While World Fixates On ‘Autopilot,’ GM Bets On A Revolution Within Reach
Detroit has a plan to change how we drive--and it hits the road in 2016
Robotic cars marketed to anyone but the wealthiest consumers are a distant dream at best, but that hasn’t stopped auto manufacturers from touting their whiz-bang appeal.
In September the pitch came from Mary Barra, CEO of General Motors, who teased a new “Super-Cruise” feature in the automaker’s 2017 Cadillac model. A month later entrepreneur Elon Musk announced “Autopilot” in his company’s upcoming “D series” Tesla S electric roadster.
These and other technologies riff on the idea of an independent, all-seeing, all-knowing, KITT-like car brain. Instead of thwarting crime, however, the goal is to thwart ourselves—machines are almost certainly safer drivers.
In practice, Super Cruise and Autopilot aren’t far removed from cruise control, since they will only make adjustments to a car’s speed and heading on the highway or other straightforward scenarios. The road to full autonomy is, as we recently detailed, a long and likely bumpy one.
However, something that stands to reduce the annoyance and danger of the daily commute, all while being more immediately deployable, cheaper, and robust, is within reach: cars that talk to one another. What’s more, GM is on track to be the first company to mass-manufacture and sell such a technology. If they win over consumers, governments, and competitors with the scheme, Google’s well-publicized innovations could become a footnote in automotive history.
Hivemind On Wheels
The industry jargon for giving cars a voice is vehicle-to-vehicle/vehicle-to-infrastructure communications, or V2V/V2I. Although it lacks the allure of Google’s robotic chauffeur, experts we spoke with believe V2V/V2I would lead to faster, greener, and safer driving for everyone—even those steering vehicles without the technology—and it could come much sooner than true universal automation. Equipped cars and trucks would constantly record and broadcast to one another data such as speeds, headings, weights, and GPS locations. Beacons installed along roadways would listen to the chatter, too, and pass that information along to traffic management systems.
GM is “impatient for the future to arrive”
Barra announced GM’s gamble on V2V/V2I in September, during the same speech at the Intelligent Transport Systems World Congress in which she teased Super Cruise. If the company follows through on incorporating the former technology in the 2017 Cadillac CTS, it could be the first V2V/V2I car to go on sale in the United States. Delphi Automotive has already announced a contract to build the systems.
GM is “impatient for the future to arrive,” Barra said, and the company should be. Tesla, BMW, Volvo, Mercedes, and others are all sinking resources into next-generation safety and navigation technologies, and each company is vying for widespread adoption of its proprietary setup.
Exactly what information GM’s system will broadcast isn’t known. Even less certain is how useful the tech will be on roadways at first, since a critical mass of capable cars and/or infrastructure are required to make use of the data. However, the nation’s largest car company just might have the foresight and wherewithal to disrupt how we drive.
The Cult Of The Self-Driving Car
Prototype Google Self-Driving Car
Google’s self-driving cars promise a sci-fi future. Just hop in the back and nap as your digital chauffeur takes the wheel. But a commercial rollout of fully automated vehicles could be decades away; the technological hurdles and cost are, for now, incredible. The hardware to support an autonomous, Google-like system, for instance, costs about $100,0000—roughly three to four times what a flesh-and-blood chauffeur can expect to make in a year in New York City. And that’s not including the cost of the car itself.
“No one wants to drive around in something that looks like a science project.”
Still, the vision of fully self-reliant cars is beguiling. GM made headlines everywhere with Super Cruise; the Los Angeles Times announced, “GM will introduce hands-free, foot-free driving in 2017 Cadillac.” Engadget wrote, “GM: A Cadillac that can (almost) drive itself is coming in 2016.” Bloomberg News put out a video with the same apologetic aside: “Cadillac Super Cruise: GM Bets on Self-Driving (Almost).”
Comparing these systems to true self-driving technology is a bit like comparing an airplane ride to a moonshot. Even GM representatives acknowledge there’s a major difference between the Super Cruise system and fully autonomous driving; i.e. it’s still up to the driver to keep the car moving in the right direction, to merge between lanes, and to get on and off at exits. Similarly, Tesla’s most significant advancement with Autopilot is guided merging. (When the driver taps the turn signal, the car will automatically shifts lanes of its own accord if and when it senses a safe opening.)
GM spokesman Jim Cain told Popular Science that Google’s full-coverage LiDAR system, which serves as their car’s robotic eyes, is perhaps the biggest stumbling block for mass adoption. No vehicle from GM, Tesla, or any other retailer features anything close to the sensor array necessary for true driverless machines. And no one has yet figured out how to adapt the technology for affordable mass production. Creative ideas to drive down the cost of autonomy-enabling hardware exist, but sticker prices remain shocking.
The aesthetics of the devices aren’t exactly pleasing, either. “No one wants to drive around in something that looks like a science project,” Cain says.
