Accident Analysis and Prevention 37 (2005) 633640

Effects of vehicle bumper height and impact velocity on type of

lower extremity injury in vehiclepedestrian accidents
Yasuhiro Matsui
Japan Automobile Research Institute, 2530 Karima, Tsukuba, Ibaraki 305-0822, Japan
Received 10 January 2005; accepted 13 March 2005

Abstract
In nonfatal passenger vehiclepedestrian accidents, the lower extremities are the most commonly injured body parts. The European Enhanced
Vehicle-safety Committee Working Group 17 (EEVC/WG17) pedestrian subsystem test method using a legform impactor has been developed
mainly for evaluation of aggressiveness of the front bumper of passenger vehicles. However, in recent years the number of sports utility
vehicles (SUV) with a high bumper has been rapidly increasing. Since the bumper height is different between a passenger vehicle and an
SUV, the type of lower extremity injury may be different. The type of lower extremity injury caused by this different bumper height should be
clarified, because the test method and vehicle safety countermeasure must take into account a certain type of injury. Furthermore, the effect
of vehicle impact velocity on the type of lower extremity injury in vehiclepedestrian accidents has not been investigated so far. Therefore,
the objective of this study is to clarify the effect of vehicle bumper height and vehicle impact velocity on the type of lower extremity injury in
vehiclepedestrian accidents. The Pedestrian Crash Data Study (PCDS), an in-depth accident database in the USA, was used for the current
analyses. The results indicate that the type of injury, i.e., to the tibia and knee ligament, could become an injury to the femur with an increase
in bumper height. Furthermore, the main injury at an impact velocity of around 2030 km/h is to the knee ligament. On the other hand, the
main injury at an impact velocity of around 40 km/h is a fracture of the lower extremities.
2005 Elsevier Ltd. All rights reserved.
Keywords: Pedestrian accident; Sport utility vehicle; Bumper aggressiveness; Lower extremity injury; PCDS

1. Introduction
Pedestrian protection is one of the key issues in the area
of vehicle safety legislation in Europe and Japan. As leg
injuries from the bumper are the most common injuries in
nonfatal pedestrian accidents (38%) (ITARDA, 2004), current investigation focuses on the accident conditions in vehicle bumper-pedestrian leg injuries. The European Enhanced
Vehicle-safety Committee (EEVC), the National Highway
Traffic Safety Administration (NHTSA), and the International Standard Organization (ISO) have discussed ways to
improve test methods for evaluating vehicle aggressiveness
against pedestrians for many years. The EEVC (WG10, 1994
and WG17, 1997) and ISO (ISO, 1996) are planning to use
a subsystem legform test setup to evaluate mainly the ag

Tel.: +81 29 856 0885; fax: +81 29 856 1121.

E-mail address: ymatsui@jari.or.jp.

0001-4575/$ see front matter 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aap.2005.03.005

gressiveness of the front bumper of passenger vehicles. The

procedure proposed by the EEVC and ISO is to assess the risk
of ligament injury and tibia fracture by means of a mechanical legform impactor representing a leg of an adult male 50th
percentile (AM 50), such as the legform impactor proposed
by the EEVC that was developed by Transport Research Laboratory (TRL) (Lawrence and Hardy, 1998). Matsui (2001,
2003) has proposed injury reference values for evaluation of
ligament injuries and tibia fracture using the TRL legform impactor. The procedure proposed by the EEVC is also available
to assess the risk of femur fracture by means of a mechanical upper legform impactor representing an upper leg of an
AM 50, such as the upper legform impactor proposed by the
EEVC that was also developed by Transport Research Laboratory (TRL) (Lawrence and Hardy, 1998). Injury reference
values for evaluation of femur fracture using the TRL upper
legform impactor were proposed by Rodmell and Lawrence
(1998) and Matsui et al. (1998). In recent years the number

