Connected cars see danger, warn others and stop crashes
Photo credit: MARK BIALEK
| Sponsored by |
| » | ||
| » | ||
| » | ||
| » | ||
| » |
Reprints
Print
Respond
DETROIT -- "It can see," says Farid Ahmed-Zaid. "We can't."
Ahmed-Zaid, an intelligent-vehicles technical expert for Ford Motor Co., is talking about the connected-vehicle technology installed on both the Ford Fusion we're in and the one his colleague, Joseph Stinnett, is driving near a facility in Farmington Hills, Mich., where the devices are being researched.
Ten times a second, each is transmitting data on speed, location, brake status and path history through a dedicated communications channel. Both cars are equipped to receive these signals and alert the driver to potential threats that are invisible to mere humans.
From what I can see -- or perhaps "feel" would be a better word -- Ahmed-Zaid is trying to kill me: accelerating with Stinnett's car stopped a few hundred feet ahead of us; going through a green light when Stinnett is running the red light the other way; trying to move into Stinnett's lane when he's in our blind spot. But then a light flashes or a buzzer sounds or the seat vibrates, Ahmed-Zaid brakes, the crash is averted, and we live to drive another day.
Which is, of course, the whole point. The connected-vehicle technology is meant to bring a dramatic reduction in traffic fatalities by helping drivers avoid the types of crashes that most often cause them: collisions in intersections, or while changing lanes, or rear-ending stopped cars. The National Highway Traffic Safety Administration estimates that this technology could lower the death rate in about 80 percent of crashes that don't involve drunk drivers.
Although he's employed by Ford, Ahmed-Zaid has worked for the last few years under the aegis of the Crash Avoidance Metrics Partnership, a consortium of eight automakers whose mission is to develop an industrywide communications system that's practical, effective and accepted by motorists.
They think they've got it. But the federal government, which is funding the lion's share of this enterprise, needs to know if the safety benefits are sufficient to require its implementation. That's why approximately 2,800 cars, trucks and buses in northeast Ann Arbor, Mich., will be tooling around until next August with some form of this technology on board.
'Real-world' results
"Does the system work in the hands of real drivers, in real vehicles, in a real environment, not a test track or a lab?" asks Jim Sayer. He's a research scientist at the University of Michigan Transportation Research Institute, which is conducting the field test for the U.S. Department of Transportation. Sayer is the project manager for what's called the Connected Vehicle Safety Pilot Model Deployment.
"Environmental issues, obstructions such as buildings, foliage on trees, can all have impacts on the quality of that safety message signal transmission," Sayer says. "We don't really know how much, and we don't know how frequently those kinds of situations will be encountered."
The systems have been developed using technology from the CAMP partners and their suppliers. Other participating automakers are Mercedes-Benz, General Motors, Toyota, Honda, Nissan, Hyundai-Kia and Volkswagen.
The technology's developers also don't know if the system will work at 70 mph, or if the assigned bandwidth is ample, or if drivers can trust that the warning they get comes from a legitimate threat and not a hacker on an overpass.
Will they tolerate another tier of buzzers and lights? Or will they think it's a pain, like a "check engine" light that won't go out? Do they have concerns about the information that's being broadcast?
Scale, privacy, security and "interoperability" -- the ability of cars made by various manufacturers to communicate -- are the four categories of issues the consortium has defined, and this test blends driver acceptance into the mix. As Sayer points out, the project will be more helpful in some areas than others.
"Interoperability testing had to be done before we could deploy the units," he says, "making sure that all of the different devices are behaving the way we expect them to and can actually communicate with each other. Scalability is only being tested to a limited degree due to the fact that only a small percentage of vehicles in this area are being equipped. With security, we are exercising protocols that are similar to what might be rolled out in a real-world deployment."
And there has to be some wiggle room regarding privacy, because this is a research study. "I have to be able to identify a device that breaks down," says Sayer, "and that isn't necessarily something that would happen in the real world, where the driver would probably get an indication from the car."
