ArticlePDF Available

Typical Pedestrian Accident Scenarios in China and Crash Severity Mitigation by Autonomous Emergency Braking Systems


Abstract and Figures

In China, nearly 25% of traffic fatalities are pedestrians. To avoid those fatalities in the future, rapid development of countermeasures within both passive and active safety is under way, one of which is autonomous braking to avoid pedestrian crashes. The objective of this work was to describe typical accident scenarios for pedestrian accidents in China. In-depth accident analysis was conducted to support development of test procedures for assessing Autonomous Emergency Braking (AEB) systems. Beyond that, this study also aims for estimating the mitigation of potential crash severity by AEB systems. The China In-depth Accident Study (CIDAS) database was searched from 2011 to 2014 for pedestrian accidents. A total of 358 pedestrian accidents were collected from the on-site in-depth investigation in the first phase of CIDAS project (2011-2014). The number of car/SUV/VAN/microbus to pedestrian (all called hereafter referred to as car-to-pedestrian accidents in this study) cases is n=265, which accounts for 74% of all collected pedestrian accidents. Statistics on all collected pedestrian crashes provided an overview understanding of the pedestrian safety situation in China. To achieve the goal of the study, 255 car-to-pedestrian cases were analyzed to figure out the most frequently scenarios, which can be the reference to set the AEB test procedure in China. Furthermore, 183car-to-pedestrian cases, with detailed information regarding accident vehicles, pedestrians and environment, were reconstructed using PC-Crash. A hypothetical autonomous braking system would activate when the pedestrian successfully detected by the sensing system and then new impact speeds will be calculated. The study documents that the most frequent situations in China are:(1) Unobscured, pedestrian walks out from nearside; (2) Unobscured, pedestrian walks from far side; (3) pedestrian walking along the road. More than 20% accidents could be avoided with an AEB system (with an instant deceleration -8.0m/s2) functioned at 1.0 second prior to crash. The mean of new impact speed for un-avoided cases decreased to 24km/h from 39km/h in real accidents.
Content may be subject to copyright.
In China, nearly 25% of trafc fatalities are pedestrians. To avoid
those fatalities in the future, rapid development of countermeasures
within both passive and active safety is under way, one of which is
autonomous braking to avoid pedestrian crashes. The objective of this
work was to describe typical accident scenarios for pedestrian
accidents in China. In-depth accident analysis was conducted to
support development of test procedures for assessing Autonomous
Emergency Braking (AEB) systems. Beyond that, this study also
aims for estimating the mitigation of potential crash severity by AEB
The China In-depth Accident Study (CIDAS) database was searched
from 2011 to 2014 for pedestrian accidents. A total of 358 pedestrian
accidents were collected from the on-site in-depth investigation in the
rst phase of CIDAS project (2011-2014). The number of car/SUV/
VAN/microbus to pedestrian (all called hereafter referred to as
car-to-pedestrian accidents in this study) cases is n=265, which
accounts for 74% of all collected pedestrian accidents. Statistics on
all collected pedestrian crashes provided an overview understanding
of the pedestrian safety situation in China. To achieve the goal of the
study, 255 car-to-pedestrian cases were analyzed to gure out the
most frequently scenarios, which can be the reference to set the AEB
test procedure in China. Furthermore, 183car-to-pedestrian cases,
with detailed information regarding accident vehicles, pedestrians
and environment, were reconstructed using PC-Crash.
A hypothetical autonomous braking system would activate when the
pedestrian successfully detected by the sensing system and then new
impact speeds will be calculated. The study documents that the most
frequent situations in China are:(1) Unobscured, pedestrian walks out
from nearside; (2) Unobscured, pedestrian walks from far side; (3)
pedestrian walking along the road. More than 20% accidents could be
avoided with an AEB system (with an instant deceleration −8.0m/s2)
functioned at 1.0 second prior to crash. The mean of new impact
speed for un-avoided cases decreased to 24km/h from 39km/h in real
Requirements of legal and New Car Assessment Program (NCAP)
tests for passive pedestrian protection have been subsequently
introduced [1, 2]. Fredriksson et al. (2010) concluded that when
developing pedestrian head injury countermeasures focus should be
on the bonnet, lower windscreen, and A-pillars [3]. Pedestrian head
injury mitigating technologies, including both structural solutions
with extra space under the bonnet as well as deployable bonnets and
various airbags, were advanced [4, 5, 6, 7, 8]. Using real-life data,
several authors have pointed out that another key factor in injury
reduction is the reduction of impact speed [9-10]. Rosén and Sander
(2009) showed that reducing impact speed from 50km/h to 40km/h
would reduce pedestrian fatality risk by 50%, while a reduction from
50km/h to 30km/h would reduce the risk by as much as 80% [11].
With forward-looking sensor systems, the possibility to detect
pedestrians in advance opens up new possibilities for injury
mitigation, and has been predicted to provide considerable reductions
in pedestrian injuries. Combining autonomous braking with driver
warning or even autonomous emergency steering could further
increase effectiveness. Both driver warning and AEB systems are
already on the market. Test methods for such active safety systems
are being developed and will be implemented in NCAP tests in the
near future, e.g. 2016 in Euro-NCAP[12], and most probably in 2018
C-NCAP tests (still in consulting phase).The setting of test conditions
involves many considerations, one of which is the desirability of
subjecting the vehicles to realistic accident conditions, i.e.
circumstances that are encountered in real accidents, or at least to
understand clearly how proposed test conditions relate to the
circumstances of real accidents.
Typical Pedestrian Accident Scenarios in China and
Crash Severity Mitigation by Autonomous Emergency
Braking Systems
Published 04/14/2015
Qiang Chen, Miao Lin, Bing Dai, and Jiguang Chen
CITATION: Chen, Q., Lin, M., Dai, B., and Chen, J., "Typical Pedestrian Accident Scenarios in China and Crash Severity Mitigation
by Autonomous Emergency Braking Systems," SAE Technical Paper 2015-01-1464, 2015, doi:10.4271/2015-01-1464.
Copyright © 2015 SAE International
Downloaded from SAE International by Bing Dai, Wednesday, July 01, 2015
Problems may appear while the trafc data-based products from
developed areas operate in complex Chinese road conditions.
Studying real-world accident data is a viable way to gain an increased
understanding of the pre-crash movements of vehicles and
pedestrians. In-depth accident analysis also provides an effective way
to evaluate products' effectiveness using real Chinese accident data.
Currently, the most detailed Chinese accident database CIDAS
includes vehicle travel and impact speeds, driver braking and steering
manoeuvres as well as detailed sketches of the accident scenes. By
combining above information it is possible to derive the pedestrian
location relative to the vehicle as a function of time during the
pre-crash phase. Such extended reconstructions can also serve to
establish the time to collision and pedestrian location at the moment
when he/she would have become detectable by a vehicle based
sensor. The aim of this paper is to describe typical accident scenarios
for pedestrian accidents based on empirical data for the reference of
AEB tests in C-NCAP. At the same time, with the purpose of
understanding of their inuence on potential system effectiveness.
