J-NCAP: TODAY AND TOMORROW

Kenji Wani
National Organization for Automotive Safety & Victims’ Aid
Susumu Ohta
Ministry of Land, Infrastructure and Transport
Hirotoshi Ishikawa
Japan Automobile Research Institute
Japan
Paper Number ID215

ABSTRACT overseas countries. Under the initiative of the

Ministry of Transport, the National Organization for
The New Car Assessment Program in Japan (JNCAP) Automotive Safety & Victims' Aid (OSA) started in
was launched in 1995 in order to improve car safety 1991 to study the type of information to be offered to
performance. According to this program, installation the users to help them choosing safer cars. Based on
conditions of safety devices and the results for this study, the first report of JNCAP containing
braking performance and full-frontal crash tests are results of full-frontal crash tests was published in
published every year. Introduction of JNCAP FY1995. The sixth report of JNCAP published in
significantly increases the installation rate of safety April 2001 contains results of full-frontal, offset
devices and contributes much in enhancement of frontal, and side impact tests as well as overall crash
safety as seen in the decrease in the average injury safety rating derived from these tests.
severity of drivers and passengers. Side impact and This paper describes reasons why respective tests
offset frontal crash tests were introduced in 1999 and were introduced, trends seen in the test results and
2000, respectively. At present, the overall crash objectives to be aimed at by JNCAP in the future.
safety rating is carried out based on the results of the
full-frontal, offset frontal, and side impact tests. 2.PURPOSE OF JNCAP

1.INTRODUCTION All cars sold in Japan meet the safety standard.

However, comparison by models reveals differences
The number of persons killed each year in traffic among them in the rate of equipping safety devices or
accidents in Japan is still over 9,000 though it has in performance of safety devices provided. Thus,
shown a decreasing trend for several years. The total JNCAP, as a neutral organization, aims at helping
number of casualties in traffic accidents has reached users to choose safer cars by offering information
as many as a million persons. Casualties who were in obtained from comparison tests on safety done by
the cars make up approximately 60% of this total and models.
the rate is increasing year after year. In order to This activity of JNCAP will not only promote wider
reduce the number of victims from traffic accidents, use of safer cars through users’ selection but also
the Ministry of Land, Infrastructureand Transport(the encourage automobile manufacturers to undertake
former Ministry of Transport) has considered it more research and development effort in producing
should be indispensable for Japan to employ the safer cars. JNCAP is also offering information on
NCAP (New Car Assessment Program) that is correct use of safety devices in order to enhance
implemented by government organizations in many users’ awareness and correct understanding on safety
Provision of J-NCAP Information
Users
*Traffic safety
education
*Promotion of use
Car users Car manufacturers
of seatbelt, etc.
*Increased awareness
of safety *More R & D effort
Cars *Selection of safer for safer cars
Road *Enhancement cars, etc.
*Facilities of safety
ensuring technologies
traffic safety *Reinforcement
of safety
standards Increased Use of Safer Cars

Figure 1 Objectives of J-NCAP

Kenji Wani 1
(see Figure 1). JNCAP report included in 1998 an article on correct
3.CONTENT OF CAR SAFETY Table 1 Trends in Safety Devices Installed in
INFORMATION Domestic Passenger Cars (excluding mini-sized
cars, cab-over and multi-purpose cars)
3.1 History of J-NCAP
Number of models surveyed and
survey years (end of December) Increase
Prior to start of actual information provision activity Safety device 1995 1996 1997 1998 from 1995
148 142 146 156 (point)
in 1995, OSA began a review in 1991 on what types models models models models
of information are to be offered and what kinds of Anti-lock brake system 38.5% 69.7% 89.0% 90.3% 51.8
Brake assist - - - 29.5% -
tests are to be conducted to collect necessary Airbag (driver’s seat) 61.3% 93.6% 96.4% 97.5% 34.4
information. OSA reviewed items relating to braking Airbag (front passenger seat) 17.4% 56.7% 82.6% 87.3% 69.9
14.4
performance and visibility performance during the Side airbag - 3.5% 16.4% 17.9% *
(1996)
two years of 1991 and 1992, and then proceeded to Adjustable belt anchor 73.2% 75.0% 77.2% 79.6% -
44.8
study items relating to collision safety in 1993 and Seatbelt pretensioner - 7.7% 38.4% 52.5% *
(1996)
continued this up to 1994. After reviewing these Seatbelt force limiter - - - 68.6% -
Seatbelt with child seat 19.9
topics from various aspects, OSA concluded that fastening function
- - 58.9% 78.8%
(1997)
*

