The Head-On Debate Over Crash Safety : Automobiles: U.S. and industry officials are locked in a controversy over which dummies best simulate a motorist’s injuries in an accident.
There’s a room in the back of First Technology Safety System’s assembly plant that looks a lot like a torture chamber.
Inside, the company’s auto crash test dummies are strapped into pendulum-like contraptions and left to the not-so-tender mercies of First Technology employees, whose job it is to drop their vinyl-skinned heads onto a metal plate, crush their stomachs with rigid steel cylinders and send 50-pound weights hurtling into their chests.
“They go flying all over the place,” says George Pitarra, president of what is now the world’s only auto dummy manufacturer of consequence. “It’s quite a mess.”
But sadism is not the object here, all appearances to the contrary. These plastic and steel creations, wired with sensors to simulate human injuries in a crash, are the keys to building safer cars.
The problem is, not all dummies are created equal.
With 45,000 people dying on U.S. roads each year, government and auto industry officials are locked in a controversy over which dummies do the best imitation of a human in a crash--a debate in which technology often takes a back seat to politics.
Critics say the dummies that government and industry rely on to reduce vehicle deaths are so dated that the information they provide about safety may be misleading. And a small circle of biomechanical engineers--whose crash-tests of cadavers and pigs have helped bring these automatons into the modern age--are frustrated as the new, more lifelike generation of dummies they had envisioned are not further developed because of lack of research and funding.
As a result, engineers complain, safety designs for today’s cars are based on anachronistic dummies that weren’t intended to operate in a world of air bags, child seats, plastic fenders, energy-absorbing bumpers and other latter-day alterations. In addition, their measurements are based on a mythical average man and do not take into account the safety needs of anyone who isn’t 5 feet, 8 inches tall and 165 pounds--in other words, most of the population.
Auto safety dummies date to the mid-1950s, when an Air Force colonel named John Stapp persuaded U.S. and European auto makers that the crude, sand-filled canvas sacks he was using to design safer planes could help keep more of their customers alive and well enough to buy new cars.
Through a process of decidedly unnatural selection, governed by shifting political interests and consumer trends, dummies have evolved into sophisticated, uncanny human surrogates. They can measure, among other things, how hard an ill-fated motorist’s head will hit the dash and how many ribs he’s likely to break during the 120 milliseconds it takes for a car to slam into a wall at 30 m.p.h.
Virtually all the world’s major auto makers now bang up whole families of dummies in hundreds of crash tests a year--most of them on laboratory sleds set up to generate crash conditions without trashing hundreds of cars in the process. The “injuries” that the dummies sustain in these tests give manufacturers an idea of how a car’s design needs to change to be safer for the occupant.
Crash-test research has for years been dominated by General Motors, the same company that recently attracted the ire of animal-rights groups for using live rodents and pigs in some safety tests. Long recognized as the industry leader in safety research, GM developed most dummies now used as standards worldwide.
Although there are about 25 different types of dummies manufactured today, the main fuss is over creatures named Hybrid II, Hybrid III and Sid.
They are the dummies that the National Highway Traffic Safety Administration picked to measure compliance with the government’s occupant safety requirements. Hybrid II and III measure injury potential in head-on collisions, while as of 1994 Sid (for side-impact dummy) will measure injury potential in side collisions.
Under the law, the dummies must register injuries to specifications set by the government in order to be sold in the United States. NHTSA does about 20 random tests a year “to keep them honest,” says George Parker, the agency’s director of research and development.
Auto makers are not technically required to use test dummies or even do their own crash tests, but government regulation and growing consumer demand for safer vehicles has in recent years fueled a five-fold increase in the dummy business--generating an estimated $9 million in sales in 1991.
“People are interested in surviving a car crash and surviving it well,” Pitarra observes.
But while each year brings new car models and designs, only a handful of auto crash dummies have been developed since the early 1970s, and even fewer have been incorporated as standards by NHTSA.
“The dummies have a way of getting frozen as soon as they are chosen as a certain device by the government,” says Dr. Ian Lau, a small, intense senior engineer at GM’s Biomedical Science Department in Warren, Mich. “The problem for us is that the (safety) questions keep changing.”
The Hybrid II, widely used for frontal collision tests, was designed in 1973, when seat belts were unfashionable and car makers’ primary safety concern was preventing an unbelted driver from ramming into the steering wheel hub. Now, in the age of buckle-up laws and air bags, it is generally believed that dummies fail to accurately mimic an occupant’s interaction with the vehicle’s modern safety contraptions.
It is hard to tell, for instance, whether a motorist wearing a seat belt will “submarine” in a crash, sliding under the belt and into the lower dashboard, because the pelvic geometry in the Hybrid II was not designed with that possibility in mind.
GM has since developed a “frangible abdomen,” a foam insert that crushes under increasing force. Measuring the deformation of what Lau refers to as the “dummy tummy” helps to detect submarining and assess risk of abdominal injury.
But the dummy tummy is used only in the Hybrid III dummy, which NHTSA doesn’t require auto makers to use. The Hybrid III is more sensitive to injury in some cases than the Hybrid II, making safety tests harder to pass. So most auto makers use the older dummy in certifying their cars, and thus don’t benefit from the latest tummy technology.