Case Sampling and Overview
In 2011, a China In-Depth Accident Study (CIDAS) project was
initiated by the China Automobile Technology and Research Center
(CATARC). This project aimed to collect 500600 on-site accidents
annually in 6 cities in China (Figure 1). In each investigated region,
approximately 100 trafc accidents were documented yearly with
1000-2000 individual pieces of data collected per accident and
entered into a database. From north to south, the 6 cities were
Changchun, Beijing, Weihai, Ningbo, Foshan and Chengdu (started in
Figure.1. CIDAS investigated cities(Five cities are shown. Cases in Chengdu
(a new started city in 2014 in southwest China) were not included in this
Specialist teams went directly to the scenes of the accidents in trafc
police vehicles to collect necessary information to complete detailed
accident reconstructions as well as medical data on how those
involved people were injured and treated. In this way, extensive
information for a wide range of research, such as ‘vehicle design for
passive and active safety’, ‘biomechanics’, ‘driver behavior’, ‘trauma
medicine’, ‘rescue services’, ‘road design’ and ‘road conditions’
could be collected. The sampling criterion for the rst phase of
CIDAS were dened as: (1) at least one 4-wheel vehicle involved in
the accidents, and (2) at least one person injured with AIS1+ injuries,
and (3) on-site information was not changed before investigators
arriving. However, in order to mitigate trafc jams, many accidents in
China (usually with slight injury, property loss and clear-cut accident
liability) were simply solved, and were thus too transient for
information collection for in-depth investigation.
From CIDAS database, 358 pedestrian accidents were searched from
July 2011 to June2014. Pedestrian accidents with (1) vehicle lost
control before collision; (2) vehicle rstly impacted with object; (3)
reverse driving were excluded. A subsample of n=337 pedestrians
was used for pre-selection. Sampling criteria for reconstruction cases
in current study were: (1) accident vehicles limited to cars, vans and
SUVs, excludingbusses and trucks; (2) principle forces during the
collision was to thecar body, thus,a pedestrian lying on the ground
prior to collision, pedestrian/e-bike sideswiped by the car and
absorbingonly a negligible impulse during the collision, were
excluded; (3) only persons taller than 150 cm or over 14. Finally
n=183car-to-pedestrian accidents were selected for reconstruction by
initially using PC-Crash to calculate impact conditions, such as
vehicle impact velocity, vehicle kinematic sequence and throwing
distance. For each accident, information on the accident spot, travel/
collision speeds and trajectories of car and pedestrian 5 seconds prior
to impact were provided by a special reconstruction database named
Pre-Crash Database (CIDAS-PCD).
In CIDAS database, injuries are coded according to the Abbreviated
Injury Scale (AIS2005). Six levels of injury severity are included,
where AIS1 denotes minor injury, 2 moderate, 3 serious, 4 severe, 5
critical, and 6 likely fatal injuries. The maximum AIS (MAIS)
indicates the severity of the worst injury if several were sustained by
the victim. In addition to AIS codes, injury severity in the CIDAS
database is also coded by ‘fatal’(dead within 7 days of the accident),
‘in-patient’ (more than 24 hours of medical treatment), ‘out-patient’
(less than 24 hours' medical treatment). Specialized medical
knowledge is necessary to identify the correct AIS code for each
injury type. To be precise, we used ‘fatal’, ‘in-patient’ and
‘outpatient’ to distinguish injury severity in this study.
In current study, lighting conditions are dened in 3 categories:
daytime, dawn and dark. According to astronomy information of the
accident city, accident investigators would identify daytime (between
sunrise and sunset), dawn (hours before sunrise) and dark (hours after
sunset) based on accident time. The duration before sunrise and after
sunset is decided by investigator according to factors such as weather
and season.
Downloaded from SAE International by Bing Dai, Wednesday, July 01, 2015
Case-by-Case Analysis
Case Information Extraction
In the current database, there is no variable to record some of the
studied items, e.g. accident scenarios for AEB test, it is necessary to
analyze each case to identify the situation if the pedestrian was
passing from far-side/near-side of the vehicle-travelling lane.
Accident causes recorded in the database up to now are roughly based
on trafc police categories, which is insufcient to understand the
real causes of the accidents. For example lots of accidents occurred
with ‘unsure safety’, which could include the situations like
‘misjudgments’, ‘unpreparedness for the accident’, ‘short distance to
other road users’ and ‘inappropriate speeds’ according to accident
description and interview.
A car-to-pedestrian accident occurring on a mix-used road in an
industrial area (dark without street light) is presented as an example
(Figure 2). A Lifan-passenger car was travelling northward when the
female pedestrian was walking in the middle of the street in the same
direction as the car moving. The driver was found impaired by
alcohol consumption. The throwing distance of the pedestrian was
about 23 meters away from the estimated collision location. The
wrap-around distance of the pedestrian's head measured 1.78meters.
According to on-site investigation, the car was stopped after 14meters
braking. Obvious damage was found on the car bumper, bonnet and
windscreen (Figure 3). The pedestrian died on-site. Car collision
speed was estimated at 55km/h with tolerance of ±5km/h.
Figure 2. Accident sketch (a) and on-site photo (b)
Figure 3. Car damage(The bumper was broken down which could relate to
pedestrian right tibiofibula fracture. The damage of bonnet could relate to the
bruise on thorax and upper limbs. Fracture of windscreen could relate to the
head injuries, which was the main fatal injury of the pedestrian by
medicolegal expertise.)
Accident Reconstruction
Reconstruction in the study included two parts: rstly collision-point
to nal-position with multi-body pedestrian model and pre-collision
phase with rigid pedestrian model. A scaled on-site sketch of the
accident scene is important for PC-Crash simulation. From the sketch
of accident we used: estimated initial impact location, rest positions
of vehicle and pedestrian, braking traces (if available) and some other
marks (such as scattering distribution, oil/water leaking trace).
Vehicle information contains the damages on accident cars, type,
model and car manufacture and so on (such as dimension, collision
mass, tyre information). Pedestrian information mainly includes the
height, weight, age, injury body regions and severity of the injuries.
Witness statements may include information about initial stance of
pedestrians at the moment of impact; however, not in all cases the
information from witnesses was available. Parametric studies
concerning the velocity of car and stance of the pedestrian, pitch
angle during the braking were performed in rene iterations to nd
the best correlations with all indications of in-depth on-site
investigations. The nal conguration that reproduced the same
impact points on the car, the same injuries and throwing distance to
the real accident was retained. Physical parameters including pre-/
post-impact velocities were calculated based on each time step for
both pre-crash and post-crash phases. The pre-crash phase was
calculated for at least 5 seconds.
Downloaded from SAE International by Bing Dai, Wednesday, July 01, 2015
Overview Statistics
Of the358 cases, 291 (81%) occurred on dry road surface and 42
(12%) on wet surface, the remaining were on icy (6%) and snow
cover surface (1%)(Table 1).
