsible for pedestrian injuries are summarized in
Table 1.
With the selected variables of vehicle front, it will
cover the vehicle types for small and big size passenger
cars, minivan, and light truck (Figure 3).
Definition of the system covers the involved
elements in vehicle-pedestrian accidents and also the
determination of the injury related parameters that can
be calculated in mathematical modeling (Table 2) and
used for evaluation the aggressiveness of the vehicle
front parts.
Table 1: The vehicle front parts and variables
for parameter study
Parts Geometry Varying stiffness
area
Bumper height
lead distance
width (in vertical
direction)
middle
side (bumper
assembly point)
Bonnet Front edge height
Length
Angle
front edge
top
bonnet fender
area
Windscreen Angle edge
Windscreen
frame A-pillar
roof frame
Fender fender top
Table 2
Injury related parameters in a lateral loading
Parameters Body
segments Tolerance levels
Force tibia
knee
femur
pelvis
4 kN*
2.5 kN (shear)
[4] kN*
4 kN (female)/
10 kN (male)
HIC adult
child 1000*
[1000]*
Linear acc head
thorax
tibia
80 g
60 g
[150] g*
Angular acc head [3000] rad/s2
Rotation angle Knee
Neck 15 degree*
[60] degree
Bending
moment knee
tibia
femur
350 Nm
200 Nm
220 Nm*
Translocation knee 6 mm*
* Acceptance levels of EEVC proposal. In [ ] Data
need a confirmation.
Yang et al., p4
Phase III
The third phase, for the design and building of
vehicle, consists of choosing the best technological
solution, fitting it with optimized definition, and
building the prototype. The prototype of the vehicle
front structure should be tested to evaluate the validity
of the proposed solution.
PARAMETER STUDY
Passenger cars are most frequently involved in
vehicle-pedestrian accidents (Ashton, 1982; Otte,
1989; ETSC, 1999). In the EU countries, the number of
pedestrians struck by passenger cars is around 60% to
80% of the reported vehicle-pedestrian accidents. In 80
- 90% of the cases the pedestrians were hit from the
side. The pedestrians were primarily impacted by the
car front with a frequency of 80%. Therefore a
parameter study was carried out with selected car front
variables. To evaluate the injury risks of pedestrians at
different impact speeds, it’s desirable to take into
account of the distributions of the different vehicle
types involved in pedestrian injury accidents.
Therefore four vehicle types have been simulated in
this study, as described in Table 5 (APPENDIX).
The parametric study involving various variables
such as impact speed, vehicle front shapes and
compliance properties is conducted with the validated
pedestrian mathematical model (Yang, 1997; Yang et
al., 2000).
Design of Parametric Study
In present study the parametric study has been
divided into three parts. The first part concerns the
influence of impact speed, taking into account of
involvement of different vehicle models including
large and compact passenger car, mini van and light
trucks. The main purpose is to predict the effect of
impact speed on the injury risk of pedestrian exposed
to the real world traffic accidents. The injury patterns
with regard to different vehicle models will also be
compared. Secondly, the effects of variations of several
vehicle front shape parameters on the impact severity
of pedestrian will be discussed with passenger car
models at impact speed of 40 km/h. It is aimed to
reveal the possible improvement of vehicle shape to
mitigate the injury severity of pedestrian. Finally the
influence of force-deformation properties of vehicle
structure will be discussed at impact speed of 40 km/h.
The four levels of impact speed are presented in
Table 3. The selected geometric and stiffness variables
and corresponding levels are listed in Table 4.
Bumper central height are chosen between knee
joint and center of gravity of the lower leg for a 50th
percentile adult male. The hood edge height varies
between hip and knee joint. The hood length are varied
at three levels in order to simulate the different vehicle
types, of which 1200 mm for large passenger car, 700
mm for compact passenger car and 500 mm for van
and light trucks. Hood slope angle also depends on the
specific vehicle models. For instance, hood slope angle
at 10 degree is assigned for passenger cars, whereas 30
degree for mini van and 45 degree for light truck.
Likewise, windshield angle is also varied to fit the
different vehicle models. In the case of mini van and
small passenger car, windshield angle has two levels at
30 and 45 degree. For large passenger car and light
truck, the windshield angle is 30 and 45 degree
respectively. These dimensions are defined in Figure 3.
By varying the bumper height, bumper lead, hood edge
height, hood length and slope angle, and windshield
angle, different vehicle models and front shapes can be
simulated, as listed in Table 5 (APPENDIX).
