The Social Network for Cars: Test of the Nation’s First Wireless Collision Avoidance System
A Boston-area security tech company and the University of Michigan are involved in one of the most ambitious—and potentially controversial—transportation projects of our time. It could have major impact on federal legislation, and almost everyone you know.
Picture this: You’re driving in your car, approaching an intersection. Maybe you’re speeding a little, going 40 mph in a 35 zone, say. Unbeknownst to you, another driver is racing down the cross street and is about to run a red light (probably texting or something). This could spell disaster. But instead, your car picks up a wireless signal from the other vehicle. A beeping sound or flashing light on your dashboard alerts you to slow down, so you hit the brakes. Disaster averted.
Now let’s take it a step further. Maybe the alert is hooked into your car’s control system, so if you don’t put on the brakes, your car does it automatically. And maybe that’s fine with you. But you might be a little worried about giving up that kind of control in life-and-death situations. After all, computers get hacked and software crashes. Not to mention, you might not want your car broadcasting its speed and location out there for all to see (especially not the cops, since you were speeding).
This scenario isn’t the future. It’s happening already—at least the driver-alert part. In six cities around the U.S., trials of about 100 drivers each are underway to see how people react to in-car alerts (such as collision warnings, do not pass, and vehicle stopped ahead). But the next step is bigger. In Ann Arbor, MI, some 3,000 cars will be equipped with onboard wireless devices for communicating with each other and signaling to drivers when there’s an imminent hazard. This 12-month pilot study, which was announced recently and starts next August, is being led by the University of Michigan’s Transportation Research Institute (UMTRI) through a $14.9 million contract from the U.S. Department of Transportation. The state of Michigan has been heavily involved as well.
The goal of the federal initiative is, ultimately, to save lives. In the U.S., auto accidents are the leading cause of death for people aged 15 to 34; more than 30,000 people are killed on the nation’s roadways each year. The hope is that with new early-warning systems in place, a sizable fraction of would-be victims could be saved—some say 80 percent of non-alcohol-related cases—especially when high speed is involved.
The idea of wirelessly connected cars isn’t new, of course. The field of vehicle telematics has been around for years, with applications in fleet management, tracking, and safety. But advances in GPS location technologies, wireless communications, sensors, hardware, and algorithms are enabling smarter, better-connected vehicles to be tested on a bigger scale. And recent breakthroughs such as autonomous road-racing vehicles and Google’s self-driving car are starting to propel the technology into the mainstream.
Nevertheless, the Michigan study raises some serious real-world concerns. “This is a massive system with tremendous security and privacy implications,” says Ed Adams, the chief executive of Security Innovation in Wilmington, MA. And that’s exactly where his software security firm comes in.
Security Innovation developed the mobile software being used in the U-M study to encrypt and authenticate the wireless messages being sent and received (and make them as anonymous as possible). In fact, the company’s chief scientist, William Whyte, has been working on the wireless standard for connected vehicles since 2003. (It has been established as IEEE 1609.2, for all the techies out there.) Whyte was also involved in a smaller proof-of-concept field study outside of Detroit that finished in 2008. He is a consultant to the broader Department of Transportation research program.
The proof of concept study showed “the technology works,” says Whyte, a crypto-security expert. Now the broader safety study will show whether the “technology works for people,” he says. “Do they find it distracting? Do they ignore it? Are they worried about privacy?”
The Michigan study will recruit mostly non-professional drivers using their regular cars on a daily basis, Whyte says. (The program manager at UMTRI didn’t respond to a request for comment.) Each car will be outfitted with an onboard device that includes a radio, antenna, and processor to broadcast its status and to receive wireless updates about the location, speed, and status of other vehicles. The radio signals, which have a range of a few hundred meters, could also be used to communicate with roadside equipment such as networked traffic lights.
Security Innovation’s software goes out of its way to “preserve drivers’ privacy as much as possible,” Whyte says. To that end, the wireless data isn’t stored anywhere in the system, and the messages (which are broadcast 10 times a second) don’t have identifying markers that link them with any particular vehicle. They have a security certificate attached, and the certificate is changed every five minutes, for every car, he says. The firm has licensing agreements with GPS vendors such as TomTom and Garmin, which sell their devices to automakers like Ford, Toyota, and Kia (they are all part of a large consortium of companies and organizations involved in connected vehicle technologies).
If the U-M study goes well—and that’s definitely an “if” at this point—the Department of Transportation is expected to seek a mandate through Congress, which would require all vehicles made in the U.S. to have onboard wireless networking equipment. Whyte estimates that such a mandate could pass by 2015, in which case all 2018 models would have the technology, and older cars could be retrofitted. Adams, the CEO, is confident that appropriate legislation will pass. “It’s going through,” he says. (Another Boston-area connection is that the Volpe National Transportation Systems Center in Cambridge, MA, is leading some of the policy work in this area.)
So, is wireless networking legislation going to be this generation’s version of seat belt laws? Possibly. One challenge is that early adopters won’t get as much benefit as later ones, because the technology requires other cars involved to be networked as well. On the other hand, if enough vehicles become equipped, a network effect should ensue (see Facebook).
Looking farther down the road, Whyte and others see computers and smart networks gradually taking over more driving responsibilities from humans. One futuristic idea gaining some popularity is to have a train of cars traveling at highway speed, just a foot or so apart from each other, so as to maximize speed, aerodynamics, and fuel efficiency. For safety reasons, that would require using autonomous vehicles that employ some form of data messaging and short-range radar, Whyte says.
“You’ll see gradual adoption for certain use cases,” he says. “But it will be 15 or 20 years before it’s the norm.”