What’s more, the current front-impact standard requires only the test of “50th percentile male” dummies. So while a 5-foot, 8-inch, 165-pound male can be relatively confident of surviving a frontal collision, the rest of the population has more cause to worry.
“It’s probably a sexist dummy,” Barry Felrice, NHTSA’s director of rule making, concedes of the Hybrid II. “Twenty years ago, most drivers were male, so they picked the average male, and that’s who they wanted to protect.”
As to why, two decades later, the “average” male is still the standard, Felrice points to what he calls the “tortured history” of the regulation itself. First proposed in 1971, the front-impact standard the dummy was designed for was killed and revived repeatedly according to the politics of the moment. It didn’t take full effect until last year.
A similar fate seems likely for the more advanced Hybrid III, equipped to measure 31 responses to crash-test forces as compared to Hybrid II’s eight. Developed by GM in 1983, it’s unlikely to be adopted before the late 1990s.
Crash-test dummies also got tangled up in political machinations earlier this year when GM tried but failed to persuade the government to use another of its dummies, called Biosid.
GM was pushing Biosid as preferable to Sid, the government’s side-impact dummy, contending that Biosid was more biofidelic, or lifelike. GM argued that Sid was too stiff, and that using it as a yardstick for side-impact protection would prompt auto makers to make car interiors stiffer, potentially endangering occupants in real-world crashes.
But critics said GM was merely using the more advanced Biosid, which was not ready for use by auto makers, to try to delay NHTSA’s enactment of new side-impact standards.
The scientific and engineering debate over the dummies turned “quite nasty,” said Dr. Lawrence Schneider, a research scientist at the University of Michigan’s Transportation Research Institute, which is working on the next generation of frontal impact dummies.
About a year ago, NHTSA head Jerry R. Curry bowed to pressure from Congress to adopt the new regulation and use Sid as the standard. The side-collision standard is expected to save 1,200 lives a year.
Curry then flew to Detroit to try to talk GM Chairman Robert C. Stempel out of challenging NHTSA’s decision. Stempel publicly backed down, and rumors flew that GM had agreed not to fight the government’s allegedly inferior dummy in return for promised NHTSA support on other issues.
At the time, GM did not respond to the allegations. After all, there was no mileage for the company in reminding motorists that the technology for making cars safer is so imperfect--or in undercutting a Bush Administration agency that is considered friendly to the U.S. auto industry. In a recent interview, Stempel said there was no deal between the government and GM.
“We don’t want to do anything to weaken (NHTSA’s) credibility, anything that would hurt that department,” Stempel says now. “The part that Jerry and I didn’t want to do was see an argument over a dummy, conducted by the media, blow our cooperative research apart. So what we really did was just simply agreed, ‘Let’s keep this thing out of the papers (and) in the laboratory.’ ”
Dismissing such political maneuverings as beyond their ken, the biomechanical engineers on both sides prefer to devote their attention instead to more fundamental problems, such as how the bumps and bruises sustained by their dummies translate into human injuries.
“All we can do is measure how hard the dummy is hit,” says Lau. “The missing link is, how do you know he got hurt?”
The Hybrid II gauges skeletal injury by measuring acceleration at one point on the spine. Lau and his colleagues bemoan the limitations of such data, which fails to account for injury to soft tissues such as the liver, spleen and heart.
Through research that literally put pigs behind the wheel, GM’s biomedical team recently came up with a method of predicting injury by measuring velocity and compression of body parts that they say will make for more accurate and meaningful results from dummy tests.
In an example of how data provided by dummies can translate into safer car designs, GM used the results to modify its “self-aligning” steering wheel, now built with six flexible metal legs that allow the wheel to bend with the driver on impact, reducing the chance of injury.
Most dummy development is based on belting cadavers into car seats and observing the injuries they sustain when slammed against walls at varying speeds in laboratory tests. But there are a number of drawbacks to modeling dummies wholly on cadavers, not the least of which is that it’s hard to measure injury to a heart that’s already stopped beating and a nervous system that’s stopped functioning. In addition, cadavers are in short supply.
That’s where animals come in--and where an animal rights group called People for the Ethical Treatment of Animals recently joined the debate. The group accuses GM of killing 20,000 animals over the last decade, including the pigs sacrificed for the steering wheel.
“There is no need for apology in my mind,” Lau says. “Dummies do not just get invented and automatically relate to human injury. We have to develop a better means of measurement, and I am very confident that with animal tests and cadaver tests, supplemented with a lot of dummy tests, we get the right answer.”
But biomechanical engineers have watched in dismay as the money to do badly needed dummy research dwindles. The federal budget deficit, anti-regulatory forces and the auto industry’s financial woes have combined to drain funds that scientists were counting on to fortify the next generation of dummies.
Several research facilities that had biomechanical research programs have abandoned the field because of the scarcity of grants, further reducing the resources available for dummy research.
“The problem over the years has been that there’s really minimal dollars going toward this major social problem that’s causing so many deaths,” Schneider says.