Table 3. Selected levels of vehicle impact speed
Levels (km/h)
Impact Speed 20 30 40 50
Table 4. Selected Factors and levels for parameter study
Levels
Geometric and Stiffness Factors -1 0 +1
BCH= Bumper Central Height (mm) 300 400 500
BS = Bumper Stiffness (N/mm) 125 250 500
BL = Bumper Lead (mm) 50 100 200
HEH= Hood Edge Height (mm) 600 700 800
HES= Hood Edge Stiffness (N/mm) 200 400 800
HL = Hood Length (mm) 500 700 1200
HTS= Hood Top Stiffness (N/mm) 75 150 300
HA = Hood Slope Angle (deg) 10 25 45
WA = Windshield Angle (deg) - 30 45
WS = Windshield Stiffness (N/mm) 300 600 800
The medium level of force-deformation properties
for different structures are consistent with the input
data of validation simulation (Yang, 2000). Some of
the data, such as bumper, and hood top are obtained
from EURO NCAP subsystem tests according to the
test procedures proposed by EEVC (1994 and 1998).
Hood edge stiffness is estimated within the corridor
reported by Ishikawa (1991). All these stiffness
variables vary from 50% to 200% to simulate the
different impact locations on vehicle front structure.
Simulation Matrix
Table 5 (APPENDIX) shows the simulation
matrix, which is based on the involvement of different
vehicle models (NHTSA, 1998), of which large and
compact passenger car has 9 samples for each (56% in
Yang et al., p5
total), whereas Van/Utilities (17%) and Light Trucks
(17%) has 6 samples for each.
Since hood edge height and bumper center height
have been recognized as two key factors affecting the
pedestrian overall kinematics and the injury severity to
knee joints. These two variables are varied in full
factorial of three levels, which lead to 9 variations for
both large and compact passenger cars as shown in
Table 5. Other variables vary in reverse levels for large
and compact passenger car, so that the corresponding
effects can be obtained through these two blocks. The
geometric and stiffness variables for van and light
truck only vary in two levels due to the obvious higher
and stiffer front profile than passenger cars. The effects
of variables calculated from van and light truck models
serve as a complementary to that from passenger cars.
The influences of geometric variables of vehicle
front end on pedestrian responses are evaluated with
the basic configuration of large passenger car at impact
speed of 40 km/h. The stiffness variables remain
constant at their middle levels as below:
− bumper stiffness: 250 N/mm
− hood lead edge stiffness: 400 N/mm
− hood top stiffness: 150 N/mm
− windscreen stiffness: 600 N/mm
The geometric variables including bumper center
height, bumper lead length, and hood edge height vary
at three levels, which makes 27 runs in full factorial
analysis.
Similarly, the effects of different stiffness variables
on pedestrian injuries are studied at impact speed of 40
km/h with a certain vehicle front shape as follows:
− bumper center height: 400 mm
− bumper lead length: 100 mm
− hood lead edge height: 700 N/mm
− hood length: 1200 mm
− hood slope angle: 10 degree
− windscreen slope angle: 30 degree
To avoid the interference of windscreen to the
pedestrian kinematics, the hood length is chosen at the
upper level of 1200mm. While the windscreen stiffness
remains constant, other three stiffness variables
including bumper stiffness, hood edge stiffness and
hood top stiffness are varied according to three-level
full factorial design, leading to 27 runs totally. The
effects of specific variables and possible interactions
on pedestrian injury can be analyzed without any
confounding effect between these variables.
Selected Injury Parameters
The injury risk of pedestrian is evaluated in terms
of injury parameters and tolerance levels listed in Table
6. Due to the absence of a more appropriate criterion,
the widely accepted HIC (Head Injury Criterion) of
1000 is assigned to predict the resultant head injury.
The chest injury criteria - TTI (Thoracic Trauma
Index) of 85 g has been proposed as the maximum
exposure for adults (NHTSA, 1993). The injury criteria
concerning the lower extremities are primarily based
on the recent report by EEVC (1998). The thigh impact
force of 4 kN represents the 20% AIS2+ injury risk
level of femur fracture, whereas tibia acceleration of
150 g indicates 40% risk of an AIS 2+ lower leg
fracture. Moreover, a proposed 15 degree of lateral
bending angle and 6 mm of knee lateral dislocation
serve as the knee joint tolerance levels.