more information relevant to collision safety should Seat with built-in child seat - - 0.7% 0.6%
-0.1
*
(1997)
be offered because users are most concerned about Note 1: "-" indicates that the subject safety device is not surveyed
* *
this subject. Note 2: (1996) or (1997) represents percentage points
increased from 1996 or 1997
In 1995 JNCAP began to offer information on safety
use of car navigation systems and child seats which
performance of cars such as collision safety, on a trial
basis, then began full-scale provision of information were being promoted by the Ministry of Transport.
using findings obtained from brake tests and full-
frontal crash tests. The side impact test was added in 3.3 Comparison Test on Safety Performance
1999. In 2000, the offset frontal crash test was added
and the program launched an overall crash safety (1) Braking test
In order to avoid accidents, brakes must be capable of
rating based on three collision tests (full-frontal,
offset frontal, and side impact). stopping cars in a short traveling distance and also
JNCAP selects models of cars to be tested starting stopping cars while maintaining them in stable
with the best selling one. Models requested by posture. This test measured stopping distances of
manufacturers can be added to the test, as well. cars with two passengers in the front seat on dry
(road surface temperature at 35.0+/-10.0oC) and wet
However, the maximum number of models allowable
(road surface temperature at 27.0+/-5.0oC) road
was limited to six per manufacturer in 2000 so that
surfaces. The brake was applied suddenly to the cars
the number of models tested may be fairly shared
tested while they were running at 100 km/h.
among manufacturers. The test was started in 1995
Deviation from the 3.5-m wide traveling lane was
with 8 cars combining small and regular passenger
also checked.
cars. This became 10 in 1996 including a mini-van
and then gradually increased to 11 in 1997 and 18 in 100km/h
1998. The target was expanded to mini-sized cars
and light-vans in 1999 implementing the test on 27 3.5m
cars. In 2000, the test was conducted on 7 mini-sized
cars, 12 small/regular passenger cars, and 5 one-box
mini-vans. Dry and wet surface
Figure 2 Braking Test
3.2 Increased Application of Safety Devices and The Braking test has been conducted in and after
Their Correct Use 1995 on cars with ABS. Figure 3 shows average
stopping distance broken down by category of car.
JNCAP runs statistics in its report on how safety
Stopping distance of passenger cars becomes shorter
devices such as ABS and airbags are provided on
as the size becomes larger. Average stopping
respective models of cars sold in Japan. JNCAP also
distance of one-box cars (mini-vans) is larger than
offers information on how to use these devices
that of passenger cars. Figure 4 shows yearly trends
correctly. JNCAP reported in 1998 statistics on the
in the average stopping distances. One-box cars,
brake assist and seatbelt force limiter that were
mini-sized cars and light-vans are excluded to
determined by the user survey as subjects of the
eliminate the influence of these categories. The
highest concern to users. Table 1 shows the
average stopping distance exhibits a decreasing trend
percentage of respective safety devices installed. The

Kenji Wani 2
both on dry and wet road surfaces. No vehicle so far In this test, cars equipped with HYBRID III (AM50)
has deviated from the traveling lane. in the driver’s and front passenger seats are impacted
against the rigid wall. From 1996 to 1999, injury
Mini-sized car severities of drivers and passengers were categorized
Passenger car A into six classes (AAA, AA, A, B, C and D) based on
the dummy’s injury values in the head and chest. In
Passenger car B addition to the above, assessment using the five-grade
Dry
system is also being conducted from 2000. It takes
Passenger car C Wet
into account injury values on other portions of the
1BOX body (neck force, neck moment, chest deflection,
femur force, and Tibia Index) as well as deformation
40 45 50 55 of cars (protrusion of the steering and brake pedal).
Stopping distance (m) Cars are also checked for ease of door opening, ease
Figure 3 Categories of Cars and Average of rescue operation of the driver and passenger, and
Stopping Distance (engine displacement of fuel leakage after collision.
passenger car A is less than 1500cc, passenger car
B is from 1500cc to less than 2000cc and
passenger car C is 2000cc or above)

