Vehicle Safety: Technologies, Challenges, and Research and	 
Development Expenditures for Advanced Air Bags (12-JUN-01,	 
GAO-01-596).							 
								 
The National Highway Traffic Safety Administration (NHTSA) issued
a rule in May 2000 requiring vehicle manufacturers to install	 
advanced air bag systems in an increasing number of cars over the
next several years starting in September 2003. This report	 
reviews the development of technologies that vehicle		 
manufacturers plan to use to comply with the advanced air bag	 
rule. GAO found that some advanced air bag technologies are	 
currently being installed in vehicles and others are still being 
developed. The principal advanced technology that is currently	 
being installed in some vehicles is air bags that can inflate	 
with lower or higher levels of power--rather than a single	 
level--depending on the severity of the crash. Although frontal  
air bag systems with these advanced technologies represent	 
improvements over previous systems, they do not contain all of	 
the features that manufacturers believe are needed to meet the	 
requirements in the advanced air bag rule, such as sensors that  
can distinguish among different types of occupants. To meet the  
requirements, manufacturers plan to introduce new technologies as
well as continue to make further improvements in current	 
technologies. The key new technologies that manufacturers plan to
introduce are occupant classification sensors that can		 
distinguish among infants and children (as well as their safety  
seats), and adults on the passenger side. The addition of these  
sensors is necessary to allow the air bag system to provide the  
appropriate deployment level--such as no deployment, low power,  
or high power--depending on the type of occupant. The primary	 
challenge in meeting the requirements in the advanced air bag	 
rule is the development of occupant classification sensors that  
are accurate, durable, and suitable for mass production.	 
Expenditures on advanced air bag research and development by	 
NHTSA and vehicle manufacturers have increased since 1998, when  
Congress mandated the installation of advanced air bags in future
vehicles. The information aggregated from four manufacturers	 
shows that these expenditures increased by about 275 percent from
1998 through 2000 and are anticipated to increase overall by	 
about 375 percent from 1998 through 2003, when the requirements  
in the advanced air bag rule take effect.			 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-01-596 					        
    ACCNO:   A01182						        
  TITLE:     Vehicle Safety: Technologies, Challenges, and Research   
             and Development Expenditures for Advanced Air Bags               
     DATE:   06/12/2001 
  SUBJECT:   Motor vehicle safety				 
	     Research and development				 
	     Motor vehicle standards				 
	     Safety standards					 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-01-596
     
Report to the Chairman and Ranking Minority Member, Committee on Commerce,
Science, and Transportation, U. S. Senate

United States General Accounting Office

GAO

June 2001 VEHICLE SAFETY Technologies, Challenges, and Research and
Development Expenditures for Advanced Air Bags

GAO- 01- 596

Page i GAO- 01- 596 Vehicle Safety Letter 1

Appendix I Scope and Methodology 23

Appendix II Selected Aspects and Status of the Advanced Air Bag Rule 24

Appendix III Multistage Frontal Air Bag Systems in Model Year 2001 Vehicles
and Model Year 2002 Vehicles in Production as of April 1, 2001 33

Appendix IV Current Availability of and Anticipated Improvements in Advanced
Air Bag Technologies 37

Tables

Table 1: Current Availability of Advanced Air Bag Technologies and Planned
Improvements to Comply With Advanced Air Bag Rule 8 Table 2: NHTSA
Expenditures on R& D for Advanced Air Bag

Technologies, Fiscal Years 1998- 2001 20

Figures

Figure 1: Air Bag Fatalities, by Type of Occupant and Belted Status, as of
April 2001 5 Figure 2: Comparison of Conventional and Advanced Frontal Air

Bag Systems 9 Figure 3: Percentage Change in Four Vehicle Manufacturers?

Advanced Air Bag R& D Expenditures, 1999 Through 2004 21 Contents

Page ii GAO- 01- 596 Vehicle Safety

Figure 4: Test Options in the Advanced Air Bag Rule to Meet Suppression and
Low- Risk Deployment Requirements for Smaller Occupants 25 Figure 5: Tests
Required in the Advanced Air Bag Rule to Improve

Protection for Occupants of Different Sizes 28 Figure 6: Timeline for Phase-
in of Requirements in the Advanced

Air Bag Rule 29

Abbreviations

NHTSA National Highway Traffic Safety Administration R& D research and
development

Page 1 GAO- 01- 596 Vehicle Safety

June 12, 2001 The Honorable Ernest F. Hollings Chairman The Honorable John
McCain Ranking Minority Member Committee on Commerce, Science,

and Transportation United States Senate

Air bags can help save lives in frontal crashes, especially when used with
seat belts. However, the power of a deploying air bag can also injure or
kill people. According to the National Highway Traffic Safety Administration
(NHTSA), an estimated 6,856 lives have been saved by air bags, but 175
fatalities- primarily children and small women- have been attributed to the
deployment of an air bag in relatively low- speed crashes as of April 2001.
Concerned about these fatalities, the Congress directed the Secretary of
Transportation in 1998 to require vehicle manufacturers to install advanced
air bag systems. Advanced air bag systems tailor the deployment of the bags
to the characteristics of the occupant and different levels of crash
severity. On May 12, 2000, NHTSA issued a rule requiring vehicle
manufacturers to install these systems in an increasing number of vehicles
over several years beginning September 1, 2003.

You asked us to report on the development of technologies that vehicle
manufacturers plan to use to comply with the advanced air bag rule.
Accordingly, we examined the (1) current availability of and planned
improvements to advanced air bag technologies; (2) challenges, if any, that
the industry faces in complying with the rule; and (3) changes in federal
and industry expenditures on the research and development (R& D) of advanced
air bag technologies since 1998.

To address these issues, we met with representatives from eight companies
that manufacture vehicles and five companies that supply air bag systems to
these manufacturers. We chose the manufacturers- BMW, DaimlerChrysler, Ford,
General Motors, Honda, Nissan, Toyota, and Volkswagen- because, according to
the Alliance for Automobile Manufacturers, their combined sales account for
over 90 percent of vehicles sold in the United States. We chose the air bag
suppliers- Autoliv, Breed, Delphi, Takata, and TRW- because they are the
primary suppliers of air bag systems for vehicles sold in the United States.
We also met with officials from NHTSA, representatives from the Insurance
Institute for Highway Safety, the National Transportation Safety Board,
university

United States General Accounting Office Washington, DC 20548

Page 2 GAO- 01- 596 Vehicle Safety

researchers, and consumer groups. For the changes in industry expenditures
on advanced air bag R& D, an industry association provided information
aggregated from four manufacturers. We did not verify this aggregated
information because individual vehicle manufacturers consider their R& D
expenditure information confidential. (See app. I for a detailed discussion
of how we conducted our work.)

Some advanced air bag technologies are currently being installed in vehicles
and others are still being developed. The principal advanced technology that
is currently being installed in some vehicles is air bags that can inflate
with lower or higher levels of power- rather than a single level- depending
on the severity of the crash. For example, four manufacturers installed
frontal air bag systems that can deploy with multiple levels of power in at
least 25 percent of their model year 2001 vehicles. Some manufacturers have
also installed other advanced technologies, such as sensors that enable the
air bag system to adapt its deployment for belted or unbelted occupants or
for drivers who are seated close to the steering wheel. Although frontal air
bag systems with these advanced technologies represent improvements over
previous systems, they do not contain all of the features that manufacturers
believe are needed to meet the requirements in the advanced air bag rule,
such as sensors that can distinguish among different types of occupants. To
meet the requirements, manufacturers plan to introduce new technologies as
well as continue to make further improvements in current technologies. The
key new technologies that manufacturers plan to introduce are occupant
classification sensors that can distinguish among infants and children (as
well as their safety seats), and adults on the passenger side. The addition
of these sensors is necessary to allow the air bag system to provide the
appropriate deployment level- such as no deployment, low power, or high
power- depending on the type of occupant.

The primary challenge in meeting the requirements in the advanced air bag
rule is the development of occupant classification sensors that are
accurate, durable, and suitable for mass production. Sensors that are
currently being developed sometimes classify an occupant inaccurately, which
could result in the air bag not deploying when it should, deploying when it
should not, or deploying with greater or less force than intended. In
addition, occupant classification sensors have not demonstrated the ability
to operate reliably over the presumed 15- year life of a vehicle or to be
consistently produced and integrated into vehicles in large quantities.
Vehicle manufacturers are working with air bag suppliers to overcome these
problems and plan to introduce these sensors in some of their vehicles by
September 2003, as required. However, the development of Results in Brief

Page 3 GAO- 01- 596 Vehicle Safety

occupant sensing technologies is taking longer than anticipated. For
example, some manufacturers previously anticipated that occupant
classification sensors would be installed in model year 2000 vehicles. These
sensors were not sufficiently developed for installation at that time, and
some manufacturers told us that these sensors have still not reached the
level of development that a new technology should have reached to be ready
by the September 2003 deadline- just over 2 years away. NHTSA and vehicle
manufacturers have discussed these issues but, as of April 2001, none of the
manufacturers have told NHTSA they will be unable to meet the deadline. Due
to the uncertainty associated with developing occupant classification
sensors, NHTSA officials told us they plan to stay abreast of manufacturers?
progress by maintaining communications with manufacturers and conducting
research on the feasibility of occupant classification technologies.

Expenditures on advanced air bag R& D by NHTSA and vehicle manufacturers
have increased since 1998, when the Congress mandated the installation of
advanced air bags in future vehicles. NHTSA?s reported expenditures
increased from about $6.3 million in fiscal year 1998 to nearly $7.0 million
in fiscal year 2000. This spending was primarily for activities related to
the development of the advanced air bag rule, such as investigations of
real- world crashes and studies to determine how people are injured or
killed by air bags. NHTSA?s expenditures are expected to increase to $7.2
million in fiscal year 2001 to, among other things, monitor the performance
of advanced air bags and develop specific advanced air bag technologies.
Individual vehicle manufacturers did not provide information on their
expenditures because they consider this information confidential. Instead,
four manufacturers coordinated through an industry association to provide
aggregated information on their advanced air bag R& D expenditures, such as
staffing, developing technology, building test facilities and conducting
tests, and integrating technologies into vehicles. According to the industry
association, these expenditures generally total between $20 million and $30
million per vehicle ?platform? (a group of vehicles that utilize the same
basic design). The information aggregated from the four manufacturers shows
that these expenditures increased by about 275 percent from 1998 through
2000 and are anticipated to increase overall by about 375 percent from 1998
through 2003, when the requirements in the advanced air bag rule take
effect.