Table 6. Tolerance levels of selected injury parameters
Injury Parameter Tolerance Level
HIC 1000
Chest Acc. (3ms) 85 g
Pelvis Impact Force 10 kN
Thigh Impact Force 5 kN
Tibia Acc. (3ms) 150 g
Knee lateral Dislocation 6 mm
Knee lateral Bending Angle 15o
RESULTS
The part of the results is presented here, which was
used to evaluate the effects of car-front parameters on
the impact force, the knee-rotation angle, the HIC
(head injury criterion) value, and head impact speed.
Effect of Impact Speed on Risk of
Pedestrian Injuries
Calculated injury parameters are presented in the
form of box diagrams to describe the injury risk of
pedestrian at different impact speeds, in which the
injury risks are expressed by the probabilities of 25%,
50% and 75% with bottom, medium and top horizontal
lines respectively. The effects of different vehicle types
with varying front shape and compliance properties are
included. From the diagrams in Figure 4, it can be seen
that the impact speed has significant influence on all
injury-related parameters.
The HIC value increases steadily with the impact
speed. The probability of HIC value exceeding 1000 is
about 50% at impact speed of 40 km/h, whereas it is
lower than 25% at 30 km/h (Figure 4a). The injury risk
of thigh exhibits a strong dependency on impact speed.
For instance, at impact speed of 30 km/h, there is only
less than 25% of all cases exceeding the tolerance level
of 5 kN, whereas more than 75% at impact speed of 40
km/h (Figure 4b). The injury risks of knee joint are
evaluated by knee lateral bending angle, representing
the bending injury mechanism. The knee joint appears
Yang et al., p6
to be the most vulnerable area under the lateral impact
loading. Even at impact speed of 20 km/h, the injury
risk of knee joint is about 25%, as shown in Figure 4c.
As the impact speed increases up to 40 km/h, the injury
risk reaches almost 50%.
0
1000
2000
3000
4000
Injury Threshold
HIC=1000
504030
20
HIC Value
Impact Speed (km/h)
(a) HIC value.
0
5
10
Injury Threshol d 5 kN
504030
20
Thigh Impact Force (kN)
Impact Speed (km/h)
(b) Thigh impact force.
0
5
10
15
20
25
30
35
Injury Threshol d 15o
504030
20
Knee Lateral Bending Angle (degree)
Impact Speed (km/h)
(c) Knee lateral bending angle.
Figure 4. The injury risks at different impact speeds.
Figure 5 describes the dependency of head impact
speed and angle on vehicle travel speed. These two
items are the essential test conditions for the head
impactor to bonnet top test (EEVC, 1998). It’s clearly
that head impact speed increases proportional with the
vehicle impact speed. The head impact angle varies
from 42 to 73 degree for large passenger car.
10 20 30 40 50 60
10
20
30
40
50
60
70
0.72
1.03
0.73
1.03
1.01
0.85
0.85
1.04
Large Passenger Car
Head/Vehicle = 1:1
Mean value o f impact speed
Head Impact Speed (km/h)
Vehicle Impact Speed (km/h)
(a) Head impact speed for large passenger car.
20 30 40 50
20
40
60
80
Large Passenger Car
Mean value of head impact angle
Head Impact Angle (degree)
Vehicle Impact Speed (km/h)
(b) Head impact angle for large passenger car
Figure 5. The influence of vehicle impact speed on head
impact speed and angle
The considerable variations in calculated injury
parameters at a given impact speed imply that factors
other than impact speed are also important in
determining the injury severity of pedestrians.
Therefore, an intensive analysis involving geometric
and stiffness variables are conducted to reveal the
corresponded effects on pedestrian injury severity.
Effect of vehicle front structure on impact
responses of pedestrians
Vehicle Types - The resultant head velocities are
greatly affected by the vehicle types as shown in
Figure 6. In case of the light truck, resultant head
velocity approximately remains constant during the
first 50 ms after the initial impact, and then decreases
sharply after the head impact with windshield structure
at 10.0 m/s (55ms), slightly lower than the travel speed
of vehicle (11.11 m/s). For the mini van and passenger
cars, the resultant head velocities increase
progressively and reach the maximum at around 13.1
and 13.9 m/s, respectively. The actual head impact
speed against vehicle structure varies from 10.8 m/s
(95 ms) for mini van, 13.1 m/s (105 ms) for compact
passenger car to 10.2 m/s (116 ms) for large passenger
car.