Year
55km/h
1995 Concrete wall
Figure 5 Full-frontal crash Test
1996
Figures 6 and 7 show the distribution of HIC and
1997 Dry chest acceleration of the driver and that of the
Wet passenger obtained from the 1998 and 1999 tests.
1998 Some injury values are higher with one-box cars. It
is considered that higher values result from their
1999 limited crushable zone compared with that of
passenger cars.
40 45 50 55 1500
Stopping distance (m) B C
Mini-sized car

Figure 4 Yearly Trends in Average Stopping Passenger car

Distance (excluding one-box cars, light-vans and 1000
mini-sized cars) 1BOX A
HIC

AA
(2) Full-frontal crash test
The purpose of the collision tests of JNCAP is to 500
compare the safety of cars among various models in
order to reduce injury severity, in particular death and
serious injuries, of drivers and passengers resulting AAA AA B
from collision. Higher vehicle acceleration 0
employable in the full-frontal crash test enables 0 15 30 45 60 75
evaluation of restraint system and vehicle structure
making it suitable for assessing serious injuries to the Chest acceleration (G)
head and chest. The full-frontal crash test has been Figure 6 Distribution of Injury Values of Person
incorporated into the safety standards since 1994 and in Driver’s Seat (1998 and 1999)
JNCAP has employed this test since it started to
publish safety information in 1996. In order to
determine the difference in safety performance by
models more clearly (to ensure higher discriminating
capability), JNCAP employs an impact velocity 55
km/h faster than that used in the safety standards.

Kenji Wani 3
1500 Figures 10 and 11 show yearly trends from 1995 to
Mini-sized car B C 1998 in the average injury values of driver’s seat and
front passenger’s seat. The HIC values and the chest
Passenger car accelerations decreased up to 1998, however those of
1000 1BOX the driver recorded a rebound in 1999.
A 1000
AA
HIC

900
800
500 700
600

HIC
500
AAA AA B 400
0 300 Driver’s seat
200 Front passenger
0 15 30 45 60 75
100 seat
Chest acceleration (G) 0
Figure 7 Distribution of Injury Values of Person 1995 1996 1997 1998 1999
in Frontal Passenger Seat (1998)
Fiscal year
Figures 8 and 9 show trends in injury values of the Figure 10 Trends in Head Injury Value HIC in
driver and passenger in full-frontal crash tests Full-frontal crash (passenger cars alone,
recorded between 1995 and 1999. Most of the cars excluding mini-sized cars)
are evaluated as “A”, the grade lower than the
standard injury value, even when 55 km/h of impact 60.0
velocity is employed. In 1998, many of the cars are
Chest acceleration (G)

accredited with the best evaluation AAA. Overall 50.0

injury for the car occupants exhibits an improving
trend as seen in the figure. In particular, that of the 40.0
passenger in the front seat was significantly improved 30.0
in 1998. The percentage of A remained unchanged
but decrease of AAA and increase of AA were 20.0 Driver’s seat
recognized in 1999, possibly because of the increase Front passenger
10.0
in cars provided for offset collision and expansion of seat
target models of the test. 0.0
1995 1996 1997 1998 1999
100% B

80% A Fiscal year

Distribution

60% A A A
B Figure 11 Trends in Chest Injury Value in Full-
A AA A
AA
frontal crash (passenger cars alone excluding
40%
AA AAA mini-sized cars)
20% AAA
AA AA
AA AAA
0%
1995 1996 1997 1998 1999
(3) Offset frontal crash test
Figure 8 Trends in Passenger Injury Assessments Employment of the frontal offset collision test by the
(passenger in driver’s seat) ECE directives, Euro NCAP, and IIHS helped wide
spreading necessity of this test. JNCAP commenced
on study of the offset frontal crash in 1999.
100% B
B A
A
Using accident data, JNCAP determined the
80% B
A relationship between the overlap ratio and frequency
Distribution