In commenting on a draft of this report, the Department of Transportation
offered only one technical comment on the availability of occupant
classification sensors. We verified that the information in our draft report
was accurate and therefore did not change it.

Page 4 GAO- 01- 596 Vehicle Safety

Air bags are one part of a vehicle?s occupant protection system, which also
includes the structure of the vehicle and seat belts. Seat belts are the
primary restraint for an occupant during a crash, and air bags are intended
to supplement this protection. In 1999, NHTSA reported that seat belt use
alone (lap and shoulder belts) reduces fatalities by 45 percent in crashes
involving an impact to the front of the vehicle, frontal air bags without
seat belts reduce fatalities by 14 percent, and the combination of seat
belts and air bags reduces fatalities by 50 percent. 1

Between 1986 and April 2001, frontal air bags have saved an estimated 6,856
lives but have caused 175 fatalities that have been confirmed by NHTSA- 19
infants in a rear- facing child seat, 85 children (not in a rearfacing child
seat), 64 drivers, and 7 adult passengers- in relatively lowspeed crashes. 2
NHTSA investigators have found that people who were killed by deploying air
bags were typically in close proximity to the air bag in one of two ways:

 The occupant was thrown forward by events that occurred before the air bag
deployed, such as sudden braking immediately before the crash or multiple
impacts. This usually occurred because the occupant was unbelted or
improperly belted.

 The occupant?s initial seating position placed them close to the air bag.
According to NHTSA, these fatalities included shorter drivers who were
belted but had moved the seat forward in order to more easily reach the
steering wheel and pedals, infants in rear- facing child seats, and children
sitting on the lap of another passenger.

The majority of people who were killed by deploying air bags in low- speed
crashes were unbelted or improperly restrained, which made them more
susceptible to being thrown into the path of the deploying air bag than
belted occupants. (See fig. 1.)

1 Fourth Report to Congress: Effectiveness of Occupant Protection Systems
and Their Use,

NHTSA, May 1999. 2 As of April 2001, NHTSA is investigating an additional 61
unconfirmed air bag- related fatality cases: 1 infant in a rear- facing
child seat, 43 children (not in a rear- facing child seat), 12 drivers, and
5 adult passengers. According to NHTSA, about 90 percent of unconfirmed
cases are eventually confirmed. Background

Page 5 GAO- 01- 596 Vehicle Safety

Figure 1: Air Bag Fatalities, by Type of Occupant and Belted Status, as of
April 2001

Note: Child fatalities in rear- facing infant seats are not considered
?properly restrained? because children in a rear- facing infant seat should
not be placed in a front seat with a passenger air bag.

Source: NHTSA.

The reported number of air bag- related fatalities increased from 1 in 1990
to 58 in 1997, as the installation of air bags in vehicles increased. 3
Since 1997, the number of fatalities has decreased; 17 fatalities were
reported in 2000. 4 NHTSA attributes the decrease in part to actions that
resulted from its November 1996 plan to address the risk of air bag- related
fatalities. These actions included a public education effort to persuade
people to properly restrain infants and children under 12 in the rear seat
and issuing

3 In September 1993, NHTSA required vehicle manufacturers to install air
bags in all passenger cars beginning in model year 1998, and in all light
trucks beginning in model year 1999. According to NHTSA, vehicle
manufacturers installed air bags in advance of the federal requirements.

4 Reported fatalities include confirmed and unconfirmed cases.

Page 6 GAO- 01- 596 Vehicle Safety

a rule in March 1997 that made it possible for manufacturers to quickly
reduce the inflation power in the air bags installed in new vehicles. From
model year 1997 through 1998, manufacturers lowered the inflation power by
an average of 22 percent in driver- side air bags and 14 percent in
passenger- side air bags. 5 NHTSA?s 1996 plan also anticipated the need for
long- term technological improvements- advanced air bag systems- to control
or prevent deployment of the air bag, as appropriate.

In June 1998, the Transportation Equity Act for the 21st Century directed
the Secretary of Transportation to issue a rule requiring vehicle
manufacturers to install advanced air bag systems. The act specified that
these systems should achieve two goals- provide improved protection for
occupants of different sizes (belted and unbelted) as well as minimize the
risk of injury or death from air bags for infants, young children, and other
occupants. On May 12, 2000, NHTSA issued a rule specifying the requirements
for such a system. Under the previous requirements, vehicle manufacturers
performed tests that involved crashing vehicles into a rigid barrier with
crash dummies- belted and unbelted- that represented average- sized males in
the driver and passenger seats. 6 To provide improved crash protection for
occupants of different sizes, the rule adds new crash tests that simulate
different types of crashes and include the use of crash dummies that
represent small adults (defined as a 5th percentile female). 7 To reduce the
risk of injury or death to children and small adults, the rule requires a
new battery of ?static? tests using dummies representing infants, young
children, and 5th percentile females. These tests involve placing the dummy
in various positions in the seat to determine if the air bag system
suppresses or activates the air bag, or placing the dummy against the air
bag module and deploying the air bag to determine if the bag deploys in a
?low- risk? manner that does not cause severe injury. Starting in the
production year beginning September 1, 2003

5 As of April 2001, NHTSA has found that depowered air bags appear to be as
effective as previous air bags in protecting occupants, including larger
occupants in medium- to highspeed crashes.

6 NHTSA?s March 1997 rule allowed manufacturers the option of certifying the
performance of air bag systems using a ?sled test? rather than crashing the
vehicle into a rigid barrier. A sled test involves placing a vehicle on a
?sled- on- rails? and accelerating the sled and vehicle very rapidly
backward.

7 Crash dummies do not sufficiently represent the human physiology, so the
rule allows the use of appropriately sized humans to test the performance of
the technologies in static suppression tests.

Page 7 GAO- 01- 596 Vehicle Safety

(approximately model year 2004) 8 and continuing over a 3- year phase- in
period, increasing percentages of each manufacturer?s vehicles must comply
with the requirements of the rule. 9 (See app. II for a more comprehensive
discussion of the rule?s requirements.)

Manufacturers have installed some of the advanced technologies that will be
needed to comply with the advanced air bag rule in certain vehicles that are
on the market today. (See table 1.) Manufacturers and companies that produce
air bags are working on the development of other needed advanced
technologies, with the aim of having them ready for installation in vehicles
by September 2003, as required.

8 The timing of the model year varies from company to company, but generally
begins between August and October. 9 The rule will be phased in during two
stages. During the first stage phase- in- from September 1, 2003, to August
31, 2006- an increasing number of each manufacturer?s vehicles must be
certified each year as passing all of the requirements in the advanced air
bag rule. During the second stage phase- in- from September 1, 2007, to
August 31, 2010- the speed for one of the tests (the belted test for the
50th percentile adult male dummy) will be increased from 30 to 35 miles per
hour. As with the earlier requirements, an increasing percentage of vehicles
must comply with the new test speed each year. Some Advanced

Technologies Are Available; Others Are Being Developed

Page 8 GAO- 01- 596 Vehicle Safety

Table 1: Current Availability of Advanced Air Bag Technologies and Planned
Improvements to Comply With Advanced Air Bag Rule

Component of advanced air bag system Advanced technologies currently
available Improvements planned to comply with advanced air

bag rule

Occupant classification sensors

Some vehicles with multistage air bags have one or more of the following:

 Driver and passenger seat belt usage sensors

 Driver seat position sensors

 Sensors to detect the presence of an occupant in the passenger seat

 Increased use of seat belt usage and seat position sensors.

 Weight- based or pattern- based sensors that can identify the type of
occupant in the passenger seat (i. e., an infant or child in a child seat, a
child outside of a child seat, or an adult).

 Possible use of seat belt tension sensors and/ or occupant position
sensors to augment information provided by above sensors. Crash sensors
Vehicles with multistage air bags have crash

sensing systems that have been refined to better discriminate among crash
severity levels. These sensors are located in the passenger compartment;
some are augmented by front (crush zone) sensors.

Crash sensing systems with a greater ability to differentiate levels of
crash severity as well as types of crashes. Use of front (crush zone)
sensors expected to increase.

Control module Vehicles with multistage air bags have more complex
computational systems designed to make timely decisions about the
appropriate level of air bag deployment based on input from crash sensors
and, as applicable, occupant sensors.

Computational systems of increased complexity that will be able to process
more inputs from crash sensors and occupant classification sensors and make
timely decisions regarding the appropriate level of air bag deployment.
Multistage inflators

Vehicles with multistage air bags have inflators with 2 or 3 levels of
deployment. Deployment decisions are based on input from crash sensors and,
as applicable, occupant classification sensors.

Inflators with 2 or more levels of deployment. Deployment decisions will be
based on input from crash sensors and occupant classification sensors.

Air bags Newer vehicles have air bags with improved designs aimed at
reducing the aggressivity of the deploying bag. Improvements have included
moving the air bag module further away from the occupant, use of tethers
within the bag, and changes in bag folding, shapes, and venting.

Further improvements may include increased use of innovative bag designs as
well as new designs aimed at enhancing the ability of the deploying air bag
to adapt to characteristics of the occupant.

Source: GAO analysis of information provided by vehicle manufacturers.

Advanced air bag systems installed in future vehicles will be much more
sophisticated than the conventional air bag systems in today?s vehicles,
because they will be capable of tailoring air bag deployment to
characteristics of the front seat occupants as well as crash severity.
Conventional frontal air bag systems deploy the air bags with a single level
of inflation output for all crashes that exceed a predetermined severity
threshold. These systems generally consist of separate components designed
to work together: crash sensors, a control module, and a driver and
passenger inflator and air bag. (See fig. 2.) The crash sensors and control
module are typically located in one unit within the passenger compartment;
the unit is often mounted within the floor between the driver and the
passenger. The crash sensors detect the occurrence and severity of crashes
and provide this input to the control module. The Components of

Conventional and Advanced Air Bag Systems

Page 9 GAO- 01- 596 Vehicle Safety

control module evaluates inputs from the sensors. If the control module
determines that a crash has occurred that exceeds the severity threshold, it
then sends a triggering signal to the inflators to deploy the air bags. The
inflators and air bags are packaged together in air bag modules, which are
located in the steering wheel on the driver side and in the instrument panel
on the passenger side. Upon receiving a triggering signal from the control
module, inflators generate or release gases that rapidly fill the air bags,
generally within 1/ 20 of a second after impact. The purpose of the inflated
air bags is to provide protective cushioning between the occupants and the
steering wheel, instrument panel, and windshield. However, the ?singlestage?

inflators in most vehicles today, in some cases, provide more inflation
power than necessary because they fill the air bags with one level of output
when deployed, regardless of the types of occupants requiring protection or
the degree of severity of the crash.