Table 1. Road surface
As shown in Table 2, 54% (n=193) of the pedestrian cases occurred
in daytime, 33% in night with streetlight and 13% in night without
Table 2. Light condition
It is shown that more than 60% of pedestrian accidents occurred on
straight road followed by cross-section (13%) and T-junction (13%).
It is worth noting that 7% collisions happened on overhead roads
(3%) and expressways (4%), which could be special cases of China
(Table 3).
Table 3. Road section
According to on-site investigation, there was zebra crossing available
in range of 50meters in 63% of the accident location. Moreover, 18%
of pedestrian accidents occurred on zebra crossings(Table 4).
Table 4. Zebra crossing
Based on on-site investigation and driver statement, about 85%
accidents were without obstacles to drivers' view. In the remaining
cases, drivers were blocked by plants, moving/waiting vehicles,
buildings. However, according to reconstruction result, challenges
could arise to accidents with obstacles by moving/waiting vehicles to
detecting devices for crash avoidance technologies at 0.51.0 second
before collision, which is regarded as the key time range for
triggering active safety products. Self-vehicle body, such as A-pillar,
was stated to block the driver view to nd the pedestrians in time in
12 cases (3%). In the evening, wrong use of high beam from the
oncoming vehicle was reported to affect the collision in 7(2%) cases
(Table 5).
Table 5. Driver view block
Table 6 shows the frequency distribution of trafc control. In more
than 73% of the cases, there was a trafc light near the collision
position and one party of the accident participants violated trafc
light sign. The second largest group is controlled by ‘right before left’
with a percentage of 13%.
Table 6. Traffic control
Accident Cause
According to accident process descriptions (police records) and
accident team investigation, six categories of accident causes listed in
Table 7 may contain 96% of those cases leading to collisions and the
remaining 4% were aggregated into ‘Others’ and ‘Unknown’. The six
categories can be describes as:(1) alcohol driving; (2) distraction
driving, such as telephone use/fatigue/thoughts etc…; (3) road
priority violation (including actions of car driver and pedestrian),
which can be classied by running yellow/red light, not giving way to
oncoming trafc with priority, stop sign violation;(4) unsure safety,
which includes misjudgments, unpreparedness for accidents (e.g.
pedestrian running behind an obstacle parking car, pedestrian
crossing street in front of bus), insufcient distance to other road
Downloaded from SAE International by Bing Dai, Wednesday, July 01, 2015
users; inappropriate speeds; (5) speed limit, which comprises over the
legal speed limit and excessive speed for safe control; (6) wrong
headlight use of approaching trafc in night.
Among the causes, the biggest category was ‘unsure safety’ (48%). In
this category, 29% drivers were unprepared for the accident, which
ranked the rst of all accident causes. The followed category was
road priority violation (26%), of which ‘not giving way to oncoming
trafc with priority’ accounted for 23% of all the accidents.
Table 7. Main accident causes
Pedestrian information
Figure 4 shows the age distribution. Pedestrians aged 40 to 50and 50
to 60 account for 22%(n=61) and 19%(n=61) respectively.
Figure 4. Pedestrian age group distribution (n=327) (n=31 unknown)
For n= 303 adult pedestrians, height ranging from 160 to 169 cm and
170-175cm comprises 80% totally (Figure.5).
Figure 5. Pedestrian height group distribution (n=303) (Only adult pedestrian
considered, n=28 unknown)
In the n= 255 car-to-pedestrian cases, n=269 pedestrians were
involved (in some cases more than one pedestrian was impacted by
one car). The number of pedestrian deaths was n= 94 (35%). Most of
pedestrians suffered inpatient injuries with number of n=135 (50%).
The remaining were with outpatient injuries (n=40, 15%).
For car-to-pedestrian accidents, n=120(47%) pedestrians can be
identied that head had obvious contact with windscreen. The contact
points were illustrated on a standard vehicle (Figure 6). For the
remaining cases, it was either without contact evidence or unsure if
the damage was caused by head or other body parts(125pedestrians
were without head contact and 10pedestrianswere indeterminable).
Figure 6. Distribution of head contact points on windscreen
Accident Scenarios
Based on the current sample, ve categories could cover all of the
cases: (1) pedestrian walks/runs from nearside; (2) pedestrian walks
out from behind obstruction;(3) pedestrian walks/runs from far side;
(4) pedestrian moves along street;(5) pedestrian walks out into the
path of the turning car. To be clear, Table 8 shows the illustration and
description of the ve categories. In this section, only car-to-
pedestrian accidents were considered.Table 9 shows the distribution
of car-to-pedestrian scenarios.
Downloaded from SAE International by Bing Dai, Wednesday, July 01, 2015
Table 8. Illustration and description of scenarios
Table 9. Distribution of car-to-pedestrian scenarios
Of the n=223 adult pedestrian accidents, n=68 were pedestrian walks/
runs from near side which account for more than 30%, of which more
than 60% (n=41) occurred in dark. Almost with the same frequency
with pedestrian from near side, n= 65 (29%) pedestrian walks/runs
from far side of which 66% (n=43) occurred in dark. Pedestrian
walking/running along the street account for 26% (n= 58), more than
50% occurred in dark. n= 20 (9%) cases can be covered by pedestrian
walks out into the path of turning car. Most of them (85%) occurred
in daytime. The situation of pedestrian walking out from behind
obstruction account for only 5% (n=12) in the current study, of which
n=2 occurred in dark. In this study, pedestrians less than14-year-old
were classied into child group. Of the n=32 child pedestrian
accidents, more than 40% (n =13) were from (walks/runs) near side.
The same percent 40% (n =13) were from far side. Of all of the child
pedestrian accidents, 47%(n=15) occurred in dark.
Reconstruction Output
Based on case-by-case analysis, n=183 car-to-pedestrian accidents
were reconstructed by the method stated in section of ‘Accident
reconstruction’. In Figure 7 (a), the cumulative distribution of car
impact speed for the sample is shown with a mean impact speed at
39km/h. For 80% cases, the collision speed was less than 60km/h.
The mean walking speed was about 5km/h(Figure 7 (b)).
Figure 7. Cumulative percent of car collision speed and pedestrian walking
Figure 8 presents the driver behaviors in the one second before
collision based on reconstruction results. Of the n=183 cars, 66%
were braked, 27% were without braking or accelerating and 7% were
in the acceleration. Of the n=121 braked cases, Figure 9 shows the
cumulative deceleration during the process of one second before
collision. It was shown that the mean deceleration was −2.8 m/s2 for
braking cases and less than 20% were fully braked (deceleration less
than −6m/s2).
Downloaded from SAE International by Bing Dai, Wednesday, July 01, 2015
Figure 8. Driver behavior in one second before collision
Figure 9. Cumulative of mean deceleration distribution in the process of one
second before collision (n=121)
The relative positions of pedestrian to car front center at 1.0 second
and 0.5 second before crash are shown in Figure 10. Fields of view
(FOV), 40 degrees with range of 35m, 50degrees with range of 30m
and 60 degrees with 25m were outlined for each occasion. A standard
car with width of 1.66m together with a motor lane with width of
3.32m (2 times of the standard car) was illustrated. Table 10
summarizes the rate of detected cases by eld of view of 40 degrees.