634

Y. Matsui / Accident Analysis and Prevention 37 (2005) 633640

of sports utility vehicles (SUV) with a high bumper has been

rapidly increasing. Since the height of an SUV bumper is
greater than that of a passenger vehicle (Matsui, 2004), the
type of lower extremity injury caused by the SUV bumper
may be different from that caused by a passenger vehicle
bumper. The type of lower extremity injury caused by the
different bumper height should be clarified, because the test
method and vehicle safety countermeasure must take into account a certain type of injury. Roudsari et al. (2004) found
from vehiclepedestrian accident analysis that light truck vehicles were associated with a three times higher risk of severe injuries in comparison with passenger vehicles. On the
other hand, the same investigations pointed out that there
was no difference among light truck vehicles, vans, and passenger vehicles in the percentage of injuries to lower extremities. Ballesteros et al. (2004) revealed that compared to
passenger vehicles, SUVs and pick-up trucks resulted in a
higher percentage of injuries to the lower extremities above
the knee. Basically, the studies by Roudsari et al. (2004) and
Ballesteros et al. (2004) investigated the effect of vehicle type
on the risk of lower extremity injuries. However, the effect
of actual vehicle bumper height on the type of lower extremity injury in vehiclepedestrian accidents has not been
investigated to date.
Takeuchi et al. (1998), Ballesteros et al. (2004) and
Roudsari et al. (2004) also found from vehiclepedestrian
accident analysis that the vehiclepedestrian impact velocity was the key factor in the most severe body injury to the
pedestrian. However, the effect of impact velocity on the type
of lower extremity injury in vehiclepedestrian accidents has
not been investigated so far. Therefore, the two-fold objective of this study is to clarify, based on accident data, both the
effects of vehicle bumper height and vehicle impact velocity
on the type of lower extremity injury occurring in real-world
vehiclepedestrian accidents.

2. Methods
2.1. Selected accident data
To analyze the effect of vehicle bumper height and vehicle
impact velocity on the type of lower extremity injury in realworld vehiclepedestrian accidents, the national accident
sampling system pedestrian crash data study (NHTSA, 1998),
an in-depth accident database in the USA (19941998), was
used. Sixty-two accident cases including information on both
the pedestrian and the vehicle involved were selected for the
present analysis. All of these accidents occurred at an impact velocity ranging from 3 to 59 km/h. The reasons for
selecting cases involving a low impact velocity in the present
study are that the information on both the pedestrian and
the vehicle at high impact velocities over 60 km/h is insufficient for present analyses and that 60 km/h corresponds to
a 1.5 times higher velocity than the 40 km/h targeted by the
EEVC/WG17 (EEVC, 1998). Pedestrians involved in these

Fig. 1. Three types of lower extremity injuries.

selected accidents were either adults with an age range defined as 16 years, or older and children taller than 150 cm.
Here, the present investigator focused on three types of injuries, i.e., femur fracture, knee ligament injury and tibia
fracture scoring more than 2 on the abbreviated injury scale
(AIS) caused by front bumper contact, because the reduction
of such injuries is the aim of the bumper safety evaluation
test proposed by the EEVC/WG17 (EEVC, 1998).
The following cases of injury to the lower extremities were
excluded from the present accident case selection:
(1) cases with tibia malleolus fracture caused by friction with
the road surface;
(2) case of a pedestrian sustaining a femur shaft fracture
despite the fact that the bumper impacted the lower legs.
The injury description and data sample are listed in
Table 1. The types of injury sustained by the lower extremity
were categorized into the following three injury groups: femur fracture, knee ligament injury and tibia fracture as shown
in Fig. 1. A limitation of the present study is the small sample:
5 cases of femur fracture, and 20 and 48 cases, respectively,
of knee ligament injury and tibia fracture (Table 1).
Bumper lower height, bumper center height, and bumper
upper height were used to express bumper height as shown
in Fig. 2, because they differ for every vehicle. Moreover, to

Fig. 2. Three bumper heights.

Y. Matsui / Accident Analysis and Prevention 37 (2005) 633640

635

Table 1
Injury description and data sample
Injury type

AIS code

Injury description

Femur fracture

851814,3
851822,3

Femur shaft fracture

Femur supracondylar fracture

840404,2
850810,2
850814,2
850826,2
853404,2
853406,2
853408,3
853420,2
853422,3

Knee ligament
injury

Tibia fracture

Total

3
2

Knee ligament injury

Knee dislocation without involving articular cartilage
Knee dislocation involving articular cartilage
Knee sprain

6
1
2
11

20

Tibia fracture
Tibia plateau fracture
tibia plateau open/displaced/comminuted fracture
Tibia shaft fracture
Tibia shaft open/displaced/comminuted fracture

1
8
12
6
21

48

express the effective bumper height in relation to the given

pedestrian knee height, the bumper lower height, bumper
center height, and bumper upper height were normalized by
the pedestrian knee height (HNBL ) using Eq. (1), because
knee height differs widely with each pedestrian as shown in
Table 2.
Normalized bumper lower height (HNBL )
=