Figure 2: Comparison of Conventional and Advanced Frontal Air Bag Systems

Source: GAO analysis of information provided by vehicle manufacturers and
literature on air bag technologies.

Future frontal air bag systems designed to meet the performance requirements
of NHTSA?s advanced air bag rule will have additional

Page 10 GAO- 01- 596 Vehicle Safety

features that will allow the deployment of the air bags to adapt to
characteristics of the front seat occupants as well as different crash
situations. Auto manufacturers anticipate that two new components will be
needed to meet the rule?s requirements: occupant classification sensors and
multistage inflators. Occupant classification sensors will provide an
additional input to the control module to detect different types of
occupants and whether or not they are belted. For example, manufacturers
anticipate installing sensors that will be able to identify whether the
front passenger seat is occupied by an infant in a rear- facing child seat,
a child, or an adult. Multistage inflators, which will replace single- stage
inflators, will provide varying levels of inflation output that can be
tailored to characteristics of the driver and front seat passenger as well
as different crash scenarios. Deployment options could include no
deployment, low- level output, and high- level output, as well as additional
levels of deployment between the low- and high- output stages. While the
occupant classification sensors and multistage inflators are the key new
features of the advanced air bag systems envisioned by auto manufacturers,
other components will also be improved. For example, manufacturers
anticipate that these systems will include crash sensors that can more
precisely discriminate among different types of crashes (such as a crash
into a rigid concrete wall versus a crash with another car), control modules
that can process the additional inputs provided by crash and occupant
sensors and make more accurate and timely deployment decisions, and air bag
designs that will allow the bag to deploy less aggressively.

These advanced air bag systems will be designed to reduce the likelihood of
the types of fatalities previously caused by air bag deployments. For
example, such systems would deactivate the passenger air bag or deploy it at
a low level if the passenger seat is occupied by an infant or small child.
10 These systems may also adjust air bag deployment if the driver or
passenger is a small adult.

10 NHTSA has stressed that children are safest in the back seat and that
this will continue to be the case even in vehicles that have advanced air
bag systems.

Page 11 GAO- 01- 596 Vehicle Safety

Some vehicles on the U. S. market today have frontal air bag systems with
multistage inflators and some other advanced features, such as seat belt
usage sensors and improved air bag designs. 11 However, no vehicles
currently on the market have air bag systems with all the features
manufacturers believe are needed to fulfill the requirements of the advanced
air bag rule. In particular, no vehicles currently have frontal air bag
systems with occupant classification sensors that can distinguish among
child seats, children, or adults. Manufacturers are not required to produce
vehicles that can meet the requirements of the advanced air bag rule until
the production year starting in September 2003 (approximately model year
2004).

Frontal air bag systems with multistage inflators started appearing on the
market in some model year 1999 and 2000 vehicles and became more widely
available in model year 2001 vehicles. While three of the eight
manufacturers we talked to installed multistage air bag systems in some or
all of their model year 1999 vehicles, seven of the manufacturers installed
this technology in some or all of their model year 2001 vehicles. Four of
these seven manufacturers- BMW, DaimlerChrysler, Ford, and Honda- installed
multistage air bag systems in at least one- quarter of their model year 2001
fleets. 12 While most of the multistage air bag systems installed in these
model year 2001 vehicles have two stages of inflation, some have three
stages. Manufacturers are planning to further increase the number of
vehicles with multistage air bag systems in their model year 2002 fleets.
(See app. III for more detailed information on the availability and features
of multistage frontal air bag systems in U. S. market vehicles.)

Most of the multistage air bag systems installed in vehicles on the market
today have one or more types of sensors that provide information about

11 Many vehicles on the market today also have newer seat belt features that
improve the performance of the occupant restraint system in protecting
occupants. For example, seat belt pretensioners retract a limited amount of
webbing to help minimize the forward movement of the occupant during a
crash. Also, energy management features prevent belts from concentrating too
much energy on the occupant?s chest during a severe crash. These include
?load limiters? built into the shoulder belt retractor and/ or ?tear
stitching? in the webbing that causes the seat belt to extend gradually.

12 BMW, DaimlerChrysler (Mercedes- Benz), and Honda introduced multistage
air bag systems in their model year 1999 fleets; Ford and General Motors
introduced them in their model year 2000 fleets; and Nissan and Toyota
introduced them in their model year 2001 fleets. The availability of
multistage air bag systems in model year 2001 vehicles, as a percentage of
each company?s fleet, was: BMW- 100 percent, DaimlerChrysler- 27 percent,
Ford- 31 percent, General Motors- 10 percent, Honda- 74 percent, Nissan- 19
percent, and Toyota- 1 percent. Some Advanced Air Bag

Technologies Are Currently Available

Page 12 GAO- 01- 596 Vehicle Safety

the front seat occupants, such as the presence of an occupant in the
passenger seat, driver seat position, and driver and passenger seat belt
use. In air bag systems with these occupant sensors installed, the control
module utilizes input from these sensors, in addition to input from the
crash sensors, in making deployment decisions.

 Three manufacturers- BMW, DaimlerChrysler (Mercedes- Benz), and Ford- have
offered some model year 2001 vehicles equipped with weightbased occupant
presence sensors on the passenger side. In these vehicles, the control
module deactivates the passenger air bag if the sensor detects that the
passenger seat is unoccupied. The main purpose of these sensors is to
prevent unnecessary deployment of the passenger air bag and save on repair
costs. The sensors are not capable of identifying what type of occupant is
in the passenger seat. 13

 One manufacturer- Ford- has offered model year 2001 vehicles equipped with
sensors that detect whether the driver?s seat is positioned forward or
rearward on the seat track. When the sensor detects that the seat is
positioned forward, indicating that the driver is seated close to the air
bag module, the control module deactivates the high- output stage of the
driver?s air bag.

 Four manufacturers- BMW, DaimlerChrysler (Mercedes- Benz), Ford, and
Honda- have offered some model year 2001 vehicles that contain, as part of
their multistage air bag systems, sensors that detect whether the occupants
are wearing seat belts. The control module deploys the air bags at a higher
crash severity threshold if the occupant is belted and a lower threshold if
the occupant is unbelted. 14

In addition to installing the new air bag technologies described above,
manufacturers have also made improvements to crash sensors, control modules,
and air bags. In currently available multistage air bag systems, the level
of air bag deployment in a crash is based on the level of crash severity,
although the occupant sensors described above also affect deployment
decisions in some vehicles. The crash sensors in these systems have been
refined to better discriminate crash severity levels.

13 However, the occupant presence sensing system in Mercedes- Benz vehicles
does have the capability of detecting the presence of a compatible child
seat in the front passenger seat, through a tag in the bottom of the child
seat. The control module deactivates the passenger air bag when the tag is
detected.

14 Although Volkswagen has not produced any vehicles with multistage air bag
systems, some of the company?s model year 2001 vehicles have these types of
belt use sensors installed.

Page 13 GAO- 01- 596 Vehicle Safety

These crash sensors are generally arranged in one of two ways. In the first
type of arrangement, which is typically used in conventional air bag
systems, a ?single- point? electronic crash sensor is located within the
control module in the passenger compartment. In the second type of
arrangement, called a ?multipoint? electronic crash sensing system, one
sensor is located within the control module and one or more sensors are
located in the front (crush zone) of the vehicle.

In all of the multistage air bag systems installed in vehicles on the market
today, the control modules contain more complex computational systems
designed to make timely decisions about the appropriate level of air bag
deployment. In multistage air bag systems that include occupant sensors and/
or multipoint crash sensing systems, the control modules must process the
additional inputs provided by these sensors in making deployment decisions.

Manufacturers have made a variety of improvements in their air bag designs
aimed at reducing the aggressivity of the deploying air bag and, therefore,
the risk of injury caused by deployment. One major area of improvement has
been to change the location of the air bag module or the size, shape, and
folding of the bag to increase the distance between the occupant and the
deploying air bag. For example, on the driver side, manufacturers now often
recess the air bag into the steering wheel and employ a fold and shape that
allows the bag to deploy laterally rather than rearward toward the driver.
Some passenger air bags in use today contain a device that directs the
initial inflation of the bag away from the occupant if he or she is in close
proximity to the bag at the time of deployment. Other improvements in bag
design that are used in some vehicles include vents that can make the bag
deploy more softly if it is obstructed by the occupant during deployment and
the use of tethers within the bag to reduce extension when deployed. (For
further information on advanced technologies currently installed in
vehicles, see app. IV.)

Vehicle manufacturers, along with companies that supply them with air bag
systems, are working now on developing frontal air bag systems that are
intended to meet the requirements in the advanced air bag rule and be ready
to be installed in model year 2004 vehicles, as required. The advanced air
bag systems envisioned by manufacturers for meeting the rule?s requirements
include new technologies that have not previously been installed in vehicles
as well as significant improvements in existing technologies. Significant
Improvements

Are Under Development

Page 14 GAO- 01- 596 Vehicle Safety

The key new technologies that manufacturers anticipate will be needed to
comply with the advanced air bag rule are occupant classification sensors
that can identify whether the passenger seat is occupied by an infant in a
child seat, a small child in or out of a child seat, or a small adult. Air
bag suppliers have been working on the development of a number of such
sensor technologies, and manufacturers are currently considering these
technologies. The primary technologies under consideration are weightbased
sensors and pattern- based sensors, which would be installed within or under
the passenger seat. Weight- based sensors attempt to classify the occupant
through various means of determining the amount of force or pressure applied
to the seat. Pattern- based sensors attempt to classify an occupant using a
mat, installed directly under the seat cover, which senses the occupant?s
applied pressure and imprint. Manufacturers are also considering augmenting
some of these technologies with seat belt tension sensors to identify
whether the amount of force applied to the seat is due in part to the seat
belt rather than the occupant?s weight. In addition to developing new
sensors for identifying the type of occupant in the passenger seat,
manufacturers plan to increase the use of driver and passenger seat belt use
sensors and driver seat position sensors. As described in the previous
section, these occupant sensor technologies are already developed and
available in some vehicles.