Yang et al., p7
0 50 100 150 200
5
10
15
20
10.0 m/s 10.8 m/s
13.1 m/s
10.2 m/s
Impact Speed 40 km/h
Resultant Head Velocity (m/s)
Time (ms)
Large Pass. Car
Compact Pass. Car
Mini VAN
Light Truck
Impact
Figure 6. Resultant head velocity against different vehicle
types at 40 km/h.
The differences in head velocity changes can be
mainly attributed to the different hood slope angles
(Table 7) of various vehicle types, which lead to
different kinematics of pedestrian after the initial
contact with bumper. Due to the large hood slope angle
of light truck, the pelvis and chest contact with hood
earlier than mini van and passenger cars. Consequently
the pedestrian body has not been rotated but pushed
forward along the direction of vehicle travel speed.
When impacted by mini van or passenger cars, the
pedestrian body is rotated downwards to the hood top,
leading to a considerably rotational movement after the
impact. The resultant head velocities thus increase after
initial impact for both mini van and passenger cars.
The extent of this rotational movement differs with the
slope angle and length of hood. Compared to the large
passenger car, the moment of head impact is earlier for
compact passenger car, which results in higher head
impact speed accordingly.
Table 7. Comparison of injury parameters with different
vehicle types at 40 km/h
Vehicle Types
Injury Paramters Large
Car Comp.
Car Mini
Van Light
Truck
HIC 890 514 1261 1317
Chest Acc. (3ms-g) 24 27 45 59
Pelvis Impact Force(kN) 5 2.8 6.7 7.3
Thigh Impact Force (kN) 6.7 3.6 5.8 6.2
Knee Lateral Angle(deg) 25 12 20 20
Knee Lateral Disl. (mm) 5.1 8.6 5.0 5.0
Table 7 summaries the mean values of various
injury parameters at impact speed of 40 km/h with
respect to different vehicle types. In the simulation
with van and light truck model, the contact locations of
pedestrian head are closed to the lower windshield
frame, where is stiffer than the center of windshield
and hood top. Therefore high injury risk of head has
been found in terms of HIC value. Moreover, large
hood slope angle results in the direct blow on
pedestrian chest and pelvis area, the injury risk to these
body parts are higher than that of passenger cars with
apparent ’slipping’ movement along hood top. The
injury risks to lower extremity of different vehicle
types are compared in terms of thigh impact force and
knee joint injury risk. Generally, pedestrian is exposed
to high injury risk to upper body area against the mini
van and light truck, while the passenger cars are more
aggressive to the lower extremity. Similar tendency has
been found in pedestrian accidents (Mizuno et al.,
2000).
Vehicle Front Shape - To avoid the effects of
hood slope angle and hood length on pedestrian
kinematics, only large passenger cars are considered to
assess the influences of bumper center height, bumper
lead length and hood edge on head and lower extremity
injuries.
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0
0
0.5
1.0
1.5
2.0
Geometric Variations (mm)
HEH BCH BL
600 400 100
800 400 100
Head
Pelvis
Foot
Knee
Vertical Displacement (m)
Horizontal Displacement (m)
(a) Influence of hood edge height on pedestrian kinematics.
50 100 150 200
0
5
10
15
Resultant Head Velocity (m/s)
Time (ms)
Influence of Hoodedge Height
Hoodedge Height Head Velcity
600 mm 11.1 m/s
700 mm 10.2 m/s
800 mm 9.4 m/s
Head Impact
Bumper Height 400 mm
Bumper Lead 100 mm
(b) Influence of hood edge height on resultant head velocity
600 700 800
9
10
11
12
Head Impact Speed (m/s)
Hood Edge Height (mm)
Bumper Center Height (mm)
300 400 500
Bumer Lead Length 100 mm
(c) Interaction between bumper height and hood edge height
on head impact speed.
Figure 7. Influence of vehicle front shape parameters on
pedestrian kinematics and head responses.
Yang et al., p8
Clearly, hood edge height has great effect on
pedestrian kinematics and resultant head velocity.
When hood edge height varies from 800 to 600 mm,
the WAD (Wrap Around Distance) increases from 1.78
m to 2.05 m, due to the considerable ’slipping’ effect of
pedestrian body, as shown in Figure 7a. The head
impact speed increases from 9.4 m/s to 11.1 ms/s
(Figure 7b).