Number

Bumper lower height

Pedestrian knee height

(1)

2.2. Distribution of three injury types for different

bumper lower height, normalized bumper lower height
and impact velocity
In the present analysis, distributions of the three injury
types (Fig. 1) for different vehicle bumper heights, and
different impact velocities were investigated. For the vehicle bumper height, a bumper lower height and normalized
bumper lower height were used.
2.3. Statistical testing of vehicle bumper heights and
impact velocity for given injury groups
The possible differences in vehicle bumper height and impact velocity between two given injury groups were examined by Wilcoxon test as shown in Fig. 3. The given injuries
were classified into femur fracture, ligament injury and tibia
fracture groups. Wilcoxon test was used to determine any significant difference in the vehiclepedestrian impact velocity

and respective bumper height (bumper lower height, bumper

center height, bumper upper height; and normalized bumper
lower height, normalized bumper center height, and normalized bumper upper height) between the two injury groups.
2.4. Comparative statistical testing of impact velocity
between bone fracture group and knee ligament injury
group at a specied location of lower extremity impacted
by bumper
The effect of vehicle impact velocity on the type of injury at a given normalized bumper lower height was investigated. The investigation started from testing statistical
significance of the differences in impact velocity between
the accidents resulting in femur fracture and those in which
ligament injury occurred. In this testing, only the accidents
with a normalized bumper lower height (HNBL ) not smaller
than 0.857 were taken into account. This is because all femur fractures occurred when HNBL 0.857 (see Table 2).
On the other hand, when testing statistical significance of the
differences in impact velocity between the accidents resulting in tibia fracture and those causing knee ligament injury,
only the cases with HNBL 1.111 were used. A justification
for this procedure is that tibia fractures occurred only when
HNBL 1.111 (see Table 2). Wilcoxon test was used to determine a significant difference in the vehicle impact velocity
between the femur fracture group and the ligament injury
group when HNBL 0.857. The possible difference between
the tibia fracture group and the knee ligament injury group
when HNBL 1.111 was also examined by Wilcoxon test.

3. Results
3.1. Distribution of three injury types for different
bumper lower height, normalized bumper lower height
and impact velocity

Fig. 3. Possible significant differences in vehicle bumper heights and impact

velocity between the two given injury groups.

The median, mean, maximum and minimum vehicle

bumper lower heights are shown in Table 2. Three types of injury (femur fracture, knee ligament injury and tibia fracture)

16
3
4
58
58
59
16
18
15
38.4
27.3
36.4
35.0
27.5
40.0
0.8571
0.5192
0.2941
0.9545 1.1111
0.7858 1.0417
0.7751 1.1111
0.8696
0.8085
0.7865
400
260
150
500
500
500
442
377
370
420
390
380
450
390
410
490
600
650
464
485
481

Maximum Minimum Median Mean SD Maximum Minimum

Median Mean Maximum Minimum Median Mean Maximum Minimum Median Mean

460
480
480
Femur fracture
5
Ligament injury 20
Tibia fracture
48

Injury type

Number Knee height (mm)

Knee lower height (mm)

Normalized bumper lower height

Vehicle-pedestrian impact velocity (km/h)

Y. Matsui / Accident Analysis and Prevention 37 (2005) 633640

Table 2
Median, mean, maximum and minimum pedestrian knee height, bumper lower height, normalized bumper lower height and vehiclepedestrian impact velocity according to three injury types