Manufacturers also plan to continue making improvements in existing
technologies for crash sensors, control modules, inflators, and air bags to
comply with the advanced air bag rule. Manufacturers and suppliers are
working on improving the ability of crash sensing systems to differentiate
levels of crash severity and types of crashes. As part of this effort,
manufacturers plan to increase the use of multipoint crash sensing systems.
Manufacturers and suppliers are also developing more complex computational
systems to be incorporated into control modules, in order to allow them to
process the additional inputs in advanced air bag systems and to make
accurate and timely decisions regarding deployment outputs. Manufacturers
will use multistage inflators that have two or more stages of inflation
output in their advanced air bag systems. Some manufacturers have already
installed inflators with more than two stages of inflation on a limited
basis, but other manufacturers have told us that they do not plan to use
them until occupant classification and control module technologies are more
fully developed. Finally, manufacturers and suppliers continue to work on
improvements in air bag design, such as venting and bag shapes, in order to
enhance the ability of vehicles to comply with the advanced air bag rule.
Further improvements may include increased use of innovative bag designs as
well as new designs that will enhance the ability of the deploying air bag
to adapt to characteristics of the occupant.

Page 15 GAO- 01- 596 Vehicle Safety

Vehicle manufacturers and air bag suppliers are also researching some other
advanced air bag technologies that are not considered necessary for
complying with the advanced air bag rule but that may be used in the longer
term to enhance the performance of air bag systems.

 Some manufacturers and air bag suppliers are researching ?dynamic?

occupant position sensing, which would continuously track the proximity of
the occupant to the air bag. Inputs from these sensors, which would be
installed in the passenger compartment, would be used by the control module
to determine when the occupant is in close proximity to the air bag and,
when this is the case, to deactivate the bag. ?Static? sensors that
periodically determine the occupant?s position may be installed on a limited
basis in the near term to augment occupant classification sensors. Although
researchers are examining various technologies for achieving dynamic
occupant position sensing, it is not yet clear whether or when this
technology will become widely used.

 Precrash sensing is another area of technology currently in the research
stage. These sensors would identify the position, speed, and mass of objects
prior to a collision and allow more time for the air bag system to respond.
The feasibility of this concept has not yet been determined; therefore, it
is not yet clear when this technology might become available.

 Some suppliers are researching inflator technologies that may produce
continuous variation in inflation, rather than inflation in discrete stages,
allowing air bag deployment to be more adaptive to inputs from crash and
occupant sensors. These may be introduced by some manufacturers during the
initial 3- year phase- in period for complying with the advanced air bag
rule.

(For further details on anticipated advancements in air bag technologies,
see app. IV.)

Page 16 GAO- 01- 596 Vehicle Safety

According to representatives of vehicle manufacturers and air bag suppliers,
the primary challenge in meeting the requirements of the advanced air bag
rule is developing occupant classification sensors for the passenger side
that are accurate, durable, and suitable for mass production before
September 2003. The rule requires manufacturers to install advanced air bag
systems that either suppress the air bag if an infant or child is seated in
the passenger seat or deploy it in a ?low- risk? manner that does not cause
severe or fatal injury, even if the infant or child is out of position. 15
If the system is designed to suppress the air bag in the presence of an
infant or child, it must deploy if the passenger is a small adult (defined
as a 5th percentile woman). To test whether a sensor accurately classifies
an occupant so the air bag can deploy appropriately, the rule specifies
tests using dummies representing infants, 3- year- old and 6- year- old
children, and 5th percentile women. The dummies have fixed weights, heights,
and stature that are easily distinguishable from each other. However, the
rule also requires that some tests be conducted using child seats, variable
seat belt tension, blankets, and with the dummies in various positions.
These added factors make it more difficult for sensors to distinguish among
the different occupants. In addition to the requirements in the rule,
manufacturer and supplier representatives told us that they are designing
occupant classification sensors for additional

?real- world? situations that further challenge the ability of sensors to
perform accurately. Such real- world situations could include variation in
the actual weight of humans, changes in weight detected by sensors as the
occupant moves forward, backward, and side- to- side, or increased weight
from objects held on laps.

Manufacturers generally require that technologies perform accurately over 99
percent of the time before being installed in vehicles. However,
manufacturer representatives told us that technologies that are currently
being developed for occupant classification sensors, such as weight- based
or pattern- based sensors, have not demonstrated the ability to consistently
distinguish among various sizes of occupants. For example, weight- based
sensors in seats have difficulty distinguishing between 6- year- old
children and small adults because a 6- year- old child can appear heavier
from additional weight (such as a booster seat and increased tension from a

15 The rule also allows manufacturers the option of suppressing the air bag
if the occupant moves out of position during a crash, as occurs in precrash
braking. However, NHTSA has not developed performance requirements and test
procedures for a system that dynamically suppresses the air bag. In
addition, manufacturer representatives told us that they are not planning to
install such a system before September 2003 because the technologies needed
for this system- dynamic occupant position sensors- are not sufficiently
developed. Occupant Sensing Is

the Primary Challenge in Meeting the Advanced Air Bag Rule?s Requirements

Page 17 GAO- 01- 596 Vehicle Safety

tightly cinched seat belt); additionally, small adults can appear lighter
because a portion of the occupant?s weight is borne by the legs resting on
the floor. Pattern- based sensors must first be programmed to recognize
various seating positions. If a child or a small adult sits in a position
that was not previously anticipated and programmed for the sensor, the
system could mistake the child for an adult or vice versa. Incorrect
classification of an occupant could result in the system mistakenly
deploying the air bag in the presence of a child, not deploying in the
presence of an adult, or deploying the air bag with greater or less force
than intended.

In addition to performing accurately, occupant classification sensors must
also be durable and capable of being consistently produced and integrated
into vehicles in large quantities. Air bag systems are expected to operate
reliably over the life span of a vehicle, which could be up to 15 years.
However, sensors are susceptible to aging and environmental influences over
that time. For example, the performance of pattern- based sensors that are
installed directly under the seat cover could be affected by deterioration
of the seat cover. Sensor performance is also affected by variations in the
manufacturing process that can affect the construction of the sensor or how
easily it can be integrated into the vehicle. The parts of a sensor must be
precisely constructed because inconsistencies in the parts can cause the
sensor to malfunction. Sensors that are integrated into a seat are also
subject to variations in how the seat is constructed. According to vehicle
manufacturer representatives, companies that produce vehicle seats will have
to significantly redesign seats and decrease the variation in the production
of seats before occupant classification sensors can consistently function
properly.

Vehicle manufacturers are working with the companies that supply air bag
systems to find solutions to these accuracy, durability, and manufacturing
issues. For example, to address the influence of seat belt tension on
weight- based sensors, some manufacturers and suppliers are developing seat
belt tension sensors that would detect when the seat belt is cinched tightly
and causing the occupant to appear heavier. Individual manufacturers are
simultaneously developing multiple occupant classification technologies with
different suppliers to increase the likelihood of finding a solution by the
deadline. Some manufacturers told us they have also postponed research on
occupant position sensors so they can focus on occupant classification
sensors.

According to representatives of some vehicle manufacturers, their goal is to
install occupant classification sensors 1 year before the September 2003
deadline in order to get real- world experience with the performance of the
sensors. However, a number of vehicle manufacturers have expressed

Page 18 GAO- 01- 596 Vehicle Safety

concerns about their ability to develop occupant classification sensors that
comply with the advanced air bag rule by the deadline- slightly more than 2
years away. Despite the fact that manufacturers have been working on
technologies for occupant classification sensors for several years, the
development of these technologies has not yet reached the level that a new
technology would normally have reached to be ready for installation within
that time frame. In 1998 and 1999, NHTSA reported that vehicle manufacturers
anticipated having occupant classification sensors installed in model year
2000 vehicles. 16 However, accuracy, durability, and manufacturing issues
were more difficult to overcome than anticipated. For example, General
Motors anticipated installing a pattern- based sensor in its model year 2000
Cadillac Seville but abandoned this plan in part because the sensor did not
perform accurately under different humidity and temperature settings, and
the production process was so variable that only 10 percent of the sensors
that were produced were suitable for installation in vehicles. More
recently, in March 2001, after 2 years of work, a company that was to be the
primary supplier of a weight- sensing system for DaimlerChrysler decided to
abandoned work on the project due to technical reasons. As a result,
DaimlerChrysler is reevaluating its options for occupant classification
sensors.

NHTSA officials have met with industry representatives to discuss their
efforts to develop advanced air bag systems. According to NHTSA officials,
although vehicle manufacturers have stated that it will be difficult to
develop occupant classification sensors by September 2003, none of the
manufacturers have indicated that they will not be able to meet the
deadline. 17

Due to the uncertainty associated with developing occupant classification
sensors, NHTSA plans to stay abreast of manufacturers? progress by holding
periodic meetings with manufacturers. These meetings may be informal
meetings that occur as NHTSA gathers information about technologies or more
formal meetings for manufacturers to provide an update on the status of
their progress. NHTSA also plans to conduct research on the feasibility of
occupant classification technologies, including laboratory research on
specific occupant classification

16 Advanced Air Bag Technology Assessment, Prepared for NHTSA and the
National Aeronautics and Space Administration by the Jet Propulsion
Laboratory, April 1998; Air Bag Technology in Light Passenger Vehicles,
NHTSA, December 16, 1999, Revision 1.

17 Should a manufacturer be unable to meet the deadline, the manufacturer
may apply to NHTSA for a temporary exemption from the requirements.

Page 19 GAO- 01- 596 Vehicle Safety

technologies and monitoring the performance of occupant classification
technologies as they are installed in vehicles.

Other technological challenges described by manufacturers include designing
an air bag that can generate enough power to protect an average adult male
as well as deploy in a manner that does not severely injure a smaller
occupant and developing crash sensors that can distinguish among the various
types of crash tests required in the rule. In addition to the technological
challenges of developing an advanced air bag system, manufacturers and
suppliers are concerned about the accuracy and repeatability of some of the
test procedures in the rule and using humans rather than dummies to test
suppression systems. These concerns were highlighted in petitions for
reconsideration of the rule filed by the manufacturers. (See app. II for
further information on these petitions for reconsideration and NHTSA?s
response.)