The main effect of hood edge height, and
interaction between bumper center height and bumper
lead length on head impact speed are described in
Figure 7c. It’s clearly that hood edge height has greater
effect than other two parameters. No significant
interaction exists between hood edge height and
bumper center height, as shown in Figure 7c.
In general, the head impact speed tends to decrease
with an increase of hood edge height and a lowering of
bumper center height, but has little relation to bumper
lead length.
300 400 500
0
5
10
15
20
25
30 Bumper Lead Length (mm)
50 100 200
Injury Threshold 15 Degree
Mean Response
Knee Lateral Bending Angle (degree)
Bumper Center Height (mm)
(a) Interaction between bumper center height and bumper lead length
on knee lateral bending angle
300 400 500
2
4
6
8
10
Mean Response
Knee Lateral Dislocation (mm)
Bumper Center Height (mm)
Bumper Lead Length (mm)
50 100 200
Injury Threshold 6 mm
(b) Interaction between bumper center height and bumper
lead length on knee lateral dislocation
Figure 8. Influence of vehicle front shape on knee joint
injuries.
The influences of the vehicle shape variables on
knee joint injuries are described in Figure 8, in which
the possible interactions among bumper center height,
bumper lead length and hood edge height were also
presented.
The results indicate that lowering bumper center
height is favorable to reduce the knee lateral-bending
angle (Figure 8a). For instance, when the bumper
center height decreases from 500 to 300 mm (nearer
the center of gravity of the lower leg) the knee lateral
bending angle was reduced by 67% (15 degree to 5
degree).
However, the knee lateral dislocation is dependent
on both bumper center height and bumper lead length
(Figure 8b). As the bumper height decreases from 500
to 400 mm in case of bumper lead length of 50 mm, the
knee lateral dislocation increased from 5.2 to 6.1 mm.
When the bumper center height decreases further to
300 mm, the knee lateral dislocation decreases to 2.4
mm.
No significant interaction exists between bumper
center height and hood edge height in terms of knee
injury parameters. Therefore, it can be concluded that
the hood edge height has little effect on both knee
lateral-bending angle and lateral dislocation. This
result was consistent with the previous studies
conducted with mechanical substitutes and computer
simulation (Ishikawa et al., 1994; Nagatomi et al.,
1996).
Influence of Vehicle Stiffness Properties The
influence of vehicle stiffness properties is analyzed in
terms of pedestrian kinematics, resultant head velocity
and injuries to head and lower extremity regions.
Figure 9a shows the influence of stiffness
variations of bumper and hood edge on pedestrian
kinematics. Although the stiffness varies between 50%
and 200%, there is slight effect on head trajectory and
speed.
Figure 9b illustrates influence of local stiffness of
contact area on the injury severity of different body
segment. It is clearly that the stiffness variables have
significant effects on head and lower extremity
injuries, with the given vehicle front shape and impact
speed. As shown in Figure 9b, a soft hood top can
provide a significant protection performance to head
injury severity, compared with other two levels of hood
top stiffness. Similar tendency is also found in terms of
thigh (Figure 9c), knee joint (Figure 9d) injury
severity.
Compared with the influence of vehicle shape
variables, stiffness proprieties have great effect on
resulted injury severity, but minor influence on
pedestrian gross motion.
Yang et al., p9
-1.5 -1.0 -0.5 0.0 0.5 1.0
0
0.5
1.0
1.5
2.0
Stiffness Variations (N/mm)
Bumper Hoodedge
125 200
500 800
HEH=700 mm
BCH=400 mm, BL=100 mm
Head
Pelvis
Foot
Knee
Vertical Displacement (m)
Horizontal Displacement (m)
(a) Influence of stiffness properties on pedestrian kinematics,
50 100100 150 200200 250 300300
400
600
800
1000
1200
1400
Mean Response
HIC Value
Hood Top Stiffness (N/mm)
(b) Influence of hood top stiffness on HIC value
200200 400400 600600 800800
2
3
4
5
6
7
Mean Response
Thigh Impact Force (kN)
Hood Edge Stiffness (N/mm)
(c) Influence of hood edge stiffness on thigh impact force.
100100 200200 300300 400400 500500
12
13
14
15
Mean Response
Knee Lateral Bending Angle (degree)
Bumper Stiffness (N/mm)
(d) Influence of bumper stiffness on knee bending angle.