636

(Fig. 1) were used for the injury categories. The highest

bumper lower height involved in these three injuries was
500 mm. On the other hand, the lowest bumper lower height
in the three injury categories ranged from 150 mm (tibia
fracture) to 400 mm (femur fracture). The median and mean
bumper lower heights in femur fracture cases were the highest
in all injury categories.
The median, mean, maximum and minimum vehicle
bumper lower heights are shown in Table 2. The highest normalized bumper lower height of these three injuries ranged
from 1.0417 to 1.1111. On the other hand, the lowest bumper
lower height ranged from 0.2941 (tibia fracture) to 0.8571
(femur fracture). Thus, the lower extremity injury type could
be affected by vehicle bumper height.
The median, mean, maximum and minimum vehicle
pedestrian impact velocities are shown in Table 2. The median
and mean impact velocities in a knee ligament injury were
lower than for a fracture to the femur or tibia, respectively.
These results suggest a possibility that the lower extremity
injury type could also be affected by the vehiclepedestrian
impact velocity.
3.2. Statistical test results on six vehicle bumper heights
and impact velocity for given injury groups
Wilcoxon test results for vehicle bumper heights and impact velocity between the femur fracture group and knee ligament injury group are shown in Table 3. The bumper lower
height, bumper center height, bumper upper height, normalized bumper lower height (i.e., bumper height normalized by
pedestrian knee height) and normalized bumper center height
were statistically different (P = 5% significance) between the
femur fracture group and knee ligament injury group. Although the normalized bumper upper height between the femur fracture group and knee ligament injury group did not
reach 5% significance (P = 0.067), it may still be taken to be a
significant difference. Thus, a higher vehicle bumper height
could cause femur fracture, whereas a lower bumper height
could cause ligament injury.
Wilcoxon test results of vehicle bumper heights and impact velocity between the knee ligament injury group and
the tibia fracture group are shown in Table 4. Wilcoxon test
results showed no significant difference in the respective six
bumper heights between the ligament injury and tibia fracture
groups. This indicated that vehicle bumper heights causing a
knee ligament injury are nearly the same as those causing a
tibia fracture. Wilcoxon test results showed a significant 5%
(P = 0.022) difference in vehicle impact velocity between the
ligament injury and tibia fracture groups.
Wilcoxon test results of vehicle bumper heights and impact velocity between the femur fracture group and the tibia
fracture group are shown in Table 5. The differences in all vehicle bumper heights between the femur fracture group and
tibia fracture group were significant (P = 1% significance).
Hence, a higher vehicle bumper height could cause a femur
fracture, and a lower bumper height could cause a tibia frac-

Y. Matsui / Accident Analysis and Prevention 37 (2005) 633640

637

Table 3
Statistical test results for six vehicle bumper heights and impact velocity in femur fracture group and knee ligament injury group
Vehicle parameters

Unit

Number in femur
fracture group

Number in ligament
injury group

Median in femur
fracture group

Median in ligament
injury group

P-value

Bumper lower height

Bumper center height
Bumper upper height
Bumper lower height/knee height
Bumper center height/knee height
Bumper upper height/knee height
Velocity

mm
mm
mm

5
5
5
5
5
5
5

20
20
20
20
20
20
20

420
480
560
0.8696
1.0435
1.2174
35.0

390
453
525
0.8085
0.9468
1.0978
27.5

0.026*
0.021*
0.022*
0.028*
0.017*
0.067
0.104

km/h

5% significance.

Table 4
Statistical test results for six vehicle bumper heights and impact velocity in knee ligament injury group and tibia fracture group
Vehicle parameters

Unit

Number in ligament
injury group

Number in tibia
fracture group

Median in ligament
injury group

Median in tibia
fracture group

P-value

Bumper lower height

Bumper center height
Bumper upper height
Bumper lower height/knee height
Bumper center height/knee height
Bumper upper height/knee height
Velocity

mm
mm
mm

20
20
20
20
20
20
20

48
48
48
48
48
48
48

390
453
525
0.8085
0.9468
1.0978
27.5

380
445
510
0.7865
0.9468
1.0851
40.0

0.248
0.211
0.073
0.376
0.366
0.209
0.022*

km/h

5% significance.

Table 5
Statistical test results for six vehicle bumper heights and impact velocity in femur fracture group and tibia fracture group
Vehicle parameters

Unit

Number in femur
fracture group

Number in tibia
fracture group

Median in femur
fracture group

Median in tibia
fracture group

P-value

Bumper lower height

Bumper center height
Bumper upper height
Bumper lower height/knee height
Bumper center height/knee height
Bumper upper height/knee height
Velocity

mm
mm
mm

5
5
5
5
5
5
5

48
48
48
48
48
48
48

420
480
560
0.8696
1.0435
1.2174
35.0

380
445
510
0.7865
0.9468
1.0851
40.0

0.003*
0.001*
0.003*
0.004*
0.004*
0.008*
0.494

km/h

1% significance.

ture. The differences in the three respective bumper heights

(lower, center and upper) when normalized at knee height between the femur fracture group and the tibia fracture group
(P = 0.004, P = 0.004, P = 0.008), proved to be statistically the
same when compared to the usual three bumper heights in the
same two groups (P = 0.003, P = 0.001, P = 0.003). Wilcoxon
test results revealed that the impact velocity was not significantly different (P = 0.494) between the femur fracture group
and tibia fracture group.

