The vehicle of the future: No drivers, no crashes
Sounds like the Jetsons — but 'the algorithm exists,' says a GM exec
Reprints
Print
Respond
Twenty years from now you will see electric cars that drive themselves, park themselves and communicate with one another. But the most fundamental advances will be in safety, predicts one of GM's foremost futurists.
"Probably the most fundamental thing I will end up having worked on in my career is the car that doesn't crash," says Larry Burns, GM vice president of research and development and strategic planning.
It's a strong prediction. But in Burns' view, if bees and birds can fly in swarms and people can avoid crashing into each other in crowded subway stations, "you know the algorithm exists."
• Adaptive cruise control
• Lane-keeping systems
• Wireless communication
• Cameras, lasers, radar
• Sensors
We're getting there
Essential to building a car that won't crash is building one that can drive itself.
General Motors demonstrated the potential of autonomous driving last year. A self-driving Chevrolet Tahoe SUV nicknamed Boss won a 55-mile race sponsored by the U.S. government's Defense Advanced Research Projects Agency.
It maneuvered the streets of Victorville, Calif., using a collection of cameras, sensors, radar and global positioning satellites. The vehicle had been modified by a team led by GM, Carnegie Mellon University, Caterpillar Inc. and Continental AG.
Burns says GM's technology roadmap sees self-driving cars by 2015 or 2020.
Some of the enabling technologies are already available or in development, such as adaptive cruise control that works at low speed as well as high speed, forward collision avoidance systems and lane-keeping assistance, which senses when a vehicle strays out of its lane and applies a small amount of force to correct steering.
"I think we can get to cars that park themselves," Burns adds, "and I'm not talking about camera parking systems. .... You pull up to that door, get out of your car, push your key fob and it goes and finds a spot in the parking deck all by itself. We call that virtual valet."
GM has demonstrated the virtual valet concept with Boss. Using sensors, Boss can judge whether a parking space is large enough and then maneuver itself into the space with no human intervention.
Chris Borroni-Bird, GM's director of advanced technology vehicle concepts, says four things have to come together to create a robust network of autonomous, noncrashing vehicles: sensing technology, dedicated traffic lanes, vehicle-to-vehicle communication and vehicle-to-infrastructure communication.
The vehicles themselves would have to be fully electric, he says, and that means more than just the powertrain. "Once we electrify the powertrain, we can begin to electrify other vehicle systems like brakes, suspension and steering and truly make an electric vehicle," he says.
Connectivity is the next step. "Where we think we can really make a quantum leap improvement in safety is when we begin to link the vehicles together wirelessly through sensor technology," Borroni-Bird adds.
Getting closer to reality
GM has been chasing this vision of autonomous driving for more than half a century.
Back in 1956, at one of the famous traveling auto shows that GM called Motoramas, the star was the Firebird II, with its radio-controlled "Auto-Way" controls. Auto-Way was supposed to be ready for the market in 1976.
As GM envisioned it, the driver would hit a switch, and an automated guidance system would whisk the vehicle along a metallic strip in the highway. It was a dream back then, but now it doesn't seem so far-fetched.
Carnegie Mellon and GM recently opened an autonomous driving lab on the university's campus.
"It's a concept that has been around for a long time, and we're starting to turn the corner where this is realistic," says Chris Urmson, an assistant research professor of robotics at Carnegie Mellon in Pittsburgh.
"It's really going to be a matter of will," Urmson says. "If there were sufficient funding I think we could push the technology to the point where it would be ready in 10 years. Whether we have the will to do that is another question."
A century of sameness
In Burns' view, the time is right for a radical change in automotive DNA.
Cars built today and cars built 100 years ago are not all that different, he asserts. By 1908, most automakers had adopted internal combustion engines and petroleum as the powertrain and fuel of choice.
"If you were an archeologist in the year 2100 and you dug up a car that was built in 1908 and you dug up a car that was built in 2008, you'd say they were the same," Burns says. "They both had four wheels. They both had an internal combustion engine. They both used petroleum. They both were controlled prevalently by hydraulics and mechanical controls."
Burns sees an overhaul of automotive DNA in the next decade. When an archeologist digs up that car, he says, he or she will find something completely different. If cars don't crash, they don't need crush zones. They don't need reinforced doors, and they may not need airbags.
"If cars don't crash, you're talking about a radical change in how cars get designed," Burns says. "Imagine how much mass you could take off."
Rather than designing a 4,000 pound vehicle for a 200-pound person, he says, you could use a vehicle that weighs 400 pounds, or even 200 pounds.
With its spinning laser scanner, sirens and other equipment, GM's self-driving Tahoe looks like a big robot from an old science fiction movie. But as the components are miniaturized, you won't be able to tell a vehicle is autonomous by simply looking at it, Urmson says.
Still, getting people to accept this leap in technology could prove GM's biggest challenge. "The biggest hurdle is going to be societal acceptance," Urmson notes. "Right now, it sounds too much like the Jetsons for people to be accepting."
You can reach Leslie J. Allen at lallen@crain.com. -- Follow Leslie on 