AA B
60%
A
AA A of occurrence in frontal collisions of cars. Figure 12
40% AA shows the results broken down by injury severity of
A AA AAA
20%
AAA
AAA
drivers. As seen from the figure, offset collisions of
AA
AAA 31% to 70% overlap-ratio are causing serious injury
0%
1995 1996 1997 1998 1999 of AIS 3+ as frequently as full-frontal collisions
Figure 9 Trends in Passenger Injury Assessments proving the importance of the offset frontal crash test.
(passenger in front seat)

Kenji Wani 4
25

AIS ≥ 1 (n=273)
20 AIS ≥ 3 (n=43)
Death (n=17)

Distribution (%)
15

10

0
1 10 11 20 21 30 31 40 41 50 51 60 61 70 71 80 81 90 91 100

Overlap ratio (%)

Figure 12 Overlap Ratio and Frequency of Occurrence in Frontal Collisions (study)
In order to determine the crash worthiness of cars impact velocity. It simulates a car-to-car offset
against offset frontal collisions, JNCAP conducted collision at 55 km/h of impact velocity and 50%
the offset frontal crash test using small-size cars A overlap ratio. HYBRID III dummies are seated in
and B and then compared the results with data the driver’s seat and front passenger’s seats. Since
obtained from the full- frontal crash test. Figure 13 this test applies an impact to front body structure,
shows injury values measured at the driver’s seat. deformation resulting in the body becomes significant,
For car A, the injury value resulting from the full- making the test suitable for evaluating the body
frontal crash test is higher than that of the offset structure and damages caused in the cabin as well as
frontal crash test. The values measured with car B assessing injury to the lower legs caused by impacts
were the reverse of the above. The car A survived applied inside the cabin. Just as in the full-frontal
the collision but the car B was significantly deformed crash test, five-grade scores are computed using the
resulting in the higher injury value. This test result dummy injury values (HIC, neck force, neck moment,
indicates that cars should not be optimized to only chest deflection, chest acceleration, femur force and
one of full-frontal or offset frontal collisions so that Tibia Index) and magnitude of body deformation
they may maintain a well-balanced collision (protrusion of the steering and brake pedal). It also
performance. It also makes us recognize the
Aluminum honeycomb
necessity for both tests. A series of the above studies
proved that this test has sufficient repeatability and 40%
discriminating capability. Based on these studies,
JNCAP introduced the offset frontal crash test in
2000 succeeding to full-frontal crash test.
In the JNCAP’s offset frontal crash test, cars are
64km/h
collided with the honeycomb prepared in compliance Concrete wall
with ECE R94 at 40% overlap ratio and 64 km/h of Figure 14 Frontal offset crash test
1.2 1.2
Car A Car A
1.0 1.0
Car B Car B
Ratio to IARV

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
HIC

deflection
deflection
HIC

Femur force

Femur force
Chest
3ms-G
Femur force

Femur force
Chest
3ms-G

Chest
Chest

Right
Right

Left
Left

Full frontal impact (55 km/h) Offset frontal impact (64 km/h)

Figure 13 Comparison of Results derived from Full-frontal crash tests and Offset frontal crash tests
(study)