NHTSA?s reported expenditures on advanced air bag R& D increased from about
$6.3 million in fiscal year 1998 to nearly $7.0 million in fiscal year 2000.
(See table 2.) According to NHTSA officials, these expenditures were
primarily for activities related to the development of the advanced air bag
rule, such as investigations of crashes involving an air bag- related injury
or fatality, evaluations of the performance and characteristics of air bag
systems, and studies to determine how people are injured or killed by air
bags. NHTSA officials estimate that expenditures on advanced air bag R& D
will increase to $7.2 million in fiscal year 2001. According to NHTSA
officials, future expenditures will focus on monitoring the performance of
advanced air bags and continuing the R& D of specific technologies. NHTSA?s
planned activities include analyzing the protection provided by advanced air
bags in real- world crashes, conducting crash tests- including tests at
various speeds and angles with belted and unbelted crash dummies- to
evaluate the performance of advanced air bags, and researching advanced air
bag technologies that are anticipated to be ready for installation in
vehicles in the next 3 to 5 years. NHTSA plans to conduct some of this
research through cooperative agreements with air bag suppliers. Federal and
Industry

Expenditures on Advanced Air Bag Research and Development Have Increased
Since 1998

Page 20 GAO- 01- 596 Vehicle Safety

Table 2: NHTSA Expenditures on R& D for Advanced Air Bag Technologies,
Fiscal Years 1998- 2001

Dollars in thousands

Category 1998 1999 2000 2001( est.)

Special crash investigations $1,331 $1,553 $1,553 $1,553 Air bag systems 1,
850 2,431 2,431 2,480 Biomechanics (crash dummies) 3, 150 3,000 3,000 3,200

Total $6,331 $6,984 $6,984 $7,233

Note: These expenditures do not include other NHTSA expenditures that
indirectly benefit air bag R& D, such as the New Car Assessment Program and
compliance testing.

Source: NHTSA.

Vehicle manufacturers did not provide information on their individual
expenditures for advanced air bag R& D because they consider this
information confidential. Instead, an industry association- the Alliance of
Automobile Manufacturers- coordinated with manufacturers to provide
aggregated information on the extent to which expenditures have changed and
are anticipated to change compared to calendar year 1998. Four
manufacturers- Ford, General Motors, Nissan, and Toyota- provided
information on expenditures for resources related to one or more of the
following categories: staffing (including expenditures to support supplier
staffing); technology development and testing; prototype parts; engineering
resources; in- house testing and data analysis, analytical performance
assessment (computer modeling), physical test properties and test costs; and
implementation and integration of technologies into vehicles. According to
the Alliance, these expenditures generally total between $20 million and $30
million per vehicle ?platform? (a group of vehicles that utilize the same
basic design). 18 The aggregated information from the four manufacturers
shows that their expenditures increased by about 275 percent from 1998 to
2000 and are anticipated to increase overall by nearly 375 percent from 1998
through 2003. (See fig. 3.) According to the Alliance, the estimated
increase from 1998 through 2003 is due to the cost of designing and
installing advanced air bag systems for an increasing number of vehicle
platforms to meet the phase- in requirements in the advanced air bag rule.
These expenditures are estimated to decrease after advanced air bag systems
have been installed in vehicles.

18 A single vehicle platform could encompass one or several models of
vehicles. According to the Alliance and vehicle manufacturers, the 8 vehicle
manufacturers included in our review have 88 vehicle platforms.

Page 21 GAO- 01- 596 Vehicle Safety

Figure 3: Percentage Change in Four Vehicle Manufacturers? Advanced Air Bag
R& D Expenditures, 1999 Through 2004

Note: 1998 is the baseline year. The Alliance adjusted each company?s
expenditures by its current market share and averaged the sum of the
manufacturers? adjusted expenditures for each year.

Source: Alliance of Automobile Manufacturers.

We provided a draft of this report to the Department of Transportation for
its review and comment. The Department did not provide an overall assessment
of our draft report. Rather, Department representatives, including the
Director of NHTSA?s Office of Vehicle Safety Research, provided one
technical comment through e- mail. Specifically, the Director suggested that
several manufacturers may have the necessary technologies for occupant
classification sensors that can distinguish among various sizes of
occupants, even though they may not have installed them in vehicles on a
large scale. We verified with auto manufacturers that they have not
installed occupant classification sensors that can distinguish among various
sizes of occupants and are still developing such sensors for Agency and
Industry

Comments and Our Evaluation

Page 22 GAO- 01- 596 Vehicle Safety

frontal air bag systems that are intended to meet the requirements of the
advanced air bag rule.

We provided portions of our draft report to vehicle manufacturers and air
bag suppliers for review to verify the accuracy of our descriptions of
advanced air bag technologies and challenges in meeting the requirements of
the advanced air bag rule. The manufacturers and suppliers generally agreed
with our draft report and offered several technical corrections, which we
incorporated as appropriate.

We are sending copies of this report to congressional committees and
subcommittees responsible for transportation safety issues; the Secretary of
Transportation; the Executive Director, National Highway Traffic Safety
Administration; the Director, Office of Management and Budget; and other
interested parties. We will make copies available to others upon request and
on GAO?s home page at http:// www. gao. gov.

If you or your staff have any questions about this report, please contact me
at (202) 512- 2834. Key contributors to this report were Judy
GuilliamsTapia, Bert Japikse, James Ratzenberger, Phyllis Scheinberg, and
Sara Vermillion.

John H. Anderson, Jr. Managing Director, Physical Infrastructure Issues

Appendix I: Scope and Methodology Page 23 GAO- 01- 596 Vehicle Safety

To determine the current availability of and planned improvements to
advanced air bag technologies, we collected and analyzed information from
eight vehicle manufacturers (BMW, DaimlerChrysler, Ford, General Motors,
Honda, Nissan, Toyota, and Volkswagen) and the five companies that are the
primary suppliers of air bag systems in the United States (Autoliv, Breed,
Delphi, Takata, and TRW). According to the Alliance for Automobile
Manufacturers, the combined sales for the eight manufacturers account for
over 90 percent of vehicles sold in the United States. We did not
independently verify the information we received from manufacturers and
suppliers. We reviewed literature on automotive technology for descriptions
of the technologies used in advanced air bag development. We also met with
officials from NHTSA, representatives from the Insurance Institute for
Highway Safety, the National Transportation Safety Board, university
researchers, and consumer groups.

To identify the challenges, if any, that the industry faces in complying
with the advanced air bag rule, we reviewed the requirements of the rule and
discussed these requirements with representatives of vehicle manufacturers
and companies that supply air bags. We also reviewed comments on the rule
submitted by manufacturers and suppliers.

To identify the changes in federal expenditures on advanced air bag R& D, we
collected data on NHTSA?s expenditures for fiscal years 1998 through 2001
and analyzed the changes in the individual categories of expenditures.
Vehicle manufacturers did not provide information on their individual
expenditures for advanced air bag R& D because they consider this
information confidential. Therefore, to identify the changes in industry
expenditures on advanced air bag R& D, we collected aggregated information
from an industry association on the extent to which four manufacturers?
expenditures have changed since 1998. We did not independently verify this
aggregated information.

We conducted our work from July 2000 through May 2001 in accordance with
generally accepted government auditing standards. Appendix I: Scope and
Methodology

Appendix II: Selected Aspects and Status of the Advanced Air Bag Rule

Page 24 GAO- 01- 596 Vehicle Safety

The advanced air bag rule requires that future air bags be designed to
create less risk of serious injury from air bags- particularly for small
women and young children- and to improve frontal crash protection for all
occupants. To achieve these goals, the rule includes requirements for
additional test procedures using different sizes of dummies than were
included in previous requirements. These new requirements will be phased in
during two stages. During the first stage phase- in- from September 1, 2003,
to August 31, 2006- an increasing number of each manufacturer?s vehicles
must be certified each year as meeting the requirements in the advanced air
bag rule. During the second stage phase- in- from September 1, 2007, to
August 31, 2010- the speed for one of the tests (the belted test for the
50th percentile adult male dummy) will be increased from 30 to 35 miles per
hour (mph) and, similar to the first phase- in period, an increasing number
of each manufacturer?s vehicles must be certified each year.

In comments to the supplemental notice of proposed rulemaking on advanced
air bags, there was a difference of opinion on whether the maximum speed for
the unbelted rigid barrier crash test should be set at 25 or 30 mph. In the
final rule, NHTSA set the maximum speed at 25 mph on an interim basis while
the agency continues to investigate whether the higher speed is more
appropriate. After the rule was issued in May 2000, consumer safety groups,
vehicle manufacturers, and air bag suppliers filed petitions for NHTSA to
consider changing certain provisions in the rule. NHTSA plans to respond to
these petitions in July 2001.

To minimize risk to infants and children on the passenger side, the rule
includes provisions for the air bag to be suppressed or deployed in a
?lowrisk?

manner that is much less likely to cause serious or fatal injury. (See fig.
4.) For newborn infants in car beds, the rule requires that the air bag be
suppressed. For 1- year- old infants in child seats, 3- year- old children,
and 6- year- old children, manufacturers are allowed to install systems
designed for suppression or low- risk deployment. Manufacturers may choose
different strategies for different occupants. For example, a manufacturer
could design an air bag system that would suppress the air bag for infants
and deploy the bag in a low- risk manner for 3- and 6- yearold children.
Appendix II: Selected Aspects and Status of

the Advanced Air Bag Rule Selected Requirements in the Advanced Air Bag Rule

Appendix II: Selected Aspects and Status of the Advanced Air Bag Rule

Page 25 GAO- 01- 596 Vehicle Safety

Figure 4: Test Options in the Advanced Air Bag Rule to Meet Suppression and
LowRisk Deployment Requirements for Smaller Occupants

a This option allows manufacturers to suppress the air bag if the occupant
(3- year- old or 6- year- old on the passenger side or small female on the
driver side) moves ?out of position,? or close to the air bag module during
a crash, as occurs in precrash braking. NHTSA has not adopted performance
requirements and test procedures for such a system that would ?dynamically?
suppress the air bag, but may do so at a later date. Several manufacturers
told us that they are not planning to install such a system before September
2003 because the technologies needed for this system- dynamic occupant
position sensors- are not sufficiently developed.