3.3. Statistical test results on differences in impact

velocity between bone fracture group and knee ligament
injury group at a specied location of lower extremity
impacted by bumper
When the lower bumper height was normalized at the knee
height of 0.857 or above, the results of Wilcoxon test conducted on impact velocity between the femur fracture and
ligament injury groups were as shown in Table 6, together

Table 6
Statistical test results of vehicle impact velocity in two given injury groups for two regions of normalized bumper lower height
Normalized bumper lower height

0.857 or above (femur fractures occurred)

1.111 or less (tibia fractures occurred)

Injury groups

Femur fracture group

Ligament injury group

Tibia fracture group

Ligament injury group

Number
Median velocity (km/h)
P-value

5
35.0
0.061

8
20.5

48
40.0
0.022*

20
27.5

5% significance.

638

Y. Matsui / Accident Analysis and Prevention 37 (2005) 633640

with those between the tibia fracture group and ligament injury group when below 1.111. The present analysis yielded an
impact velocity P-value of 0.061 between the femur fracture
group and the knee ligament injury group at the normalized
bumper lower height of 0.857 or above. This strongly suggests that the impact velocity can be treated as statistically
different between the femur fracture group and knee ligament
injury group. Where the bumper lower height normalized at
the knee height was below 1.111, the Wilcoxon test result
revealed a 5% significant difference (P = 0.022) in impact
velocity between the tibia fracture group and the ligament
injury group.

Table 7
Injury mechanism to lower extremity influenced by impact velocity level
Impact velocity level
High (40 km/h)

Low (2030 km/h)

Thigh impacted by
bumper

4. Discussion and conclusions

4.1. Strength
In the present study, three vehicle bumper heights, i.e.,
bumper lower height, bumper center height, and bumper upper heights, were used to investigate the effect of the bumper
height on the type of injury to the pedestrian lower extremity.
The differences in the vehicle bumper heights between the
two given injury groups were examined by Wilcoxon test,
and found to reach statistical significance between the femur
fracture group and the knee ligament injury group in particular. Moreover, the three bumper heights were also statistically
different between the femur fracture group and the tibia fracture group. This would suggest that any of the three vehicle
bumper heights can be used to distinguish the vehicle type
causing a different type of injury. The EEVC/WG17 proposed two tests in terms of bumper height for bumper safety
evaluation in 1998 (EEVC, 1998). One test is for when a vehicle bumper collides with the lower leg, and the other for
when a vehicle with a high bumper strikes the femur region.
In selecting the vehicles, the bumper lower height was used
as the criterion. Based on the present statistical test results
using real-world vehiclepedestrian accident data, the vehicle bumper lower height referred to by the EEVC/WG17 is
an appropriate parameter to distinguish between the normal
passenger vehicle and high bumper vehicle.
The present study investigated the effect of the
vehiclepedestrian impact velocity on the type of injury at a certain bumper height. Where the bumper lower
height normalized at the knee height was 0.857 or above, the
Wilcoxon test result showed that the difference in impact velocity between the femur fracture group (median 35.0 km/h)
and the ligament injury group (median 20.5 km/h) did not
reach 5% significance (P = 0.061), but that it may still be
taken to be a significant difference (Table 6). This result suggests that when the pedestrian thigh is impacted by a bumper
at a high-impact velocity (close to 40 km/h), a femur fracture
would usually occur. It also suggests that when the pedestrian
thigh is impacted by the bumper at a low impact velocity
(from 20 to 30 km/h), ligament injuries would usually occur.
Where the bumper lower height normalized at the knee height

