When an aircraft goes down, the first thing the National Transportation Safety Board looks for is the infamous “black box.” Though the term is a misnomer-there are two boxes and they’re both orange-the idea of capturing what went wrong in an accident is so compelling that three companies are adapting the idea to automobiles. While the concept is intriguing, its practical application has yet to traverse several obstacles.
To create what’s known as an “event data recorder” for cars, these companies plan to harness the output of the sensors that already exist to deploy airbags and control anti-lock braking systems. The recorder would use half a kilobyte of flash memory to record ten seconds of data-five seconds before the crash and five seconds after the crash. It would tally such information as speed at impact, how long the brakes had been applied, how many times the car was hit-all factors that can help determine the cause of a crash and who was at fault.
“Right now, in the world of crash investigation,” says Ricardo Martinez, CEO of Atlanta-based Safety Intelligence Systems, one of three manufacturers of event data recorders, “they use the BOGSAT method. That stands for a bunch of guys sitting around talking,’ coming up with their best guess.”
Efforts are underway to change this. Last August, the National Highway Traffic Safety Administration ran a Ford F-150 truck into a wall at 30 mph in order to determine how accurately test units from three companies-Safety Intelligence Systems, Independent Witness of Salt Lake City, and DriveCam Video Systems of San Diego, as well as an unnamed internal recorder supplied by Ford-tracked the delta in the vehicle’s velocity. The agency compared the devices against three of its own Endevco accelerometers, and found Independent Witness’s device to be the closest to its baseline devices; it suggested that the Safety Intelligence Systems and DriveCam devices needed further development (see the report).
Importantly, the test proved that data from an event data recorder could provide objective evidence for crash investigations, reduce the reliance on eyewitness reports, and reveal design flaws more quickly. If Ford and Firestone had been able to gather event data on accidents involving Sport Utility Vehicle tires, they might have avoided their recent liability and public relations disaster.
To that end, Martinez, an emergency room physician who served as head of the National Highway Traffic Safety Administration, is also working with IBM and the Insurance Services Office of Jersey City, NJ, an aggregator of statistical and actuarial data for the insurance industry. The group is creating a database of accident records to be used collectively by car manufacturers, law enforcement agencies, and insurance companies.
Crash investigators or auto body shops could offload the contents of the flash memory into this database, dubbed the Global Data Safety Vault. Originally scheduled to launch in mid-2002, the database won’t begin to be populated until at least the beginning of 2003. Though the car manufacturers have already agreed on a standard plug for downloading recorder data, agreement on standard data formats has proved harder to reach.
For instance, of the hundreds of onboard diagnostics codes that the Environmental Protection Agency has mandated for emissions testing, few if any are consistent among the Big Three. The Global Safety Vault initiative faces a similar normalization challenge, and the Institute of Electrical and Electronics Engineers, a leading standards body, has joined the effort to get input on what data should be captured for its P1616 standard.
There’s also the question of where to place the recorder in a vehicle. Martinez suggests putting it near the airbag module, which is close to the crush zone of the passenger compartment, since both have to not only withstand a collision but keep working afterwards (the same sensors that trigger the airbags also trigger distress calls from certain telematics systems). In the test, Ford centered its recorder behind the dashboard and under the radio.
Even with these unanswered questions, Martinez believes that event data recorders will first appear in the 2007 model year, the earliest year in which they could be added. He estimates the cost at only around $5. The benefits to recorders and the planned Global Data Safety Vault are compelling: police could investigate accidents quickly and accurately; manufacturers could identify potential recall problems more quickly; and insurance companies could more easily identify fraudulent accident claims (currently about 20 percent of all claims filed). “These are the dark ages of crash data,” insists Martinez, “and we’re trying to replace bad information with good information.”
Saudi Arabia plans to spend $1 billion a year discovering treatments to slow aging
The oil kingdom fears that its population is aging at an accelerated rate and hopes to test drugs to reverse the problem. First up might be the diabetes drug metformin.
Yann LeCun has a bold new vision for the future of AI
One of the godfathers of deep learning pulls together old ideas to sketch out a fresh path for AI, but raises as many questions as he answers.
The dark secret behind those cute AI-generated animal images
Google Brain has revealed its own image-making AI, called Imagen. But don't expect to see anything that isn't wholesome.
A quick guide to the most important AI law you’ve never heard of
The European Union is planning new legislation aimed at curbing the worst harms associated with artificial intelligence.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.