Source: NHTSA.

To test for suppression on the passenger side, the dummies are placed in
their appropriate child seats that are, in turn, placed on the passenger
seat. 1 These tests may be conducted under various scenarios: using any of
several models of safety seats; with the passenger seat in the forward,
middle, or rear position; unbelted or belted with up to 30 pounds of tension
on the belt; with any handles and sunshields on infant safety seats in fully
open and fully closed positions; or with a towel or blanket on the infant
safety seats. For the 3- and 6- year- old dummies, tests will also be
conducted with the dummies unbelted and in various positions, such as

1 Newborn infant dummies are placed in car beds. One- year- old dummies are
placed in rearfacing or forward- facing child safety seats. Three- year- old
dummies are placed in forwardfacing child seats or booster seats. Six- year-
old dummies are placed in booster seats.

Appendix II: Selected Aspects and Status of the Advanced Air Bag Rule

Page 26 GAO- 01- 596 Vehicle Safety

sitting back or sitting on the front edge of the seat. 2 The rule requires
that the car have a ?telltale? light that, after the dummy is in place,
indicates whether the air bag is suppressed or activated. Following each
suppression test with an infant or child, a dummy representing a small (5th
percentile) adult female will be placed in the passenger seat to ensure that
the air bag is not suppressed for small adults.

To test for low- risk deployment on the passenger side, a 1- year- old dummy
is placed in one of several models of rear- or forward- facing child seats
on the passenger seat in the forward position on the seat track. The seat
belt may be cinched with up to 30 pounds of tension. For the 3- and 6- year-
old dummies, the unbelted dummy is placed ?out of position? with their head
or chest on the air bag module to simulate the situation where an unbelted
child is close to the instrument panel due to sudden braking immediately
before a crash. After the dummy (infant or child) is in place, the air bag
is deployed. The amount of ?injury? that occurs to the head and neck of the
dummies (and the chest of the 3- and 6- year- old child dummies) must be
below criteria specified in the rule.

To minimize risk to small drivers, the rule includes provisions to deploy
the air bag in a low- risk manner, similar to the low- risk deployment tests
for 3- and 6- year- old children. 3 The tests are conducted by placing the
5th percentile adult female dummy out of position, with the chin on the
steering wheel rim or on the air bag module. The air bag is then deployed
and the resulting ?injury? to the head, neck, chest, and legs is measured.
NHTSA has determined that, when all of the combinations of the various
testing scenarios are considered, there are 129 tests for suppression and
low- risk deployment: 95 suppression tests for infants in a car bed or child
seat, 28 suppression tests for 3- and 6- year- old children, 4 low- risk

2 Crash dummies do not sufficiently represent the human physiology, so the
rule allows the use of appropriately sized humans to test the performance of
the technologies in static suppression tests. The University of Michigan
Transportation Research Institute is developing dummies that more accurately
represent the human shape.

3 The rule also allows manufacturers the option of suppressing the air bag
if the occupant moves out of position during a crash, as occurs in precrash
braking. However, NHTSA has not developed performance requirements and test
procedures for a system that dynamically suppresses the air bag. In
addition, manufacturer representatives told us that they are not planning to
install such a system before September 2003 because the technologies needed
for this system- dynamic occupant position sensors- are not sufficiently
developed.

Appendix II: Selected Aspects and Status of the Advanced Air Bag Rule

Page 27 GAO- 01- 596 Vehicle Safety

deployment tests for 3- and 6- year- old children, and 2 low- risk
deployment tests for 5th percentile female drivers.

To improve protection in frontal crashes for occupants of different sizes,
the rule includes seven tests that involve crashing vehicles into barriers
at different speeds and angles and with dummies representing average (50th
percentile) adult males and 5th percentile women, belted and unbelted. (See
fig. 5.) Four of the tests are conducted with dummies that represent 50th
percentile adult males and three are conducted with dummies that represent
5th percentile adult females. After the crash test, the resulting injury to
the head, neck, chest, and legs of the dummies must not exceed the limits
specified in the rule. The offset deformable barrier test was included in
the requirements to ensure that manufacturers upgrade their crash sensors as
necessary to prevent late air bag deployments in crashes that are less
abrupt than those into rigid barriers. NHTSA did not include a requirement
for an unbelted crash test at an oblique angle using a 5th percentile female
dummy because the agency determined that the requirement for this type of
crash using a 50th percentile male dummy would result in an air bag that is
sufficient to protect smaller occupants as well.

Appendix II: Selected Aspects and Status of the Advanced Air Bag Rule

Page 28 GAO- 01- 596 Vehicle Safety

Figure 5: Tests Required in the Advanced Air Bag Rule to Improve Protection
for Occupants of Different Sizes

a Maximum speed will increase to 35 mph after September 1, 2007. Source:
NHTSA.

The rule will be phased in during two stages. The first stage phase- in-
from September 1, 2003, to August 31, 2006- requires an increasing number of
vehicles to be certified as passing all of the above tests each year. (See
fig. 6.) During the second stage phase- in- from September 1, 2007, to
August 31, 2010- the speed for the belted test for the 50th percentile adult
male dummy will be increased from 30 to 35 mph. As with the earlier
requirements, an increasing percentage of vehicles must comply with the new
test speed each year.

Appendix II: Selected Aspects and Status of the Advanced Air Bag Rule

Page 29 GAO- 01- 596 Vehicle Safety

Figure 6: Timeline for Phase- in of Requirements in the Advanced Air Bag
Rule

Source: GAO analysis of information from NHTSA.

In the comments to the supplementary notice of proposed rulemaking for the
advanced air bag rule, there was a significant difference of opinion on
whether the top speed for the unbelted rigid barrier crash test should be
set at 30 mph or 25 mph. 4 Comments from those who favored setting the
maximum test speed at 30 mph- safety groups such as Public Citizen,

4 While the speed of a rigid barrier crash test at 30 mph is 20 percent
higher than a test at 25 mph, it represents a 44 percent increase in crash
energy. Comments on the

Maximum Speed for the Unbelted Rigid Barrier Test

Appendix II: Selected Aspects and Status of the Advanced Air Bag Rule

Page 30 GAO- 01- 596 Vehicle Safety

Center for Auto Safety, Consumers Union, and Parents for Safer Air Bags-
included:

 half of all fatalities in frontal crashes involve a change in velocity
greater than 30 mph, so a maximum test speed of 30 mph represents
significantly more potentially fatal crashes than a test speed of 25 mph;

 in crash tests conducted by NHTSA, almost all vehicles with redesigned air
bags passed the 30 mph unbelted rigid barrier test with the 50th percentile
male dummy, so air bags would not have to be more aggressive (and
potentially more risky to small occupants) to meet a 30 mph test;

 lowering the test speed to 25 mph would not offer improved protection, as
required in the Transportation Equity Act for the 21st Century;

 advanced technologies can be used to enable all vehicles to meet
requirements for high speed protection and risk reduction; and

 a 25 mph test speed would not encourage the use of advanced technologies.

Comments from proponents of a 25 mph maximum test speed- such as vehicle
manufacturers, air bag suppliers, the Insurance Institute for Highway
Safety, and the National Transportation Safety Board- included:

 redesigned air bags work well and there has been no loss in protection;

 a 25 mph test speed allows flexibility to design air bags for all
occupants;

 a return to a 30 mph test would require a return to overly powerful air
bags;

 there are significant technological challenges in meeting a 30 mph
requirement for both the 50th percentile adult male dummy and the 5th
percentile adult female dummy; and

 advanced technologies are not currently available that address problems
posed by air bags designed to a 30 mph test. NHTSA concluded that, given the
uncertainty associated with simultaneously achieving improved protection for
occupants of all sizes without compromising efforts to reduce the risks of
injury to smaller occupants, a conservative approach should be taken.
Consequently, NHTSA set the maximum speed for the unbelted rigid barrier
test at 25 mph. However, the agency issued that part of the rule as an
interim final rule and announced that it would issue a final rule after it
monitors the performance of advanced air bags and determines whether
increasing the maximum test speed to 30 mph would offer any advantages over
a test speed of 25 mph. To monitor the performance of advanced air bags,
NHTSA plans to, among other things, evaluate real- world crash data, perform
compliance testing and publish an annual report on the extent to

Appendix II: Selected Aspects and Status of the Advanced Air Bag Rule

Page 31 GAO- 01- 596 Vehicle Safety

which advanced air bags comply with requirements, conduct crash tests, and
conduct research on specific advanced air bag technologies.

After the final rule was issued in May 2000, consumer safety groups, vehicle
manufacturers, and air bag suppliers submitted petitions to NHTSA for the
agency to reconsider certain aspects of the rule. Consumer safety groups-
Center for Auto Safety, Consumer Federation of America, Parents for Safer
Air Bags, and Public Citizen- jointly filed a petition. The petition states
that separate phase- in schedules for cars and sport utility vehicles are
feasible because some cars can already meet a 30 mph unbelted rigid barrier
crash test, but sport utility vehicles have more difficulty in complying
with this test due to their stiffer frame, which produces a harder crash
pulse and requires a more aggressive air bag than passenger cars.
Specifically, the petition requests NHTSA to require the industry to meet a
30 mph unbelted rigid barrier test for passenger cars and a 25 mph test for
sport utility vehicles, to be increased to 30 mph at a later date. The
petition also requests NHTSA to: (1) add requirements for tests to simulate
lower- speed, softer crashes in which the air bag deploys late and strikes
an occupant who has moved forward before the air bag deploys, (2) require
manufacturers to meet a 35 mph belted barrier test with the 5th percentile
female dummy as well as the 50th percentile male dummy, and (3) require that
manufacturers conduct all barrier tests in both the perpendicular and
oblique modes, including tests using the 5th percentile female dummies.

Petitions from some vehicle manufacturers and air bag suppliers state that
the directions for some tests in the rule, particularly those related to the
positioning of dummies in suppression and low- risk deployment tests, need
to be clarified. For example, some petitioners claimed that the procedure
for positioning the child dummies for the low- risk deployment test do not
always result in the dummies being against the air bag module, as intended.
Other issues raised in petitions from manufacturers include: (1) limiting
the amount of time required to collect data on the dummies? injuries during
a low- risk deployment test in order to minimize inclusion of injury from
interior components other than the air bag, (2) requesting that a generic
child restraint test device be developed so that humans will not have to be
used to test air bag suppression systems, and (3) reducing the upper limit
on the amount of tension that can be applied to a seat belt.