Kenji Wani 5
checks the ease of opening doors and rescue studies on side collisions in 1998.
operations as well as fuel leakage after the collision. Since side collisions occur more frequently between
Target models of the offset frontal crash test are vehicles, testing by use of a moving deformable
limited to passenger cars because it is difficult to barrier (MDB) is considered suitable for this purpose.
reproduce body deformation resulting from car-to-car Figure 15 shows the distribution of impact velocities.
collision in this test with commercial vehicles Risk recognition speeds up to 55 km/h make up more
provided with a frame. Expansion of the target than 90% of minor injuries or more and more than
models is one subject that requires further study. 60% of serious injuries or more.
Factors that can affect the dummy injury value
(3) Side impact test include speed, stiffness, weight, structure, and road
Among fatalities in traffic accidents in Japan, those clearance. Considering that factors other than the
resulting from side collisions of vehicles make up speed are MDB-unique problems, JNCAP focused its
approximately 24% - a large figure. In order to study on speed. Figure 16 shows results of the side
reduce and prevent such accidents, JNCAP conducted impact tests conducted on the small cars A and B at
25.0 100

90

20.0 80

Cumulative distribution (%)

Minor injury or more
Serious injury or more
70
Cumulative distribution (%)

Minor injury or more

(Cumulative frequency)
15.0 Serious injury or more 60
(Cumulative frequency)
50

10.0 40

30

5.0 20

10

0.0 0
0 10 20 30 40 50 60 70 80 90 100 120 ≥121

Traveling velocity (km/h)

Figure 15 Distribution of Traveling Velocities at which Risk was recognized in Car-to-Car Collision
(research survey)
Rib deflection Abdominal
HPC PSPF
(Maximum) force (Total)
1200 60 3 6 γ=4.26
γ=38.1 γ=1.87
σ=0.39
Car A 1000 50 σ=1.17 σ=0.31 5
800 40 2 4
γ=341.3
600 σ=71.1 30 3
400 20 1 2
200 10 1
0 0 0 0
50 55 55 55 55 60 50 55 55 55 55 60 50 55 55 55 55 60 50 55 55 55 55 60

Impact velocity (km/h)

γ=4.62
σ=0.85
1200 γ=856.5 60 3 γ=2.44 6
σ=115.9 σ=0.04
1000 50 γ=30.4 5
800 40 σ=0.68 2 4
Car B
600 30 3
400 20 1 2
200 10 1
0 0 0 0
50 55 55 55 55 50 55 55 55 55 50 55 55 55 55 50 55 55 55 55

Impact velocity (km/h)

Figure 16 Impact Velocity and Injury Values of Passenger in Side Impact Tests (research survey)

Kenji Wani 6
impact velocities of 50, 55 and 60 km/h. When the the overall crash safety rating based on the full-
impact velocity is set at 55 km/h, injury values at frontal, offset frontal and side impact tests.
various portions of the dummy increase compared to Dummy injury values obtained from the respective
results obtained at 50 km/h, making it easier to tests are converted into scores based on the
determine differences between models. At impact probabilistic injury scale and, further, they are
velocity of 60 km/h, however, no significant weighted taking into consideration of the frequency
differences were observed in injury value on the of the injuries occurring to respective portions on the
dummy compared with 55 km/h. To make matters body. For instance, scores ratio of the head, neck,
worse, a problem occurred in the test. Namely, chest, and lower legs in the frontal crash is 4:1:4:4.
deformable elements on MDB were used and this In the side impact, scores ratio of the head, chest,
could not simulate the front side stiffness of the car abdomen, and waist is 4:4:2:2, totaling 12 points. In
anymore during the test. At any rate, 55 km/h was the full-frontal or offset frontal, a computed score is
selected as the impact velocity based on these studies modified by the steering and brake deflection in order
and this test was added to JNCAP in 1999 and is used to take influences from the body deformation.
to this day. Taking into consideration of frequency of accidents
JNCAP’s side impact test method is developed based occurring in each collision type, full-frontal, offset
on the safety standards (equivalent to ECE R96) (see frontal, and side collision are weighted by 1: 1: 1,
Figure 17). Weight and road-clearance of MDB are respectively. Overall crash safety rating of the
950 kg and 300 mm, respectively. EUROSID-1 driver’s seat is rated out of 36.
seated in the collision side of the front seat assesses
injury severity for driver’s seat based on HPC, chest 4.STUDY
deflection, abdominal force and force to waist.
4.1 Child Seat Safety Assessment
Aluminum honeycomb
Truck Use of a child seat for children aged less than 6-year
old has become mandatory since April 2000. In
order to determine safety performance of child seats
and reliability of installing them, their assessment is
urgently required. Dynamic impact test and the
55km/h static test to check ease of handling will be conducted
to evaluate child seats. Preliminary research is to be
Figure 17 Side Impact Test Method completed during 2000. Test results will be
published from 2001.
Figure 18 shows results of the test conducted in 1999. The study recommends employing the sled test using
High HPC values result from contact between the a cut body of production car. Sled acceleration or
head and B-pillar. Chest deflection is mostly delta V (50 or 55 km/h) will conform to the safety
distributed around the standard value. It is standard (equivalent to ECE R44). In order to ensure
considered that injury values can be significantly fidelity to the human body, a study on injury values is
affected by the relationship between MDB and the conducted using an P3/4 dummy and Hybrid III 3YO
dummy’s seat height or position of the armrest. dummy. Referencing information obtained from
These values are smaller in one-box cars. these tests, studies have been continued to solve
The five-grade assessment system was started from problems involved in implementing the assessment
1999 computing scores based on these injury values. test (such as the belt and seat replacement frequency),
Ease of opening the doors, ease of rescue operations, usefulness of the test (usefulness of the data
and fuel leakage after collision are also checked. generated), reproducibility of the results, and the
Further studies will be continued on how to describe test’s capability of discerning differences among the
tipping over in the record, how to test one-box cars child seat products.
(vans) and how to modify test items other than speed.