Lower leg impacted

by bumper

was less than 1.111, the Wilcoxon test result revealed a 5%

significant difference (P = 0.022) in impact velocity between
the tibia fracture group (median 40.0 km/h) and the ligament
injury group (median 27.5 km/h) (Table 6). Therefore, when
a lower leg is impacted by a bumper, a tibia fracture would
usually occur at a high impact velocity (close to 40 km/h), and
ligament injuries would usually occur at a low impact velocity (from 20 to 30 km/h). Thus, the mechanism by which the
lower extremity is injured could be influenced by the impact
velocity levels as shown in Table 7. If the impact velocity
(impact energy) is insufficient to cause a lower extremity
bone fracture to the tibia or femur, a high tensile force would
be exerted on each ligament, possibly causing injury. This
result was confirmed earlier by experimental studies using
cadavers (Kajzer et al., 1997, 1999; Matsui et al., 2004). In
an experiment performed by Kajzer et al. (1997, 1999), the
lower legs of completely intact cadavers were impacted by
a ram laterally at low (20 km/h) and high (40 km/h) impact
velocities. The results revealed that a low-velocity impact
mainly caused ligament injuries, whereas a high-velocity
impact caused ligament injuries and bone fractures. In an
experiment performed by Matsui et al. (2004) the femur
mid-shafts of the right and left thighs of completely intact cadavers were laterally impacted only once by a ram at a speed
of 35 km/h using various impact energy. The test results
showed that a knee ligament injury occurred at the impact energy of 450 J (left thigh), while a femur fracture was observed
at 600 J (right thigh). The type of injury was also found to be
affected by the impact energy level. In future, it would be of
interest to confirm the lower leg injury mechanism by which
the impact velocity (energy) influences the type of injury
from an computer simulation study using a human finite

Y. Matsui / Accident Analysis and Prevention 37 (2005) 633640

element model with various impact velocities or impact

energy.
One of the main findings of the present study is that the
type of lower extremity injury sustained is much affected
by the impact velocity (Table 6). When a lower extremity
is impacted by a bumper, ligament injures usually tended to
occur at a low impact velocity. On the other hand, only the
femur fracture risk can be evaluated in the currently proposed
EEVC test procedure (EEVC, 1998) for the case of impact
between the bumper and thigh. In future, the knee ligament
injury should also be considered as a type of injury for a
bumper safety evaluation test simulating an impact between
a bumper and a thigh.
4.2. Limitations
In the present study, an in-depth database of accidents that
occurred in the USA between 1994 and 1998 was used to investigate the effects of vehicle bumper height on the type of
lower extremity injury in vehiclepedestrian accidents. The
limitation of the present study is the small number of cases
involving femur fracture, i.e., only 5 compared to the 20 knee
ligament injury cases and 48 tibia fracture cases. This limitation, however, does not compromise the present results,
because the author employed a non-parametric statistical test
when investigating the possible significance of differences in
vehicle bumper height and impact velocity between the two
given injury groups.
Due to sample size, we used all the available data to investigate the effects of vehicle bumper height on the type of
lower extremity injury. This implies that accidents that occurred within the impact velocity range of 060 km/h were
used to investigate these effects. As one cannot rule out the
possibility that the results of this investigation were affected
not only by the bumper height but also the velocity, one possible direction of further studies would be to obtain additional
data under constant velocity and varied bumper height.
When selecting data for the present analysis, we excluded
one case in which a pedestrian sustained a femur shaft fracture
despite the fact that the bumper impacted only the lower leg. It
was deduced that this fracture must have occurred as a result
of femurbonnet leading edge impact. The basis for this estimation is that experimental studies using cadavers (Bhalla
et al., 2003; Kerrigan et al., 2003; Ivarsson et al., 2004;
Matsui et al., 2004) clearly indicate that in vehiclepedestrian
accidents, femur fractures nearly always occur only as a consequence of direct lateral impact to the femur.

Acknowledgements
The author is greatly indebted to Tsutomu Doi, Assistant
Professor of Ibaraki Christian College, Dr. Adam Wittek, research fellow of the University of Western Australia, Mr.
Masahiro Ito, senior researcher of the Institute for Traffic
Accident and Data Analysis (ITARDA), and to Mr. Akira

639

Sasaki for their valuable comments. The author would also

like to thank the anonymous reviewers as well as Prof. Karl
Kim, editor of this journal, for their helpful suggestions. This
study was funded by the Japan Automobile Manufacturers
Association.