According to NHTSA officials, NHTSA is drafting a final response to these
petitions and plans to issue the response in July 2001. Some consumer safety
groups and vehicle manufacturers have told us that they are concerned about
the timeliness of NHTSA?s response to the petitions. Some manufacturers have
raised concerns that the issues with the compliance test procedures may not
be resolved in time for them to Petitions for

Reconsideration

Appendix II: Selected Aspects and Status of the Advanced Air Bag Rule

Page 32 GAO- 01- 596 Vehicle Safety

finalize their advanced air bag system designs and perform the required
testing to certify that the vehicles meet the requirements in the advanced
air bag rule.

Appendix III: Multistage Frontal Air Bag Systems in Model Year 2001 Vehicles
and Model Year 2002 Vehicles in Production as of April 1, 2001

Page 33 GAO- 01- 596 Vehicle Safety

Stages of inflation Types of occupant sensors that affect air bag deployment

Manufacturer/ make/ model year/ vehicle line

Percentage of U. S.

fleet a Driver Passenger Type of crash

severity sensor b Occupant

presence sensor on passenger

side c Seat

position sensor on

driver side d

Belt use sensor on driver and passenger

sides e BMW

All 2001 vehicles 100 2 2 Single- point electronic

DaimlerChrysler

Chrysler/ Dodge/ Jeep

2001 Sebring/ Stratus Sedan 5 3 3 Single- point electronic 2001 Sebring
Convertible 2 2 2 Single- point electronic 2001 Caravan/ Voyager/ Town and
Country Minivans

16 3 3 Single- point electronic 2002 Jeep Liberty 7 3 3 Multipoint
electronic

Mercedes- Benz

2001 C- class 1 2 2 Single- point electronic f

2001 S- class 1 1 2 Single- point electronic f

2001 CL- class <1 1 2 Single- point electronic f

2001 ML- class 2 2 2 Single- point electronic f

Ford Motor Company

Ford/ Mercury/ Lincoln

2001 Taurus/ Sable 12 2 2 Multipoint electronic

2001 Crown Victoria/ Grand Marquis

6 2 2 Multipoint electronic

2001 Town Car 2 2 2 Multipoint electronic

2001 Windstar 5 2 2 Multipoint electronic g

Volvo

Appendix III: Multistage Frontal Air Bag Systems in Model Year 2001 Vehicles
and Model Year 2002 Vehicles in Production as of April 1, 2001

Appendix III: Multistage Frontal Air Bag Systems in Model Year 2001 Vehicles
and Model Year 2002 Vehicles in Production as of April 1, 2001

Page 34 GAO- 01- 596 Vehicle Safety

Stages of inflation Types of occupant sensors that affect air bag deployment

Manufacturer/ make/ model year/ vehicle line

Percentage of U. S.

fleet a Driver Passenger Type of crash

severity sensor b Occupant

presence sensor on passenger

side c Seat

position sensor on

driver side d

Belt use sensor on driver and passenger

sides e

2001 S40 1 2 2 Single- point electronic

2001 S60 <1 2 2 Single- point electronic

2001 S80 1 2 2 Single- point electronic

2001 V40 1 2 2 Single- point electronic

2001 V70 1 2 2 Single- point electronic

2001 Cross Country <1 2 2 Single- point electronic

Jaguar

2001 XK Coupe and Convertible 2 2 2 Single- point electronic

General Motors

2001 Chevrolet Impala 4 2 2 Multipoint electronic 2001 Chevrolet Monte Carlo
1 2 2 Multipoint electronic 2001 Buick Lesabre 3 2 2 Multipoint electronic
2001 Oldsmobile Aurora 1 2 2 Multipoint electronic 2001 Pontiac Bonneville 1
2 2 Multipoint electronic 2002 Cadillac Seville 2 2 2 Multipoint electronic
2002 Buick Rendezvous 1 2 2 Multipoint electronic 2002 Chevrolet Trail
Blazer 2 2 2 Multipoint electronic 2002 GMC Envoy <1 2 2 Multipoint
electronic

2002 Oldsmobile Bravada <1 2 2 Multipoint electronic

Honda

Honda

2001 Accord 32 2 2 Single- point electronic

2001 Civic 28 2 2 Multipoint electronic with mechanical front sensors

Appendix III: Multistage Frontal Air Bag Systems in Model Year 2001 Vehicles
and Model Year 2002 Vehicles in Production as of April 1, 2001

Page 35 GAO- 01- 596 Vehicle Safety

Stages of inflation Types of occupant sensors that affect air bag deployment

Manufacturer/ make/ model year/ vehicle line

Percentage of U. S.

fleet a Driver Passenger Type of crash

severity sensor b Occupant

presence sensor on passenger

side c Seat

position sensor on

driver side d

Belt use sensor on driver and passenger

sides e

Acura

2001 CL 5 1 2 Single- point electronic

2001 MDX 3 1 2 Single- point electronic

2001 RL 1 1 2 Single- point electronic

2001 TL 5 1 2 Single- point electronic

2002 RL N/ A 1 2 Single- point electronic

2002 TL N/ A 2 2 Single- point electronic

Nissan

2001 Maxima/ Infiniti I30 19 2 2 Multipoint electronic 2002 Q45 N/ A 2 2
Multipoint electronic

Toyota

2001 LS430 1 1 2 Multipoint electronic Legend: N/ A = not available Note:
Table includes U. S. market vehicles only. The checkmarks in the columns for
occupant sensors indicate that these sensors have been installed on the
corresponding vehicle. a These figures are the approximate percentage that
each specified vehicle line represents of the

company?s light vehicles produced during the model year for the U. S.
market. b Single- point electronic crash sensors are generally located
within the control module in the

passenger compartment. In a multipoint crash sensing system, one electronic
crash sensor is located within the control module and one or more electronic
or mechanical crash sensors are located in the front (crush zone) of the
vehicle. c An occupant presence sensor detects weight on the passenger seat
to determine if the seat is occupied. If the sensor does not detect an
occupant over a specified weight, the system deactivates the passenger air
bag. d Seat position sensors identify whether the driver?s seat is forward
or rearward on the seat track.

When the seat is positioned forward, indicating that the driver is seated
close to the steering wheel and the air bag module, the system deactivates
the high output stage of the driver?s air bag. e Belt use sensors detect
whether the front seat occupants are wearing their seat belts. The system

deploys the air bag at a higher crash velocity threshold if an occupant is
buckled and a lower threshold if an occupant is unbuckled.

Appendix III: Multistage Frontal Air Bag Systems in Model Year 2001 Vehicles
and Model Year 2002 Vehicles in Production as of April 1, 2001

Page 36 GAO- 01- 596 Vehicle Safety

f The occupant presence sensing system in these vehicles has the capability
to detect the presence of a compatible child seat in the front passenger
seat. When the sensor detects the presence of such a child seat, through a
tag in the bottom of the child seat, the system deactivates the passenger
air bag. g An occupant presence sensing system was incorporated in the front
passenger seat of model year

2001 Windstars beginning in May 2001. Source: GAO analysis of information
provided by vehicle manufacturers.

Appendix IV: Current Availability of and Anticipated Improvements in
Advanced Air Bag Technologies

Page 37 GAO- 01- 596 Vehicle Safety

Description: These sensors detect a crash and differentiate levels of crash
severity. They provide this input to the control module, which uses it to
make decisions about whether the air bag should be deployed and, if so, what
level of deployment is appropriate.

Advanced technologies currently available: The crash sensors in vehicles
equipped with multistage frontal air bags have been refined to better
discriminate crash severity levels, so that the appropriate level of air bag
deployment can be determined. Some of these vehicles have a ?singlepoint?

electronic crash sensor, which is generally located within the control
module in the passenger compartment. Others have ?multipoint?

crash sensing systems, in which one sensor is located within the control
module and one or more sensors are located in the front (crush zone) of the
vehicle. In some of these multipoint systems all of the sensors are
electronic while in others the sensor in the control module is electronic
and the sensors in the front of the vehicle are mechanical. Some
manufacturers have placed additional sensors in the front of the vehicle in
order to produce more information on the crash earlier in the crash event,
allowing additional time to determine crash severity and make the
appropriate deployment decision. However, others have not yet installed up-
front crash sensors in their multistage air bag systems, because the
performance of these sensors in ?real- world? conditions can be affected by
irrelevant ?background noise,? such as extraneous vibrations that occur
during a crash event. (See app. III for information on the types of crash
sensing systems installed in model year 2001 and early model year 2002
vehicles with multistage air bags.)

Anticipated improvements: In order to enhance the ability of their vehicles
to comply with the requirements in the advanced air bag rule, manufacturers
plan to improve the ability of their crash sensing systems to distinguish
among levels of crash severity as well as identify the type of crash, such
as a frontal rigid barrier crash, a pole crash, or an offset deformable
barrier crash. As part of this effort, manufacturers plan to refine and
increase the use of multipoint crash sensing systems. Appendix IV: Current
Availability of and

Anticipated Improvements in Advanced Air Bag Technologies

Technologies Currently Available and Planned Enhancements to Meet
Requirements Crash Sensors

Appendix IV: Current Availability of and Anticipated Improvements in
Advanced Air Bag Technologies

Page 38 GAO- 01- 596 Vehicle Safety

Description: These sensing devices installed in the interior of the vehicle
are intended to identify characteristics of the occupants, such as their
belted status and size. They provide this input to the control module, which
uses it to make decisions about whether the air bag should be deployed and,
if so, what level of deployment is appropriate.