Figure 9. Influence and interactions of vehicle front
stiffness variables on lower extremity injuries
DISCUSSION
The influences of impact speed, vehicle front
shape and compliance on pedestrian responses are
evaluated in terms of the calculated injury parameters.
The mean values of injury parameters are plotted in
terms of different geometric and stiffness parameters.
The main effects of these variables and possible
interactions are examined.
The results from parameter study indicated the
significant effects of the bumper height, bumper
stiffness, bumper-lead distance, and hood-edge height
on responses of the knee-leg complex in a lateral
impact to the leg. The head responses appear to be
dependent primarily on hood-edge height. Injury to the
lower extremities and head is influenced by car front
parameters. Thus it is necessary to perform an
optimization study on new car front design.
The tendency of reduction of the pedestrian
fatalities during past years is mainly due to improved
traffic planning in built-up areas. New aerodynamic car
designs may also have contributed to the reduction of
pedestrian injury. The modern cars with new
aerodynamic design have changes in front-end shape
by rounded bonnet edges, smooth surfaces and a low
bumper which is in accordance with the principle of an
improved car design for pedestrian protection. The
findings from experimental studies (Ashton, 1982;
NHTSA, 1993; EEVC, 1998; ETSC, 1999) suggest that
a potential benefit could be obtained from changes in
car front-end.
Effective counter-measures require knowledge
about pedestrian injuries and injury mechanisms in
accidents. Knowledge about injury mechanisms of
pedestrians in car impacts has mainly been achieved
from tests with PMHS. PMHS tests, however, can not
be used for extensive study of safety counter-measures
due to ethical problems and high cost of such tests.
Several types of pedestrian dummies have therefore
been developed to evaluate new counter-measures. So
far none of them have been found appropriate to
simulate responses of pedestrians in car impacts due to
biofidelity and repeatability problems. The subsystem
test procedure is an important measure to determine the
aggressiveness of car front parts in pedestrian
accidents. The kinematics in car-pedestrian impacts are
quite complex due to successive impacts to the body
segments in a large relative movement between
pedestrian and moving car front. There is a need of a
validated pedestrian mathematical model that can be
used for the prediction of the kinematics behavior and
injury risks of pedestrians as well as for the evaluation
of safety concepts in the early stages of car designs.
Yang et al., p10
Improving pedestrian protection via new
vehicle front design
A general principle for pedestrian protection is to
reduce local stiffness of the car-front components and
minimize impact energy in collisions. It could be
implemented by using energy-absorbing materials,
removing sharp edges and increasing the distance
between the bonnet and engine components, as well as
spreading impact forces over as wide a body area as
possible and preferably over the strongest skeletal
structure. Modifying the construction of the bonnet,
especially the rear part of the bonnet and bonnet-fender
regions, and the under-bonnet clearance provides head
impact protection. Lowering the bumper height to a
suitable level and increasing the compliance and width
of the bumper system can reduce risk of leg and knee
injuries.
Pedestrians may also benefit from other crash
safety strategy such as automatic radar braking
systems, which can detect the presence of an object in
the path of a car and automatically brake a car to a
lower speed.
The third phase in proposed procedure for research
into pedestrian protection have not yet implemented in
the present study. It can be conducted with the findings
from the parameter study by developing a prototype of
pedestrian friendly vehicle. As a positive effect to
pedestrian safety, a pedestrian friendly vehicle can
minimize the risks of pedestrian injuries in
combination with suitable speed limit in city build up
area.
CONCLUSIONS
• The validated pedestrian model based on human
body characteristics is an effective means to study
the complex collision event between vehicle
structure and human body. The results from
parametric study have indicated that the injury
severity of pedestrian is strongly affected by
impact speed and vehicle design, and can be
greatly reduced by altering vehicle front shape and
structure stiffness properties.
• Impact speed has the critical significant effect on
the pedestrian injury severity. As the impact speed
decreases from 40 to 30 km/h, the probability of
severe head injury (HIC > 1000) will decreases
from 50% to lower than 25%, whereas the injury
risks to knee joints remains above 50%.
Considering the possible improvement of vehicle
front structure to mitigate the injury severity to
knee joint, it is concluded that a speed limit of 30
km/h in urban area is effective means to reduce the
pedestrian injury risk.
• The kinematics and injury patterns of pedestrian
vary considerably with the vehicle models. For van
or light truck, the transnational movement of
pedestrian body determines the injury patterns and
distributions, whereas rotational movement is
significant for the passenger cars. Consequently
pedestrian is exposed to high injury risk to upper
body area against the mini van and light truck,
while the passenger cars are more aggressive to the
lower extremity than mini van and light truck.