3.4 Overall Crash Safety Rating by Models

It was considered necessary to present an overall

assessment based on two or more collision tests to the
users in order to ensure wider use of information
available from JNCAP. Thus, in 2000 JNCAP started

Kenji Wani 7
1.Head

1000
Mini-sized cars Passenger cars A Passenger cars B Passenger 1BOX&Mini v an Commer-
cars C cial cars
800

600

400

200

2.Chest
42
40

30

20

10

3.Abdomen

2.5

2.0

1.5

1.0

0.5

0.0

4.Pubi s

6.0

5.0

4.0

3.0

2.0

1.0

0.0

Figure 18 Results of Side Impact Test (fiscal 1999)

4.2 Death and Injury Rate by Models determine the effectiveness of JNCAP, it is important
to compare the occupant injury assessments derived
In order to ensure reliability of the test and from JNCAP’s collision tests to the ocuupant injury
assessment methods employed, as well as to rate resulting from actual accidents. Research in this

Kenji Wani 8
field will be continued to refine the accident-data make up approximately 27% of those killed in traffic
based methods for analyzing death and injury rate as accidents. Thus, another critical subject to be
well as to improve the JNCAP test assessment addressed urgently is how to protect pedestrians.
methods. JNCAP is now planning a test for pedestrian
protection along with one for verifying the head
5.FUTURE J-NCAP restraint. Expanding coverage of safety information
will help to promote a wider distribution of safer cars
It is now six years since the first car safety and to reduce the number of victims from accidents.
information was published. During the course of
these years, safety measures for cars against frontal 6.CONCLUSION
collisions have been improved as exhibited by the
increase of models accredited with the AAA rating. JNCAP has been contributing to improving
Introduction of the frontal offset collision test, side automobile safety against collisions since it was first
collision test, and overall assessment will help to published in FY1995. We are going to expand test
increase the number of suitable models provided for items and implement overall assessments better
various types of collisions. suited to real world accidents so that information
In Japan, minor injuries are increasing year after year. from JNCAP may become more helpful to users and
In particular, the percentage of whiplash sufferers more effective in reducing the number of victims
from rear-end collisions is becoming large, requiring from traffic accidents.
a significant amount of social costs. Pedestrians
Table 2 JNCAP’s Tests and Agenda(Fiscal year)
1995 1996 1997 1998 1999 2000 2001 2002
Full-frontal
Published
crash
Offset
Study Published
frontal crash
Side impact Study Published
Child seat Study Published
Pedestrian
protection Study
test
Head
Study
restraint
High-speed
brake Published
Death rate
Study
by models

Kenji Wani 9