The concept of automatic restraint systems for passenger vehicles has been around for more than 40 years. However, it is in only the last 10 years that the system has found wide acceptance in the form of airbags and, indeed, is mandated for all cars and light trucks sold in the United States. The National Highway Traffic Safety Administration (NHTSA) estimates that airbags have saved more than 3,800 lives, but in the last few years it has been shown that airbags can, under certain conditions, cause serious injury or fatalities. NHTSA has proposed a series of new rules to address this problem.

The Need For Improved Air Bag Safety

The first airbag patent in the U.S. was awarded to John Hetrick in 1952. Since that time, supplemental restraint systems have proved to be fertile ground for inventors, with more than 20,000 patents granted worldwide to date. However, the adoption of airbag technology occurred at a far slower pace. The first vehicle with optional airbags was produced by General Motors in 1978. Mercedes-Benz offered the first standard air bags in 1980. Later, the U.S. Federal Government adopted Federal Motor Vehicle Safety Standard 208 (FMVSS 208), which called for mandatory airbags for drivers and passengers to be phased in by the 1999 model year. The additional protection was embraced by the vast majority of car buyers, and the installation of airbags quickly ramped up ahead of the required dates. Through 1998 the Insurance Institute for Highway Safety (IIHS) reports that 3.9 million air bags have deployed in the US, saving 3,808 lives. However, in late 1996 reports began to surface of airbags causing serious injury or death in certain circumstances. The population most at risk was shown to be children in rear-facing infant seats placed in front of an airbag; unbelted, out-of-position occupants; and drivers sitting very close to the steering wheel. Through 1998 the Insurance Institute reports 125 deaths caused by airbags inflating in low severity crashes. NHTSA and the industry quickly began investigating the cause of these incidents and concluded that the root cause was a vehicle occupant's close proximity to the deploying airbag. The region of high risk is defined by a "keep-out" zone, which is related to the vehicle's air-bag design, and the size, position, and fragility of the occupants. Contributing to the risk is that most airbag systems in the U.S. fleet are "one-size-fits-all" -- optimized for a male passenger without a safety belt in a 30 mph rigid frontal barrier crash. But real-world crash events are highly variable in nature -- they can happen at any speed, with different levels of severity, and a wide variety of occupant sizes, ages, and positions. Compounding the problem are the relatively fast timelines involved. During a frontal crash, unrestrained vehicle occupants continue to move forward under their own momentum until they impact some portion of the vehicle's interior. A properly deployed airbag provides a much softer impact surface than a steering wheel or other interior surface, but the bag must be fully inflated before impact to provide the maximum benefit. For the nominal 30 mph barrier crash event, the time from impact to full deployment is in the order of 50 milliseconds. During this time the occupant will move about 5 cm forward relative to the vehicle. It takes the bag about 30 milliseconds to deploy, leaving 20 milliseconds for the sensor system to determine the crash profile and begin the deployment sequence. Because of the short amount of time allowed for deployment for the nominal crash pulse, the bag must inflate aggressively. An occupant who comes in contact with an air bag in the early stages of deployment (in the keep-out zone) is at risk of injury from the air bag. However, only a small percentage of frontal crashes mirror the 30 mph rigid barrier conditions. Many types of crashes are into deformable barriers such as other vehicles, collapsible barrels, or vegetation. These have the effect of a longer crash pulse, which increases the crash timeline. Precrash braking may also bring the occupants forward in the vehicle--in effect decreasing the timeline because they are closer to the keep-out zone before the actual beginning of the crash. In either case, the presence of a seatbelt acts as an improvement--acting early in the crash or during precrash braking to restrict free movement relative to the vehicle.

What Can Be Done?

To improve airbag performance, more information is needed about the crash and the occupants, along with a means of tailoring the deployment profile. Crash sensors that can better analyze the need for occupant protection are the first step. These improvements can come in the number and placement of sensors, sensor bandwidth, improved crash software algorithms, and more powerful processors to handle the additional sensor load. The second step would be a means to identify which occupants are wearing seat belts. Third is a sensor or suite of sensors to determine occupant position relative to the keep-out zone. These may be augmented with sensors to determine the occupant's size (weight), seat position, and a means to detect a rear-facing infant seat. Once this information is in hand it is possible to control the time delay from the first crash instant to the onset of deployment. However, to broaden the effective safe area dramatically, the airbag inflation (and deflation) rate must be variable--either continuously or in discrete steps.