• As to the kinematics and resultant head velocity of
pedestrian struck by passenger cars, hood edge
height has been identified as the dominant factor.
The effect of bumper center height and bumper
lead length is slight. In general, head impact speed
decreases with an increase of hood edge height and
a lowering bumper center height.
• The injury severity to the knee joint is strongly
influenced by the bumper central height and
stiffness properties. Bumper lead and hood edge
height only has slight effect on lateral bending
angle of knee.
• The local stiffness of head contact area greatly
controls the resulting injury severity of head. The
influence of force-deformation properties of
bumper and hood edge on pedestrian kinematics
and head impact speed is slight, but significant for
the resulting injury severity to the involved body
segments.
Further research is to be focused on:
• Development of a series of pedestrian models that
cover different ages of child pedestrian groups and
to be used for study on child pedestrian protection.
• The protection priorities of child head/neck/thorax
injuries, adult head/neck injuries, as well as leg
and knee injuries.
• Development of a prototype of pedestrian friendly
car based on findings from parameter study.
• Possible speed limit in city area considering a
vehicle fleet designed for pedestrian protection.
ACKNOWLEDGEMENT
This study is sponsored by the Swedish National
Road Administration (Vägverket).
REFERENCES
Aldman, B., Kajzer, J., Anderlind, T., Malmqvist, M.,
Mellander, H. and Turbell, T. (1985). Load Transfer
From the Striking Vehicle in Side and Pedestrian
Impacts. Proc. of the 10th Int. Technical Conf. on
Yang et al., p11
Experimental Safety Vehicle, Oxford, England, July
1-4, pp. 620-637.
Ashton, S. J. (1982). Vehicle Design and Pedestrian
Injuries. Chapter 6, in Pedestrian Accidents. Edited
by Chapman et al. John Wileys & Sons Ltd.
Cavallero, C., Cesari, D., Ramet, M., Billault, P.,
Farisse, J., Seriat- Gautier, B. and Bonnoit, J.
(1983). Improvement of Pedestrian Safety:
Influence of Shape of Passenger Car-Front
Structures Upon Pedestrian Kinematics and
Injuries: Evaluation Based on 50 Cadaver Tests.
SAE Paper 830624.
Cesari, D., Bouquet, R., Caire, Y. and Bermond, F.
(1994). Protection of Pedestrians Against Leg
Injuries. Proc. of the 14th Int. Technical Conf. on
Experimental Safety Vehicle, Paper No. 94-S7-O-
02, Munich, Germany, May 23-26, pp. 1131-1138.
EEVC (1994). Proposals for Methods to Evaluate
Pedestrian Protection for Passenger Cars, Report,
European Experimental Vehicle Committee,
working group 10.
EEVC (1998). Improved Test Methods to Evaluate
Pedestrian Protection Afforded by Passenger Cars,
European Experimental Vehicle Committee,
Working Group 17. Document No. 100.
ETSC (1999). Safety of Pedestrian and Cyclists in
Urban Areas, European Transport Safety Council.
Higuchi, K. and Akiyama, A. (1991). The Effect of the
Vehicle Structure’s Characteristics on Pedestrian
Behavior. Proc. of the 13th Int. Technical Conf. on
Experimental Safety Vehicle, Paper No. 91-S3-O-
10, Paris, France, Nov 4-7, pp. 323-329.
Ishikawa, H., Kajzer, J., Ono, K. and Sakurai, M.
(1994). Simulation of Car Impact to Pedestrian
Lower Extremity: Influence of Different Car-Front
Shapes and Dummy Parameters on Test Results.
Accident Analysis and Prevention, 26(2). pp. 231-
242.
Ishikawa, H., Yamazaki, K., Ono, K. and Sasaki, A.
(1991). Current Situation of Pedestrian Accidents
and Research into Pedestrian Protection in Japan.
Proc. of the 13th Int. Technical Conf. on
Experimental Safety Vehicle, Paper No. 91-S3-O-
05, Paris, France, Nov. 4-7, pp. 281-293.
MacLaughlin, T. F., Hoyt, T. A. and Chu, S. M. (1987).
NHTSA’s Advanced Pedestrian Protection
Program. Proc. of the 11th Int Tech. Conf on
Experimental Safety Vehicles, Washington, May
12-15, pp. 771-776.