Letting cars communicate with one another could be much cheaper than letting them drive themselves, and solve many of the same problems.
If cars could talk, then they could help drivers and onboard safety systems prevent crashes. Imagine, for example, a distracted human at the helm of a 20,000-pound semi-truck barreling down a country road. He or she might not notice a two-door car backing out of a driveway a quarter of a mile away. But if the truck and the coupe could talk to each other, they could determine they’re on a collision course. Alarms might sound in the cockpit of the tractor-trailer, alerting the driver to the danger. Meanwhile, the coupe’s braking-assist feature might freeze the car in its tracks until the truck passes.
A whole host of possibilities emerge in places where connected infrastructure is in place. Getting on the road adds an anonymous data source for real-time traffic management. Together with thousands of other drivers on their way to work in the morning, your data stream could inform real-time changes in speed limits, lane use, and intersections. The city might share information with GPS guidance systems to help direct drivers to the paths of least resistance and manage loads. In yet another example, V2I systems in an ambulance could re-time traffic lights to clear a route to the hospital.
The above is still fiction, however, and no players have established a clear plan to make it happen. Getting from hoping to doing will take action and leadership. GM is betting it can provide both.
A Need To Lead
General Motors’ impatience for a connected future may spout from its desire to cut ties with the past.
The company’s gamble on V2V/V2I follows a painful period in their once-indomitable history. Scandal plagued the company in early 2014, costing GM what little positive press it had recouped after a $51 billion bailout in 2009: a defect in GM vehicles had endangered drivers for years, and the company knew about the problem as long ago as 2001. Public revelations prompted congressional hearings (and a comparison between the automaker and cannibal serial killer Jeffry Dahmer on The Daily Show). The ongoing recall is the largest in automotive history.
GM’s public stake on networked roadways puts its professional reputation on the line once again. Similar to cell phone and computer networks, the value of the technology depends on achieving a critical mass of users and infrastructure. Even mandating V2V/V2I in all of GM’s production models wouldn’t be enough; other car manufacturers and state and federal governments need to get on board. It’s unclear how many more failures GM can afford to add to its pile. In effect, Barra and GM risk joining a long list of oddball utopians who predicted transportation futures that never came to pass.
In A Perfect World
Cain believes the company’s risk will pay off. “We’re hopeful that our announcement will encourage other automakers to follow suit,” he says. “Someone had to go first. We’re happy to do it.”
Certain state and federal agencies seem to be cautiously signaling that GM is headed in the right direction. The National Highway Traffic Safety Administration (NHTSA), which oversees roadways for the federal government, issued something called an Advance Notice of Proposed Rulemaking on V2V technology in August. The motion means that, given time (experts say at least a few years), NHTSA could require automakers to include V2V technology in all production cars as a matter of public safety. They could even mandate retrofits of old cars with the technology, according to John Maddox, who helps lead an extensive test bed for intelligent vehicles at the University of Michigan.
“There will be hot spots and there will be dead spots. It will be critical that in those initial hot spots drivers see value from day one.”
At the same time, he cautions against too much optimism. A V2V mandate would put America ahead of the rest of the world on car-to-car communication, but infrastructure takes governmental resources and action of a kind that’s hard to imagine being implemented from the top down.
“We won’t see a national Big Bang of the kind we did last time with the federal interstate highways program under Eisenhower,” Cain said. “We won’t see that again given that we can’t even seem to finance the current system.”
When we drive down an interstate funded by the government, it’s easy to forget the feds don’t oversee the majority of the nation’s roads. That’s handled by state and local governments, which often have different priorities. While departments of transportation in Michigan, California, and Texas invest millions in preparing for a connected era in driving, other states have taken baby steps—or none at all.
Studies by the NHTSA indicate places that adopt the technology will see returns from increased efficiency and safety, and lower pollution as cars move in better tune with the roadways. (Something similar happened with the advent of automated tollbooths, which result in less idling time for drivers.) Yet as with all futuristic schemes, the exact costs are hard to predict.
Paul Lewis, an analyst with the nonpartisan Eno Center for Transportation, says for connected vehicles to make an impact, the government has to match the energy of the private sector and clarify issues such as liability for crashes connected to faults in the system.
Maddox says clearing all the roadblocks will be a matter of gathering momentum. Like cellular phones a few decades ago, he says, “there will be hot spots and there will be dead spots. It will be critical that in those initial hot spots drivers see value from day one.” He also notes that GM’s focus on V2V will have more trouble attaining a useful critical mass than V2I would if implemented because it relies on other cars—not infrastructure concentrated along highly trafficked routes.
For the most part, policymakers, experts, and GM reps share the same view: If V2V/V2I gets off the ground, it will be the biggest revolution in driving in the century-plus history of the automobile. We have the tools to do it now, and companies like GM might have the clout. The trick is getting everyone—including their cars—to listen.