It was shown that a eld of view of 40 degrees enabled detection of
93% (n=144) of pedestrians in the group of ‘on obstacle’ at the
moment of 1.0 second prior to crash. The detected cases decreased to
n=134 (86%) at 0.5 second prior to crash. Concerning the accidents
that were investigated with information of obstruction, the
pedestrians can be detected at 1.0 second prior crash if the obstruction
was coming from xed objects (such as buildings, plants) by
reconstruction results. However when the obstruction was from
temporary objects, e.g. moving/waiting vehicles, it was difcult to
identify the exact location where the obstruction were located (when
the investigator arrived at accident scene, generally the waiting/
moving vehicles had already left). To study these cases, we assumed
that the motor lane of the car travelling was not blocked by temporary
objects and the pedestrian can be ‘seen’ if he/she was in the lane.
Based on the assumption, it is shown that only 45% (5 of 11) can be
detected at both1.0 and 0.5 second before crash. With FOV
40degrees, situation of low detecting rate also appears that only 50%
(9 of 18) could be visible by sensors at both 1.0 and 0.5 second
before crash. At 1.0 second prior to collision, view detection range
30m was sufcient for pedestrians.
Figure 10. Relative position of pedestrian to the car front center (a) at 1.0
second prior to collision, (b) at 0.5 second prior to collision(Green square
presents driver view blocked by vehicle. Circle with cross inside present car
turning cases. Detection range: 25m/30m/35m; FOV: 40degrees
(red)/50degrees (blue) /60 degrees (black))
Table 10. Rate of detected pedestrian in case of FOV 40°
Crash Severity Mitigation by AEB
In Figure 11, the cumulative distribution of with/without(original
impact speed) for the sample is shown. The mitigation prediction
with autonomous braking was applied to all cases regardless of
visibility and eld of view. This represents the greatest possible level
of mitigation given the unrealistic assumption of perfect information.
The AEB was assumed to function at 1.0 second before collision with
instant deceleration of −8 m/s2 [14]. According to the prediction,
n=38 cases (21%) can be totally avoided with new collision speed
equals 0km/h. For the remaining cases (n=145), the mean collision
speed was 24km/h, which was decreased by 38% comparing to the
original mean speed 39km/h.
Downloaded from SAE International by Bing Dai, Wednesday, July 01, 2015
Figure 11. Cumulative percent of car collision speed with/without AEB (AEB
functioned at 1 second before collision with instant deceleration −8m/s2. n =
145 cases calculated and n= 38 cases avoided (collision speed = 0) for the
situation with AEB)
For on-scene investigated accidents, a special accident investigation
team came to the accident scene upon receiving an alarm from the
emergency calling center with on-duty trafc policemen. Accident
sketches were drawn based on on-scene measurements. Detailed
vehicle damage information was inspected in a parking lot, in an area
provided for damaged vehicles in accidents. Injury information was
provided by clinical reports. It is important that accident
investigations are of high quality to assure impact speed estimations.
Systematic and random errors are expected to be kept to a minimum
level. It has been shown that in only 47% of cases pedestrian heads
contacted the car body causing obvious damage to the windscreen.
Distribution of head contact points illustrated a ‘U’ shape around the
A-pillars and the bottom of windscreen with a range of about 150mm
from the frame. In China, ‘unsure safety’ with situations of
misjudgment and unpreparedness for the accident were the most
frequent causes of pedestrian accidents. ‘Not giving way to
participants with priority’ comprises the most percentage of the cause
of road priority violation. For pedestrians, incorrectly crossing roads,
walking on freeways and trafc light violation were the most
common failures. According to the dominating accident causes,
enhancing safety consciousness of all road users and strict trafc law
enforcement could provide an effective way of decreasing accidents
and fatalities in China.
According to the studies on European and US accidents by Avery
[13], test scenarios selected to represent greatest frequency of real
world crashes were: (1) Pedestrian walks from nearside; (2)
Pedestrian walks out from behind obstruction; (3) Pedestrian walks/
runs from far side; (4) pedestrian walks along in the road; (5)
pedestrian walks out into the path of turning car. Combining Chinese
data from other international sources is shown in Table 11. It is found
that pedestrian walking along the street account for 26% in China,
which is unique comparing to developed countries. According to the
current sample, the proposal of AEB test scenarios in China could be:
(1) Unobscured pedestrian walks from nearside; (2) Unobscured
pedestrian walks from far side; (4) pedestrian walks along the road,
which comprise 85% of the pedestrian accidents scenarios. It is
necessary to underline that more than 50% pedestrian accidents
occurred in night, which should be considered in tests in China.
Table 11. Combining accident data from other international sources
Undoubtedly, reducing speed should be the primary focus for
lowering the risk of trafc fatality and injury. Currently, active
systems such as automatic braking are being rapidly developed. And
the rst systems for pedestrian safety are now on the market. The
systems consist of a pre-crash sensor that detects potentially
dangerous situations. The system will automatically brake the car if
the detected danger is unnoticed by the driver. In real-life trafc, auto
braking implies a rather abrupt intervention, which may affect the
driver's ability to maintain control of the vehicle. A discrepancy
between the need for early activation and minimum false activation is
unavoidable. Active safety products for pedestrians are now prevalent
in developed countries. However, the related products in China are
still in the debugging phase. Problems may appear while the trafc
data-based products from developed areas operate in complex
Chinese road conditions. In-depth accident analysis provides an
effective way to evaluate product efciency using Chinese data. A
previous study of pedestrian safety with Germany GIDAS data shows
that with a eld of view equal to 40°, 53 of 57 (93%) fatally/severely
injured pedestrians will be visible to the vehicle one second prior to
impact (not considering obstacles)[14]. In our study in 2014 with less
pedestrian cases (n=108) [15], the percentages of covered cases are
92% for fatal/in-patient injured pedestrians, if one second prior to
impact a similarly FOV is applied. In this study with the extended
sampling cases (n=183), similar result we got that a eld of view of
40 degrees enabled detection of 93% of pedestrians in the group of
‘on obstacle’ at the moment of 1.0 second prior to crash and the
detected percentage could decrease at 0.5 second prior to crash
(86%).With FOV 40 degrees, only 45% and 50% pedestrian could be
visible to sensors for cases that moving/waiting vehicles blocked
driver's view and when accident car was turning at 1.0second prior to
crash. The percentage would not change at 0.5 second prior to crash.
Downloaded from SAE International by Bing Dai, Wednesday, July 01, 2015
Based on the current sample, 66% drivers braked in the process of 1.0
second before collision; however, in many cases the effect of the
braking was very small with a mean deceleration −2.8 m/s2. Based on
the assumption that AEB was activated at1.0second before collision
with instant deceleration of −8.0m/s2, 21% of the studied accidents
can be avoided with the new collision speeds equal 0km/h. However,
more cases could be avoided in situations where the pedestrians could
already leave the travelling lane if car speed can be decreased. For the
remaining cases, the mean collision speed could decrease to 24km/h
and 80% of collision speeds were less than 40km/h, which could be
much safer to all kinds of pedestrian injuries.