Matsui, Y. and Ishikawa, H. (1998). Validation of
Pedestrian Upper Legform Impact Test -
Reconstruction of Pedestrian Accidents. Proc. of the
16 Int. Technical Conf. on Experimental Safety
Vehicle, Paper No. 98-S10-O-05, Ontario,
CANADA, May 31- June 4, pp. 2152-2167.
Mizuno, K. and Kajzer, J. (2000). Head Injuries in
Vehicle-Pedestrian Impact. SAE Paper 2000-01-
0157.
Nagatomi, K., Akiyama, A. and Kobayashi, T. (1996).
Bumper Structure for Pedestrian Protection. Proc.
of the 15th Int. Technical Conf. on Experimental
Safety Vehicle, Paper No. 96-S4-O-02, Melbourne,
Australia, May 13-16, pp. 593-601.
NHTSA (1993). Pedestrian Injury Reduction Research,
National Highway Traffic Safety Administration,
US Dept. of Transportation, Washington DC, USA.
NHTSA (1998). Taffic Safety Facts - Pedestrian,
National Highway Traffic Safety Administration,
Department of Transportation, U.S.A.
Otte, D. (1989). Influence of vehicle front geometry on
the injury situation of injured pedestrians. Road
traffic accident research, Medical University of
Hannover.
TAPSM (1997). Road Traffic Accident Data in 1996
from P.R. China. Traffic Administration of Public
Security Ministry, Report, August 1997.
TNO (1999). MADYMO User's Manual 3D. Version
5.4. Delft, Netherlands, TNO Road-Vehicles
Research Institute.
Yang, J. K. and Kajzer, J. (1992). Computer Simulation
of Impact Response of the Human Knee Joint in
Car-pedestrian Accidents. Proc. of 36th STAPP
Conf., SAE Paper 922525, pp. 203-217.
Yang, J.K. (1997). Injury Biomechanics in Car-
Pedestrian Collisions: Development, Validation and
Application of Human-Body Mathematical Models.
Doctoral thesis. Dept. of Injury Prevention,
Chalmers University of Technology, Göteborg,
Sweden
Yang, J. K., Lövsund, P., Cavallero, C. and Bonnoit, J.
(2000). A Human-Body 3D Mathematical Model
for Simulation of Car-Pedestrian Impacts. J. Crash
Prevention and Injury Control, Vol.2(2), pp.131-
149.
Yang, J. K., Rzymkowski, C. and Kajzer, J. (1993).
Development and Validation of a Mathematical
Breakable Leg Model. Proc. of the IRCOBI Conf.,
Eindhoven, Netherlands, Sept 8-10, pp 175-186.
Yang et al., p12
APPENDIX
Table 5. Simulation matrix for influence of impact speed
Vehicle Types BCH BS BL HEH HES HTS WA WS
-1 1 1 -1 1 1 0 1
0 0 0 -1 0 0 0 0
1 -1 -1 -1 -1 -1 0 -1
-1 1 1 0 1 1 0 1
0 0 0 0 0 0 0 0
1 -1 -1 0 -1 -1 0 -1
-1 1 1 1 1 1 0 1
0 0 0 1 0 0 0 0
Large Passenger Car
HL= 1.2 m, HA= 100
WA=250
1 -1 -1 1 -1 -1 0 -1
-1 -1 -1 -1 -1 -1 1 -1
0 0 0 -1 0 0 0 0
1 1 1 -1 1 1 1 1
-1 -1 -1 0 -1 -1 0 -1
0 0 0 0 0 0 1 0
1 1 1 0 1 1 0 1
-1 -1 -1 1 -1 -1 1 -1
0 0 0 1 0 0 0 0
Compact Passenger Car
HL=0.7 m ,HA=100
1 1 1 1 1 1 1 1
0 0 1 0 0 1 0 1
1 1 0 0 1 0 1 0
0 0 1 0 0 1 0 1
1 1 0 1 1 0 1 0
0 0 1 1 0 1 0 1
Van/Utilities
HL=0.5m, HA=300
1 1 0 1 1 0 1 0
0 0 0 1 0 1 1 1
1 1 1 1 1 0 1 0
0 0 0 1 0 1 1 1
1 1 1 0 1 0 1 0
0 0 0 0 0 1 1 1
Light Truck
HL=0.5 m, HA=450
1 1 1 0 1 0 1 0