References
Ballesteros, M., Dischinger, P., Langenberg, P., 2004. Pedestrian injuries
and vehicle type in Maryland, 19951999. Accident Anal. Prev. 36,
7381.
Bhalla, K., Bose, D., Madeley, J., Kerrigan, J., Crandall, J., Longhitano,
D., Takahashi, Y., 2003. Evaluation of the response of mechanical
pedestrian knee joint impactors in bending and shearing loading. In:
Proceeding of Eighteenth International Technical Conference on Enhanced Safety of Vehicles (CD).
Department of Transportation National Highway Traffic Safety Administration (NHTSA), National Center for Statistics and Analysis, 1998.
National Accident Sampling System Pedestrian Crash Data Study
(PCDS) 1998 Data Collection, Coding and Editing Manual. Washington, DC, USA.
European Experimental Vehicles Committee (EEVC), 1994. Proposals for
Methods to Evaluate Pedestrian Protection for Passenger Cars, EEVC
Working Group 10 Report.
European Enhanced Vehicle-safety Committee (EEVC), 1998. Improved
Test Methods to Evaluate Pedestrian Protection Afforded by Passenger
Cars, EEVC Working Group 17 Draft Report.
Institute for Traffic Accident Research and Data Analysis of Japan
(ITARDA), 2004. Annual Traffic Accident Report, Tokyo (in
Japanese).
Ivarsson, J., Lessley, D., Kerrigan, J., Bhalla, K., Bose, D., Crandall, J.,
Kent, R., 2004. Dynamic response corridors and injury thresholds of
the pedestrian lower extremities. In: Proceedings of the 2004 International Research Council on the Biomechanics of Impact. pp. 179
191.
ISO/TC22/SC10/WG2, 1996. Passenger Cars and Light Commercial
VehiclesPedestrian ProtectionImpact Test Method for Pedestrian
Lower Leg and Knee, ISO/TC22/SC10/WG2 Working Draft #11.
Kajzer, J., Schroeder, G., Ishikawa, H., Matsui, Y., Bosch, U., 1997.
Shearing and bending effect at the knee joint at high speed lateral
loading. SAE Trans. J. Passenger Cars 106, 36823696.
Kajzer, J., Matsui, Y., Ishikawa, H., Schroeder, G., Bosch, U., 1999.
Shearing and bending effect at the knee joint at low speed lateral
loading. SAE Trans. J. Passenger Cars 108, 11591170.
Kerrigan, J., Bhalla, K., Madeley, N., Funk, J., Bose, D., Crandall, J.,
2003. Experiments for Establishing pedestrian-impact lower limb injury criteria. In: Proceedings of the 2003 SAE International Congress
& Exposition, Society of Automobile Engineers, paper 2003-01-0895.
Lawrence, G., Hardy, B., 1998. Pedestrian safety testing using the eevc
pedestrian impactors. In: Proceeding of the Sixteenth International
Technical Conference on Enhanced Safety of Vehicles, pp. 2131
2144.
Matsui, Y., Ishikawa, H., Sasaki, A., 1998. Validation of pedestrian upper legform impact testreconstruction of pedestrian accidents. In:
Proceeding of the Sixteenth International Technical Conference on
Enhanced Safety of Vehicles, pp. 21522167.
Matsui, Y., 2001. Biofidelity of TRL legform impactor and injury tolerance of human leg in lateral impact. Stapp. Car Crash J. 45, 495510.
Matsui, Y., 2003. New injury reference values determined for TRL legform impactor from accident reconstruction test. Int. J. Crashworthiness 8 (2), 179188.
Matsui, Y., 2004. Evaluation of pedestrian subsystem test method using
legform and upper legform impactors for assessment of high-bumper
vehicle aggressiveness. Traffic Inj. Prev. 5 (1), 7686.

640

Y. Matsui / Accident Analysis and Prevention 37 (2005) 633640

Matsui, Y., Schroeder, G., Bosch, U., 2004. Injury pattern and response of
human thigh under lateral loading simulating carpedestrian impact.
In: Proceedings of the SP-1878 Vehicle Aggressivity and Compatibility, Structural Crashworthiness, and Pedestrian Safety in 2004 SAE
International Congress & Exposition, Society of Automobile Engineers, paper 2004-01-1603.
Rodmell, C, Lawrence, G., 1998. Further Pedestrian Accident Reconstructions With The Upper Legform Impactor. EEVC WG17 document
113, TRL.

Roudsari, B., Mock, C., Kaufman, R., Grossman, D., Henary, B., Crandall,
J., 2004. Pedestrian Crashes: Higher Injury Severity and Mortality
Rate for Light Truck Vehicles Compared with Passenger Vehicles.
Inj. Prev. 10, 154158.
Takeuchi, K., Bunketorp, O., Ishikawa, H., Kajzer, J., 1998. Carpedestrian accidents in gothenburg during ten years. In: Proceeding
of the 1998 International IRCOBI Conference on the Biomechanics
of Impacts, pp. 8798.