Sources of uncertainty for this study range from the uncertainties of
accident investigation and inaccuracy of accident reconstructions.
The data used in this study was not weighted. The reason is that in
China the only national accident statistics is not transparent and
incomplete. Momently we have no reliable representative sample to
be weighted to. However, representativeness is very important to
predictive studies, which is a disadvantage for this study. In the
current study, fatality rate for car-to-pedestrian crashes is about 35%.
Accidents with no or slight injuries were always solved by accidents
participants themselves. For these cases, it could be no report to
policeman. To avoid trafc jam, accidents with slight injuries,
property loss and clear-cut accident liability are required to take quick
on-site photos and remove the accident scene before policeman
reaching. These are part of reasons that the current database is
severely-injured biased.
Scenarios: (1) Unobscured, pedestrian walking out from nearside; (2)
Unobscured, pedestrian walking from far side; (3) pedestrian walking
along the road, are the most frequent situations in China. Especially
‘pedestrian walking along the road’ is quite unique comparing to
developed countries. More than 50% pedestrian accidents occurred in
dark, it is recommended to have a scenario test in dark. FOV of
40degrees with range around 30m is sufcient to detect 93%
pedestrians that the car is not in situations of turning and blocking by
moving/waiting vehicles.
We thank Dr. Yong Chen (from Volkswagen Research China) for his
support with accident reconstruction and analysis as well as paper
writing and edition. The conclusions and points stated in this paper
are only that of CATARC.
1. EC, 2003. Directive 2003/102/EC of the European Parliament
and of the Council of17 November 2003 relating to the
protection of pedestrians and other vulnerable road users
before and in the event of a collision with a motor vehicle and
amending council directive 70/156/eec.
2. EC, 2009. Regulation (EC) No 78/2009 of the European
Parliament and of the Council of 14 January 2009 on the Type
approval of Motor Vehicles with Regard to the Protection
of Pedestrians and other Vulnerable Road Users, Amending
Directive 2007/46/EC and repealing Directives 2003/102/EC
and 2005/66/EC.
3. Fredriksson, R., Rosén, E., Kullgren, A., 2010. Priorities of
pedestrian protection - areal-life study of severe injuries and car
sources. Accident Analysis & Prevention 42 (6), 1672-1681.
4. Fredriksson, R., Haland, Y., Yang, J., 2001. Evaluation of a new
pedestrian head injury protection system with a sensor in the
bumper and lifting of the bonnet's rear part. In: Proceedings of
17th ESV Conference, Amsterdam, Netherlands.
5. Maki, T., Asai, T., Kajzer, J., 2003a. Development of future
pedestrian protection technologies. In: Proceedings of 18th
International Technical Conference on the Enhanced Safety of
Vehicles (ESV), Nagoya, Japan.
6. Nagatomi, K., Hanayama, K., Ishizaki, T., Sasaki, S., 2005.
Development and full-scale dummy tests of a pop-up hood
system for pedestrian protection. In: Proceedings of 19th
International Technical Conference on the Enhanced Safety of
Vehicles (ESV), Washington DC, USA.
7. Oh, C., Kang, Y., Kim, W., 2008. Assessing the safety benefits
of an advanced vehicular technology for protecting pedestrians.
Accident Analysis & Prevention 40 (3), 935-942.
8. Pinecki, C., Zeitouni, R., 2007. Technical solutions for
enhancing the pedestrian protection. In: Proceedings of 20th
International Technical Conference on the Enhanced Safety of
Vehicles (ESV), Lyon, France.
9. Yao, J., Yang, J., Otte, D., 2008. Investigation of head injuries
by reconstructions of real-world vehicle-versus-adult-pedestrian
accidents. Safety Science 46 (7),1103-1114.
10. Otte, D., “Use of Throw Distances of Pedestrians and Bicyclists
as Part of a Scientific Accident Reconstruction Method,” SAE
Technical Paper 2004-01-1216, 2004, doi:10.4271/2004-01-
11. Rosén, E., Sander, U., 2009. Pedestrian fatality risk as a
function of car impact speed. Accident Analysis & Prevention
41 (3), 536-542.
12. Rosén, E., Autonomous Emergency Braking for Vulnerable
Road Users. In: Proceedings of IRCOBI 2013 Conference,
Gothenburg pp. 13-71.
13. Avery, M., AEB test development, autonomous emergency
braking, and further details at
14. Rosén, E., Kallhammer, J., Eriksson, D., Nentwick, M.,
Frediriksson, R., Smith, K., Pedestrian injury mitigation
by autonomous braking, Accident Analysis and Prevention
42(2010) 1949-1957.
15. Chen, Q., Chen, Y., Bostrom, O., Ma, Y. et al., “A Comparison
Study of Car-to-Pedestrian and Car-to-E-Bike Accidents: Data
Source: The China In-Depth Accident Study (CIDAS),” SAE
Technical Paper 2014-01-0519, 2014, doi:10.4271/2014-01-
Downloaded from SAE International by Bing Dai, Wednesday, July 01, 2015
Contact Information
Bing Dai
The Engineering Meetings Board has approved this paper for publication. It has successfully completed SAE’s peer review process under the supervision of the session organizer. The process
requires a minimum of three (3) reviews by industry experts.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or
otherwise, without the prior written permission of SAE International.
Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE International. The author is solely responsible for the content of the paper.
ISSN 0148-7191
Downloaded from SAE International by Bing Dai, Wednesday, July 01, 2015
... Motor vehicles overtaking pedestrians account for 10% of all pedestrian fatalities in Europe [47,81,82], and 24%, 27%, and 26% of all types of fatal pedestrian crashes in the UK, USA, and China [83][84][85]. AES systems are designed to steer the vehicle automatically around the detected obstacle in front to avoid a likely collision. A four-phase model (approaching, steering away, passing, and returning) typically applied to describe overtaking cyclists [86] may be adapted to model an AV overtaking VRUs on a roadway including highways, as shown in Figure 6 (top). ...
Full-text available
This study aims to provide evidence to support the development of automated vehicle (AV) safety assessments that consider the possible presence of non-motorized vulnerable road-users (VRUs) on limited-access highways. Although limited-access highways are designed to accommodate high-speed motor vehicles, collisions involving VRUs on such roadways are frequently reported. A narrative review is conducted, covering the epidemiology of VRUs crashes on limited-access highways to identify typical crash patterns considering collisions severity and the underlying reasons for the VRUs to use the highway. The review results show that occupants alighting from a disabled or crashed vehicle, people seeking help or helping others, highway maintenance zones, police stops, and people crossing a highway should be given priority to ensure VRU safety on limited-access highways. The results are summarized in figures with schematic models to generate test scenarios for AV safety assessment. Additionally, the results are discussed using two examples of traffic situations relevant to the potential AV-VRU crashes on highways and the current performance of autonomous emergency braking and autonomous emergency steering systems. These findings have important implications for producing scenarios in which AV may not produce crashes lest it performs worse than human drivers in the proposed scenarios.