Regulatory Action

US government regulations tend to drive the automotive industry, because there is no cross-national equivalent regulations in the other large automotive consuming regions of the world. However, this does not imply that there is no interest in airbag safety in these regions. Quite the contrary, many of the technology innovations that will be used to meet these requirements come from these areas. Also, it should be pointed out that most of Europe has had mandatory seat belt laws and regulations prohibiting children in the front passenger seat for several years--greatly improving air bag safety statistics. In 1997 NHTSA amended FMVSS208 to allow for the depowering of airbags. Depowering reduces the inflation rate and pressure of the air bag. By decreasing the keep-out zone, the risk to drivers who are close to the steering wheel and drivers and passengers who are out of position is reduced. However, depowering may negatively impact air-bag effectiveness for large, unbelted occupants. NHTSA is also allowing disconnection of airbags (both permanently and temporarily, with a switch) for people who are unavoidably at risk from airbag inflation. Disconnection must be done by petition, and authorization is required before these modifications are made. NHTSA made both of these rulings temporary--until such a time as advanced airbag technology is available to implement so-called "smart airbags." In September 1998, NHTSA released its Notice for Proposed Rulemaking (NPRM) on advanced airbag technology (Federal Docket No. NHTSA 98-4405). Like FMVSS208, which the NPRM proposes to modify, the NPRM does not specify how to build an advanced airbag system but rather defines the test conditions under which an airbag system will be approved for production. The NPRM defines several new test cases designed to minimize the dangers to children (including those in rear-facing child seats), short-statured drivers, and out-of-position occupants. In some cases it allows the manufacturer to select from one of multiple procedures, thus providing for a wide variety of possible implementations. It also calls for tests using a wider range of crash pulses. The NPRM also defines the phase-in time for the U.S. fleet and eliminates depowering and deactivation as the new requirements are met. The phase-in time includes early implementation credits--giving manufacturers some control of implementation across their fleet. The phase-in timeframe is from the model year beginning Sept. 1, 2002, with full compliance in the model year beginning Sept. 1, 2005.

Potential New Technologies

As part of the investigation leading up to the Advanced Air Bag NPRM, NHTSA commissioned the NASA Jet Propulsion Laboratory to study airbag performance and assess the potential technologies that may be available to OEMs in the near future. This expansive report contains a detailed overview of the many technologies that may be useful to the safety system designer in the future. JPL concludes that the following technologies may be available for Model Year 2001: *Two-stage inflators that inflate softly for lower-threshold crashes and full inflation for high-threshold velocity. *Crash sensor/control systems with improved algorithms for better air bag deployment discrimination, better threshold control, and proper inflation level for two-stage inflators. *Belt-use sensors that allow the deployment threshold to be raised when belts are in use. *Seat position and seat belt spool-out sensors that give an approximate measure of occupant size and proximity to the air-bag module. *Static-proximity (occupant-position) sensors that identify occupants in the keep-out zone. (JPL indicated that an aggressive development program would be needed to meet this timeframe and would have some limitations for out-of-position occupants.) *Automatic suppression that would prevent inflation when sensors determine that an occupant is in the keep-out zone. *Compartmented air bags, radial deployments, and lighter-weight bag fabrics to reduce the size of the keep-out zone. *Advanced belts, including pretensioners that better couple the occupant to the seat, and load limiters that reduce the maximum belt loads on the occupant. A long list of technologies may be available for Model Year 2003: *Continued improvement in crash sensor/control system algorithms. *Wide use of belt-use sensors. *Integrated occupant and proximity sensors that will identify occupants in the keep-out-zone or those who will be forced into it prior to air-bag deployment. *Precrash sensors -- although their application requires further investigation. *Automatic suppression used with proximity sensors to prevent inflation. *Multistage inflators that tailor a response for occupants and crash severity. *Continued improvement in bag design and materials -- further reducing the keep-out zone. *Increasing use of pretensioners and load limiters. *Inflatable belts that provide both pretensioning and better belt-load distribution. In each of these areas there is significant industry work to reduce these technologies to practice. An example is in the area of occupant and proximity sensors. Devices have been demonstrated that use pressure, strain gauge, piezo effect, infrared, ultrasonic, capacitive, and radio-frequency (RF) sensors, to name a few. Some of these are beginning to be fielded on production vehicles. It is certain that in a few years the safety designer will have a wide variety of technologies to choose from. However, all this airbag-related content will somehow have to be packaged into the vehicle interior, which will pose a problem in itself.