Advanced technologies currently available: Some vehicles currently equipped
with multistage frontal air bag systems have occupant sensors that provide
information such as seat position, occupant presence, seat belt use, and
identification of a child seat. However, occupant sensors currently
available in U. S. market vehicles are not capable of distinguishing among
different sizes of occupants, such as whether the passenger is a child or an
adult. A limited number of vehicles with multistage air bags contain seat
position sensors on the driver side that identify whether the seat is
forward or rearward on the seat track. If the seat is positioned forward,
indicating that the driver is seated close to the steering wheel and the air
bag module, the system deploys the air bag with reduced force. Some vehicle
models contain weight- based sensors on the passenger side that identify
whether the seat is occupied. If the sensor does not detect an occupant over
a specified weight, the system deactivates the passenger air bag. These
sensors are intended to prevent unnecessary deployment of the passenger air
bag. Some vehicle models also contain, as part of their multistage air bag
systems, seat belt use sensors on the driver and passenger sides that
identify if the occupant is wearing the seat belt. The system deploys the
air bags at a higher crash velocity threshold if the occupant is buckled and
a lower threshold if the occupant is unbuckled. Finally, a limited number of
vehicles with multistage air bags contain child seat sensors that identify a
tag in the bottom of a compatible child seat. The system deactivates the
passenger air bag when it detects the tag. NHTSA considers these child seat
sensors to be an excellent supplement to other occupant classification
systems. However, NHTSA will not allow manufacturers to use this sensing
system alone to comply with the rule?s requirements because it would be
difficult to ensure that tags would be properly installed on the wide
variety of child seats used by the general public. (See app. III for
information on the types of occupant classification sensors installed in
model year 2001 and early model year 2002 vehicles with multistage air
bags.)

Anticipated improvements: To comply with the advanced air bag rule,
manufacturers anticipate increasing the use of driver seat position sensors
and driver and passenger seat belt use sensors. In addition, sensors based
on weight classification and/ or pattern recognition will be installed on
the passenger side to distinguish among adults, children, and child seats.
Technologies being considered primarily include load cells, pressure
Occupant Classification

Sensors

Appendix IV: Current Availability of and Anticipated Improvements in
Advanced Air Bag Technologies

Page 39 GAO- 01- 596 Vehicle Safety

bladders, and pattern/ pressure mats. 1 Load cells are electro- mechanical
devices located at each attachment point of the seat frame to the vehicle.
They estimate the force applied to the seat, allowing the system to classify
an occupant based on their seated weight. A pressure bladder is a fluid- or
air- filled bladder located under the seat cushion and above the seat frame.
The system classifies an occupant based on the amount of pressure applied to
the bladder. A pattern/ pressure mat contains multiple sensor elements and
is located between the seat cushion and upholstery. The system classifies an
occupant based on the amount of pressure applied to the mat and the pattern
of the occupant?s imprint on the seat. Some of these technologies may need
to be augmented by seat belt tension sensors to identify whether the amount
of force applied to the seat is due in part to the seat belt rather than the
occupant?s weight. This is important information for identifying child
seats. Occupant position sensors, which are described in the section below,
may also be used to enhance occupant classification sensors.

Description: This central processing unit stores the vehicle?s sensing
algorithms, computational systems that interpret and analyze inputs from the
crash sensors and occupant sensors to determine whether the air bag should
be deployed and, if so, what level of deployment is appropriate. In order to
deploy the air bags in time to restrain the occupants, the control module
must predict during the initial part of the crash whether a crash is
occurring that exceeds a predetermined severity threshold. The control
module generally triggers deployment from 10 to 100 milliseconds after the
start of the crash, depending on the type of crash.

Advanced technologies currently available: Multistage air bag systems
contain control modules with sensing algorithms of increased complexity that
can determine the appropriate level of air bag deployment, based on
available inputs. In multistage air bag systems that include occupant
sensors and/ or multipoint crash sensing systems, the algorithms process the
additional inputs provided by these sensors in making deployment decisions.

Anticipated improvements: To comply with the advanced air bag rule, control
modules will require algorithms of greater complexity that will be able to
interpret and analyze additional inputs concerning crash scenarios

1 Manufacturers are also considering other technologies, such as capacitive
coupling sensors, which are embedded in the seat cushion and classify the
occupant by detecting the disruption in a low energy electric field. Control
Module

Appendix IV: Current Availability of and Anticipated Improvements in
Advanced Air Bag Technologies

Page 40 GAO- 01- 596 Vehicle Safety

and types of occupants and use this information in making appropriate
deployment decisions. Manufacturers also intend to make further improvements
in control modules to increase the speed of processing of inputs and the
accuracy of deployment decisions. As algorithms become more complex, it may
be necessary in the longer term to move from a centralized control module to
a system in which the processing and decision- making functions are
decentralized, because of dramatic increases in the amount of information
being input and in the computations needed.

Description: Multistage inflators have two charges that can generate two or
more stages of inflation. Firing one charge generates low- level deployment;
firing both charges simultaneously or in sequence generates higher levels of
deployment.

Advanced technologies currently available: A number of vehicle models
currently have frontal air bag systems with multistage inflators. Most of
these vehicles have inflators with two stages of inflation (low- and
highlevel deployment), while a limited number have inflators with three
stages of inflation (low-, medium-, and high- level deployment). In most
cases, the multistage inflators are on both the driver and passenger sides,
but in some cases only the passenger side has a multistage inflator. In
currently available multistage air bag systems, the deployment level is
triggered based on crash severity and, in some cases, driver seat position.
In addition, as previously explained, some of these systems deploy the air
bags at different crash severity thresholds for belted or unbelted occupants
and/ or deactivate the passenger air bag if a sensor detects that the
passenger seat is empty or contains a child seat. (See app. III for
information on the multistage air bag systems installed in model year 2001
and early model year 2002 vehicles.)

Anticipated improvements: To comply with the advanced air bag rule,
manufacturers will use multistage inflators with two or more stages of
inflation. Some manufacturers have told us that their introduction of
inflators with more than two stages of inflation depends on further
advancements in crash sensors, occupant sensors, and the control module in
order to be able to reliably determine the appropriate level of inflation.
Various inflation technologies are under development that may provide
continuous variation in inflation, rather than inflation in discrete stages,
allowing greater adaptiveness to inputs provided by crash and occupant
sensors. For example, one such technology would use a variable electric
current to continuously control the rate of gas generation during inflation
of the air bag. Inflators with these technologies may be introduced by
Multistage Inflators

Appendix IV: Current Availability of and Anticipated Improvements in
Advanced Air Bag Technologies

Page 41 GAO- 01- 596 Vehicle Safety

some manufacturers during the 3- year phase- in period for complying with
the advanced air bag rule.

Description: In addition to characteristics of the inflators, some air bag
design features can reduce the aggressivity of the deploying air bag and,
therefore, the likelihood of serious injury caused by deployment. These
features include the location of the air bag module (which contains the
inflator and the air bag) and characteristics of the bag itself, such as
folding, shape, compartments, tether straps, and venting.

Advanced technologies currently available: Manufacturers have made a variety
of changes in bag design in order to make deployment less likely to cause
injury. The location, folding, and shape of frontal air bags have been major
areas of design change. On the driver?s side, air bag modules have been
recessed into the steering wheel in many vehicles to add space between the
driver and the deploying air bag. Also, many driver air bags now have a fold
pattern and shape that allows the bag to deploy in a radial manner, so that
the initial ?burst out? inflation force will inflate the bag laterally
rather than rearward toward the driver. On the passenger?s side,
manufacturers often locate the bag module in a ?top- mount? position on the
instrument panel, to increase the distance between the occupant and the
deploying bag, or use a smaller- sized bag if it is located in a ?midmount?

position in front of the passenger on the instrument panel. Some newer fold
and shape designs and venting schemes can make the deploying air bag
adaptive to the position of the occupant. For example, some passenger air
bags in use today contain a fabric flap attached to the bag, known as a
?bias flap,? which directs the initial burst out inflation of the bag to the
side of and away from the occupant if he or she is out of position (in close
proximity to the bag) at the time of deployment. Some bags have variable
venting designs that inflate the bag more softly if it is obstructed during
deployment, indicating that the occupant is out of position. Tethers, which
are strips of fabric connecting the front and back panels of the bag, have
been incorporated into some bag designs on both the driver and passenger
sides to reduce extension of the bag and help position it more quickly when
it is deployed.

Anticipated improvements: Although air bag design is a relatively mature
technology, manufacturers and air bag suppliers continue to work on
improvements in this area in order to enhance the ability of air bag systems
to comply with the advanced air bag rule. Concepts under development that
may become available in the longer term include venting systems that will
work with multistage inflators to increase the Air Bag Features That

Minimize Risk to Occupants

Appendix IV: Current Availability of and Anticipated Improvements in
Advanced Air Bag Technologies

Page 42 GAO- 01- 596 Vehicle Safety

adaptability of deployment (by controlling inflator output based on input
from sensors) and bags with multiple compartments that inflate sequentially.

Description: These sensors are intended to determine the proximity of an
occupant to the air bag. Sensing devices installed in the interior of the
vehicle would enable the system to suppress the air bag if an occupant is
out of position and too close to the air bag.

Anticipated technologies: Infrared, ultrasonic, capacitive, and optical
technologies are being researched to develop ?dynamic? sensors that can
continuously track an occupant?s position with respect to the air bag
module. Infrared sensors utilize an array of invisible infrared light beams
projected across the passenger compartment to identify the position of an
occupant. For ultrasonic sensors, ultrasonic transducers emit sound waves
and the sensors monitor the sound waves that are reflected by an occupant.
Capacitive sensors utilize an electric field to identify the position of an
occupant by detecting moisture in the body and optical sensors monitor the
position of an occupant. While ?static? ultrasonic sensors that periodically
determine the occupant?s position may be installed on a limited basis to
augment occupant classification sensors before 2003, researchers are not yet
certain whether or when dynamic occupant position sensing will become widely
used.

Description: These sensors would identify the position, approach angle,
velocity, and mass of objects prior to a collision and allow more time for
the air bag system to respond.

Anticipated technologies: Radar (radio wave) is the principal technology
being researched. Precrash sensing technologies are still in the early
stages of research to determine if the concept is feasible. Therefore, it is
not yet clear whether or when this technology might become available. The
primary challenge in developing this technology is identifying the
Additional

Enhancements to Advanced Air Bag Systems Anticipated for the Longer Term

Occupant Position Sensors Precrash Sensors

Appendix IV: Current Availability of and Anticipated Improvements in
Advanced Air Bag Technologies

Page 43 GAO- 01- 596 Vehicle Safety

mass of an object in the path of the vehicle. ?Adaptive cruise control,?
which automatically reduces a vehicle?s speed to maintain a safe distance to
vehicles in front, is currently available on some vehicles and is considered
by some to be a precursor of precrash sensing.

(348239)
*** End of document. ***