... For each accident case sampled, this research conduct five-step analysis method 19 for accident causation and reconstruct the accident process with PC-Crash. PC-Crash is a powerful program for the simulation of motor vehicle accidents, covering many different accident situations. ...
Vehicle-to-TW accidents are one of the largest components of traffic accidents system in China, which always leads to severe accident consequences because the cyclists are vulnerable road users and drive at a high collision velocity. With the development of information and communication technology, automated Emergency Braking (AEB) system has been applied to modern vehicles, which can perform emergency braking automatically in dangerous situations and mitigate the consequence of accident. The purpose of this study is to propose a method to identify the mechanism that affects the effectiveness of AEB functions under real-life vehicle-to-TW traffic accident scenarios with Chinese county characteristics. Through the analysis and reconstruction for the sampling accident case-by-case, the whole process of the accident has been reproduced. The trajectory analysis program determines the accident triggering conditions and then generates the virtual physical parameters of the triggering conditions through the virtual sample generation method based on the initial sample. In parallel, the AEB effectiveness simulation program has been established. Plug in parameters of AEB system generated by the Latin hypercube sampling to the AEB effectiveness simulation program for getting the mechanism that affect the effectiveness of AEB functions. The AEB system parameters generated by Latin hypercube sampling are inserted into the AEB effectiveness simulation program to obtain the mechanism that affects the effectiveness of the AEB function. Under the multi-objective optimization problem of accident avoidance rate, technical cost and occupant comfort, the optimal parameters and multi-objective values of AEB are obtained by the NSGA-II algorithm. These results can adapt to Chinese actual traffic conditions to a certain extent, and provide a reliable basis for the research and development of AEB system in China.
... How the vehicle learns in these scenarios and responds to these low-frequency but fatal malfunctions of major vehicular components would be challenging to the full automation driving system. erefore, the accident reconstruction of traditional vehicles would also be important [76]. ...
Full-text available
Autonomous vehicle (AV) is regarded as the ultimate solution to future automotive engineering; however, safety still remains the key challenge for the development and commercialization of the AVs. Therefore, a comprehensive understanding of the development status of AVs and reported accidents is becoming urgent. In this article, the levels of automation are reviewed according to the role of the automated system in the autonomous driving process, which will affect the frequency of the disengagements and accidents when driving in autonomous modes. Additionally, the public on-road AV accident reports are statistically analyzed. The results show that over 3.7 million miles have been tested for AVs by various manufacturers from 2014 to 2018. The AVs are frequently taken over by drivers if they deem necessary, and the disengagement frequency varies significantly from 2 × 10-4 to 3 disengagements per mile for different manufacturers. In addition, 128 accidents in 2014-2018 are studied, and about 63% of the total accidents are caused in autonomous mode. A small fraction of the total accidents (∼6%) is directly related to the AVs, while 94% of the accidents are passively initiated by the other parties, including pedestrians, cyclists, motorcycles, and conventional vehicles. These safety risks identified during on-road testing, represented by disengagements and actual accidents, indicate that the passive accidents which are caused by other road users are the majority. The capability of AVs to alert and avoid safety risks caused by the other parties and to make safe decisions to prevent possible fatal accidents would significantly improve the safety of AVs. Practical applications. This literature review summarizes the safety-related issues for AVs by theoretical analysis of the AV systems and statistical investigation of the disengagement and accident reports for on-road testing, and the findings will help inform future research efforts for AV developments.
... same scenario is the second most predominant pedestrian crash scenario, accounting for 27 % of all vehicle-pedestrian crashes (Yanagisawa et al., 2017). In China, the scenario of a pedestrian moving along the road (in the center or beside a two-lane road) accounted for 26 % of the pedestrian fatalities between 2011 and 2014 (Chen et al., 2015). ...
Full-text available
For pedestrians, the risk of dying in a traffic accident is highest on rural roads, which are often characterized by a lack of sidewalks and high traffic speed. In fact, hitting the pedestrian during an overtaking attempt is a common crash scenario. To develop active safety systems that avoid such crashes, it is necessary to understand and model driver behavior during the overtaking maneuvers, so that system interventions are acceptable because they happen outside drivers’ comfort zone. Previous modeling of driver behavior in interactions with pedestrians primarily focused on road crossing scenarios. The aim of this study was, instead, to address pedestrian-overtaking maneuvers on rural roads. We focused our analysis on how drivers adjust their behavior with respect to three safety metrics (in order of importance): 1) minimum lateral clearance when passing the pedestrian, 2) overtaking speed at that moment, and 3) the time-to-collision at the moment of steering away to start the overtaking maneuver. The influence of three factors on the safety metrics was investigated: 1) walking direction (same as the overtaking vehicle or opposite), 2) walking position (on the edge of the vehicle lane or 0.5 m away from the edge on the paved shoulder), and 3) oncoming traffic (absent or present). Seventy-seven overtaking maneuvers in France from the naturalistic driving study UDRIVE and 297 maneuvers in Sweden from field tests were analyzed. Bayesian regression was used to model how minimum lateral clearance and overtaking speed depended on the three factors. Results showed that drivers maintained smaller minimum lateral clearance and lower overtaking speed when the pedestrian was walking in the opposite direction, on the lane edge, or when oncoming traffic was present. Minimum lateral clearance and time-to-collision were only weakly correlated with overtaking speed. The regression models predicted distributions similar to those actually observed in the data. The time-to-collision at the moment of steering away was comparable in value to the time-to-collision used by Euro NCAP for testing active safety systems in car-to-pedestrian longitudinal scenarios since 2018. This study is the first to analyze driver behavior when overtaking pedestrians, based on field test and naturalistic driving data. Results suggest that pedestrian safety is particularly endangered in situations when the pedestrian is walking opposite to traffic, close to the lane, and when oncoming traffic is present. The Bayesian regression models from this study can be used in active safety systems to model drivers’ comfort in overtaking maneuvers.
Traffic crashes are the result of the interaction between human activities and different socio-economic, geographical, and environmental factors, showing a temporal and spatial relationship. The temporal and spatial correlations must be characterized in crash severity studies, for which the geographically and temporally weighted ordered logistic regression (GTWOLR) model is an effective approach. However, existing studies using the GTWOLR model only subjectively selected a type of kernel function and kernel bandwidth, which cannot determine the best expression of the spatiotemporal relationship between crashes. This paper explores the optimal kernel function and kernel bandwidth considering the aforementioned problem to obtain the best GTWOLR model to analyze the crash data based on the crash data of rural highways in Anhui Province, China, from 2014 to 2017. First, the GTWOLR models with Gaussian or Bi-square kernel function and fixed (the spatiotemporal distance remains constant of local sample) or adaptive (the quantity of the local sample is constant) bandwidth are compared. Second, the log-likelihood and Akaike information criterion are used to compare the GTWOLR model with the ordered logistic regression (OLR) model. Finally, the spatial and temporal characteristics of the contributing factors in the best GTWOLR model are analyzed, and corresponding countermeasures for improving traffic safety on rural highways are proposed. Model comparison results reveal that although the difference was insignificant, the Bi-square kernel function with fixed bandwidth (BF)- GTWOLR model has a better goodness of fit than the GTWOLR models with other types of kernel function and bandwidth and the OLR model. The BF-GTWOLR model estimation results showed that eight factors, including pedestrian-vehicle crash, middle-aged driver, hit-and-run, truck, motorcycle, curve, slope and mountainous, passed the non-stationary test, indicating their varying effects on the crash severity across space and over time. As a crash severity modeling approach that effectively quantifies the spatiotemporal relationships in crashes, the BF-GTWOLR model, which adapts to crash data, may have implications for future research. In addition, the findings of this paper can help traffic management departments to propose progressive and targeted policies or countermeasures, so as to reduce the severity of rural highway crashes.