• The dedicated short-range communications system functions similar to WiFi.
• Transmitted signals travel over wireless communication channels allocated by the Federal Communications Commission for automotive use.
• Signals are emitted 10 times per second and have a range of 900 feet in all directions.
• The systems do not identify specific drivers.
• Members of the Crash Avoidance Metrics Partnership are Ford, General Motors, Honda, Hyundai-Kia, Mercedes-Benz, Nissan, Toyota and Volkswagen. They used hardware from suppliers.
Studying driver behavior
Of the 2,800 vehicles in the study, 64 cars and three trucks are fully integrated, with safety systems installed during production that both transmit and receive data and warn drivers of potential threats. Another 300 have aftermarket safety devices that do the same, as do 16 trucks and three transit buses. The rest have an aftermarket vehicle awareness device that sends basic safety messages.
Although all the participants are volunteers, Sayer says they're "a pretty representative cross-section of the population." Who got which vehicles depended on who drove the most in the project area, roughly the northeast quadrant of Ann Arbor.
"We're not really looking for a positive or negative outcome," he says. "We're not expecting to see an impact on the crash rate in Ann Arbor, for example. There aren't enough of them. The project is evaluating the state of the technology and initial user acceptance. That's why the DOT wants this data; they will examine it to look for potential changes in driver behavior."
The current study is actually the second phase of the Safety Pilot of Connected Vehicle Technology. The first was a series of six driver acceptance clinics conducted over a six-month period ending last January. The 688 subjects were exposed to two hours of situations, like the ones Ahmed-Zaid provided me, at test tracks around the country.
"I spent six months on eggshells" waiting for the results, he says. They were pretty astounding: 89.4 percent of the participants said they would like to have the technology in their vehicles, and 90.9 percent said they believed it would improve safety.
There was plenty of initial skepticism, says Ahmed-Zaid. "I had people tell me there was nothing better than defensive driving," he recalls, "then they'd get into a blind scenario and say, 'You know, that happened to me just the other day.'"
Cooperation vs. competition
When Ahmed-Zaid is wearing his CAMP hat, he's working on the technology that allows Fords to "talk" to Toyotas and Chevrolets. As a Ford employee, his focus is on developing proprietary methods for how its vehicles use that technology.
"I walk a very fine line between cooperation and competition," he says. "I have to, though, because we need to develop this technology as quickly and affordably as possible."
One knowledgeable source said that with volume production, the system should cost about $100 per car.
Once the capacity to warn the driver is established, what will differentiate the automakers' products is how and when those warnings are communicated. Will the driver hear buzzers, horns, or bells and whistles? Or see a red light? Will the seat vibrate? And how grave a situation will it take to trigger whatever the methods may be? Or, in technical parlance, what is the threat assessment algorithm?
Ahmed-Zaid has been down this road before. Early in his career he worked on adaptive cruise control, which was launched in 1999 and is now is in wide use.
"It really hits you when your neighbor starts talking about things such as radar technologies that are now pretty common but were once highly confidential projects you worked on," he says.
How far off that day is for vehicle-based warning devices is one of the things the Safety Pilot will help determine, but the ball is being moved. Federal officials are scheduled to decide next year whether to require such systems, although researchers say it could be at least five years before they appear on production vehicles.
"I think this is a big piece of the puzzle," says Sayer, "if for no other reason than it gives all the parties involved a level of confidence moving forward in addition to helping to address technical questions and refine the technology. It's not just the simulation of a crazy idea that worked on the bench or worked on the test track.
"Thirteen months in, I'm more geeked than ever."
• 91% said the system was desirable; 1% said it was undesirable.
• 78% said they fully understood the system; 22% said they understood it in large part but still had questions.
• Among age groups, 83% in their 20s and 30s said they fully understood the system, compared with 78% of people in their 40s and 50s and 71% of people in their 60s and 70s.
Source: National Highway Traffic Safety Administration study
You can reach Jeff Mortimer at (Unknown address).