Crash safety of electric two-wheelers (ETWs) has been one of the most important safety issues in China due to their high proportion of involvement in traffic accidents. Automated Emergency Braking (AEB) systems have proven to be effective in reducing the number of fatalities and injuries in traffic accidents. Providing test scenarios is one of the fundamental tasks required for establishing a set of AEB test programs for ETWs. Compared to traditional in-depth accident data, accident data accompanied with video recordings provide more accurate accident information prior to a crash as both the traffic environment and the crash process can be observed from the video. In this study, a set of typical AEB test scenarios for ETWs was developed using accident data with video information. Video recordings of 630 car-to-ETW crashes in China from 2010 to 2021 were selected from the VRU Traffic Accident database with Video (VRU-TRAVi). A K-medoids¹ cluster analysis was carried out based on variables including the collision time, visual obstruction, motion of the car and ETW before the collision, relative motion direction between the car and ETW, and the ETW type. The velocity information of cars and ETWs was also accounted for in each clustering scenario. Seven typical pre-crash scenarios were obtained, including five electric-scooter (E-scooter) scenarios (representing two scenarios where the ETWs are approaching the car from the left side, two scenarios where the ETWs are approaching the car in the same direction and another scenario where the ETWs are approaching the car in the opposite direction) and two electric-bike (E-bike) scenarios where the E-bikes are approaching the car in the perpendicular direction. Both E-bike scenarios are consistent with the E-scooter scenario except for the ETW type and velocity range; therefore, by combining the E-bike and E-scooter scenarios, five ETW scenarios were finally recommended as AEB test scenarios. By comparing with typical scenarios extracted based on the China In-Depth Accident Study (CIDAS) data and the China New Car Assessment Program (C-NCAP) test scenarios, the results show that future AEB test scenarios for ETWs should focus on scenarios with visual obstructions and scenarios where either the car or the ETW is turning, with a velocity range of 15–30 km/h for ETWs.
Road safety remains a challenge with numerous Vulnerable Road Users (VRUs) suffering from injuries and death every year. Pedestrian protection using active safety systems, such as Automated Emergency Braking (AEB), is an effective measure to combat the situation. Furthermore, the perception of precrash scenarios plays an important role in active safety research. It is essential to understand and define precrash scenarios. This study aimed to apply the obtained typical car-to-pedestrian precrash scenarios from Chinese severely injured pedestrian traffic accidents to develop and test active safety systems. The National Automobile Accident In-Depth Investigation System (NAIS) recorded 467 cases from 2011 to 2018 in China, and 12 items were selected from the NAIS database as description variables for the precrash scenario. The items were divided into four categories: car, pedestrian, road, and environment. Group decision theory was applied to evaluate the importance of each variable in its category. A total of 34 basic scenarios were defined and obtained according to the extracted significant variables. These basic scenarios represented diverse fatal scenarios in China which are crucial for autonomous driving. The frequency distribution of the scenarios demonstrated that the top five scenarios covered 85.3 % of the total. Five scenarios were identified to have the common characteristic of cars going straight. Additionally, 13 detailed scenarios were obtained from the five basic scenarios by using cluster and frequency analyses. In contrast to the New Car Assessment Program (NCAP) test scenarios, weather and lighting conditions were considered in these 13 scenarios, and the driving speed before the crash were mostly distributed in the range of 40–80 km/h (20–60 km/h in the NCAP). Meanwhile, both walking and running were commonly recorded for pedestrians to cross the street from the nearside, compared with records of walking only to cross from the nearside in the NCAP. These results contribute to a reference for test scenarios of pedestrian AEB or Forward Collision Warning (FCW) in China.
Pedestrians suffer significant injuries from ground contact, but attempts to reduce them have been limited. A recent study using multi-body simulations showed ground contact injury may be improved by controlling vehicle braking. The aim of the study is to assesses whether controlled braking would be beneficial in real-world collisions. PC-Crash was used to reconstruct 150 freely available real-world vehicle pedestrian video collisions. Pairwise comparison of actual versus controlled braking was then performed for each case. Pedestrian injury evaluation metrics included head injury criterion HIC15 and pelvis contact force, and another metric named separation distance between the vehicle and pedestrian at the instant of ground contact was used to test the feasibility of external airbags for mitigating ground contact. Substantial head and pelvis injury reductions may be possible through the application of controlled braking. Full vehicle braking after first vehicle-pedestrian contact has little influence on overall pedestrian injury outcome. In controlled braking, an external semi-elliptic airbag with a range 2 m can prevent 83.1% ground contact (only 46.6% in actual cases) and there is a high probability (50%) that the available space to control braking is limited if the vehicle deviates from its original lane. Various kinematic explanations are presented, consistent with the principle that the vertical component of the impact on the ground can be reduced by controlling the vehicle braking. Despite the variation in real-world pedestrian collisions, controlled braking may generally provide a benefit for ground contact injuries.
The problem of injury protection for pedestrian is a research hot spot in traffic safety, and simulations are important research tools. In the process of design of simulation experiments, pedestrian gaits become more and more important. Hence in the research three pedestrian gait serials (walking, running and emergency) were found through observing and reconstructing 150 vehicle-pedestrian collision videos. Further studies had found that pedestrian speed and pedestrian injury are different among the three gait serials and there is significant difference between the Pedestrian Speed Before the Collision (PSBC) and the Pedestrian Speed at the Instant of the Collision (PSIC) in the same gait serial; for the emergency gait serial, there is significant difference between the estimated and reconstructed pedestrian heights; and there is high relevance between angles of the left extremities and the corresponding right extremities. Furthermore, based on results observed from the 150 videos and those existing research results, ten pedestrian gaits in each gait serial were proposed, simulation results shown that the pedestrian injury in different gait serial and in different gaits in the same gait serial are all different. The observed pedestrian gaits will help us to design a more objective simulation experiments in the future in analyzing pedestrian injury.
A simple, but realistic, model of an autonomous emergency brake (AEB) system was studied. Using Matlab, the model was applied to 543 car-to-pedestrian and 607 car-to-bicyclist real-world collisions gathered from the highly detailed German In-Depth Accident Study Pre-Crash Matrix (GIDAS PCM) and weighted for representativeness. All collisions were to the front of the car. The aim was to investigate how AEB performance was influenced by varying some of the most relevant system parameters. A reference system was predicted to provide very high effectiveness in saving lives and mitigating severe injuries. However, the effectiveness was substantially impaired by imposing restrictions on functionality in darkness and high speeds. Further, effectiveness was highly sensitive to timing of brake activation and deceleration provided by the AEB system. Combining all these restrictions (darkness, high speed, timing and deceleration) led to a tenfold decrease of effectiveness compared to the reference system.
The aim of the study was to investigate the difference between car-to-e-bikes and car-to-pedestrian accidents. The China In-depth Accident Study (CIDAS) database was searched from 2011 to 2013 for pedestrians and e-bikes struck by car, van and SUV fronts, which resulted in 104 pedestrian and 85 e-bike cases where information was sufficient for in-depth analysis. Reconstruction by PC-Crash was performed for all of the sampled cases. Pre-crash parameters were calculated by a MATLAB code. Focus was on prototypical accident scenarios and causes; speed as well as possible prevention countermeasures. It has been shown that traffic light violations, road priority violations, and unsure safety (these situations included misjudgments, unpreparedness, proximity to other road users, inappropriate speeds, etc.) are the main causes in both the VRU groups. Distinctions were found for aspects of car collision speed, accident scenario, distribution of head contact points and so on. Pre-impact braking/warning systems could help drivers take pre-crash measures and mitigate crash severity, but a larger field of view (FOV) for sensors is of greater necessity for e-bikers than for pedestrians.
Since October 2005, the European regulation for pedestrian protection is applicable to new vehicles. Four impactors have been developed: leg, femur, child and adult heads for testing predefined areas on the front face of the vehicle. This paper presents the technical strategy and the set of solutions which place PSA Peugeot Citroën as one of the best manufacturers for pedestrian protection with in particular Citroën C6, first and unique vehicle achieving 4 stars in EuroNCAP pedestrian protection assessment. The scenario of head and leg protection is articulated around two requirements: -keeping a space between the bonnet and the various hard elements of the engine, and behind the front bumper so that the impactors do not come into contact with rigid elements, -softening the bonnet and the front bumper elements in order to generate a more progressive head and leg deceleration during the impact. The level of constraint induced by these requirements penalizes heavily the style and the overhang of the vehicles. Massive development efforts have been invested in both fields of leg and head protection. The physical characteristics of the components and the design constraints have to be optimized under advanced computational analyses with finite elements model. The protection of the leg requires the installation of two absorbers (upper and lower). The head protection requires complex tuning of the stiffness of the bonnet and some components inside the engine compartment. For executive cars with long hood, like C6, it also implied the development of an active bonnet, triggered by fusible optic sensors, which is not only a technical challenge but also addresses outstanding issues in the field of quality and reliability. The paper provides technical descriptions of the methods deployed by PSA Peugeot Citroën, associating numerical simulations and physical tests, for developing innovative solutions in the field of passive and active safety.
A pop-up hood system has been developed to reduce the severity of head injuries to pedestrians in pedestrian-to-automobile accidents. The system employs sensors located on the bumper to detect impact with a pedestrian. If an impact occurs, a signal is sent to an actuator to raise the rear portion of the engine hood approximately 100mm. This provides a space between the engine and other hard components and the hood, resulting in reduced pedestrian head injuries. Previous studies have mainly employed headform impactors to evaluate the head injury criteria (HIC) values for pop-up hoods. This report describes studies of the effect of the pop-up hood on injury parameters and kinematics using the POLAR pedestrian dummy. The effectiveness of the pop-up hood system was confirmed by the significant reduction of HIC values in impact tests using the POLAR dummy.
A sizeable proportion of adult pedestrians involved in vehicle-versus-pedestrian accidents suffer head injuries, some of which can lead to lifelong disability or even death. To understand head injury mechanisms, in-depth accident analyses and accident reconstructions were conducted. A total of 120 adult pedestrian accident cases from the GIDAS (German in-depth accident study) database were analyzed, from which 10 were selected for reconstruction. Accident reconstructions initially were performed using multi-body system (MBS) pedestrian and car models, so as to calculate head impact conditions, like head impact velocity, head position and head orientation. These impact conditions then were used to set the initial conditions in a simulation of a head striking a windshield, using finite element (FE) head and windshield models. The intracranial pressure and stress distributions of the FE head model were calculated and correlated with injury outcomes. Accident analysis revealed that the windshield and its surrounding frames were the main sources of head injury for adult pedestrians. Reconstruction results indicated that coup/contrecoup pressure, Von Mises and shear stress were important physical parameters to estimate brain injury risks.
The objective of this study was to calculate the potential effectiveness of a pedestrian injury mitigation system that autonomously brakes the car prior to impact. The effectiveness was measured by the reduction of fatally and severely injured pedestrians. The database from the German In-Depth Accident Study (GIDAS) was queried for pedestrians hit by the front of cars from 1999 to 2007. Case by case information on vehicle and pedestrian velocities and trajectories were analysed to estimate the field of view needed for a vehicle-based sensor to detect the pedestrians one second prior to the crash. The pre-impact braking system was assumed to activate the brakes one second prior to crash and to provide a braking deceleration up to the limit of the road surface conditions, but never to exceed 0.6 g. New impact speeds were then calculated for pedestrians that would have been detected by the sensor. These calculations assumed that all pedestrians who were within a given field of view but not obstructed by surrounding objects would be detected. The changes in fatality and severe injury risks were quantified using risk curves derived by logistic regression of the accident data. Summing the risks for all pedestrians, relationships between mitigation effectiveness, sensor field of view, braking initiation time, and deceleration were established. The study documents that the effectiveness at reducing fatally (severely) injured pedestrians in frontal collisions with cars reached 40% (27%) at a field of view of 40 degrees. Increasing the field of view further led to only marginal improvements in effectiveness.
The aim of this study was to aid the optimisation of future, vehicle based, pedestrian injury countermeasures. The German In-Depth Accident Study (GIDAS) database was queried for pedestrians impacted by the front of a passenger car or van. A total of 1030 cases from 1998 to 2008 were studied including 161 severely (AIS3+) injured pedestrians. Considering the severe injuries, the most frequent injury mechanisms were "leg-to-front end", "head-to-windscreen area", "chest-to-bonnet area", and "chest-to-windscreen area". For children, a "head-to-bonnet area" impact was the second most common source of injury. With safety systems targeting these five injury mechanisms, 73% (95% confidence interval [CI], 65-81%) of the severely injured pedestrians would be provided protection from all of their vehicle-induced severe injuries. Omitting the windscreen area, this figure is decreased to 44% (CI, 36-53%). Furthermore, 31% of the surviving pedestrians were estimated to sustain a permanent medical impairment at any level. For more severe impairment, head was the dominating body region. The study shows that when developing countermeasures for the windscreen area to mitigate head injuries, attention should be paid to the structural parts of the windscreen area with a special focus on brain injuries. Finally, the incidence and risk of severe injury were derived as functions of impact speed for different body regions and injury sources.