Traffic piles up on the Gardiner Expressway as commuters head home during the evening rush hour in Toronto, Ont. March 14/2011.Kevin Van Paassen/The Globe and Mail
It's not often that a driver will scramble to take responsibility for a car accident.
But after a minor fender-bender in San Francisco on Aug. 5, the owner of the energy-efficient Toyota Prius loudly announced that the driver was to blame and emphasized that the driver was a person.
The car was outfitted with an intricate display of video sensors mounted on its roof, and witnesses had quickly outed it as the property of search engine giant Google.
"We regret that a Google driver recently caused a minor accident," the company said in a statement. "One of our goals is to prevent fender-benders like this one, which occurred while a person was manually driving the car."
The car is part of Google's prototype fleet of driverless vehicles, technology designed with the idea of eradicating the human error responsible for so many highway accidents and traffic jams.
Earlier this year, the company announced that its "automated cars" had already driven from Silicon Valley to Santa Monica, Calif., over the Golden Gate bridge and navigated the winding lanes of the Pacific Coast Highway. According to the company, the cars have logged more than 140,000 miles, or more than 225,000 kilometres, an achievement Google called "a first in robotics research."
And while the company has placed safety at the forefront when explaining its desire to put a Google computer behind every steering wheel, it also is floating the idea that driverless cars could have a profound impact on commuting times.
The Google project is led by Stanford University robotics pioneer Sebastian Thrun, who theorized during a speech at the TED conference this year that the world's congestion problems could be solved by taking human drivers out of the equation.
"Do you realize we could change the capacity of highways by a factor of two or three if we didn't rely on human precision, and staying in the lane, but on robotic precision, and thereby drive a little bit closer together, in a little bit tighter lanes, and do away with all traffic jams," he said.
Mr. Thrun predicted that the United States could save four billion hours in wasted commuter time by freeing people from the burden of driving with driverless cars that would make more efficient use of road capacity.
The cars use video cameras, radar sensors and a laser range finder to "see" other traffic, while simultaneously referring to detailed maps that let them navigate the road ahead.
Tom Vanderbilt, author of Traffic: Why We Drive The Way We Do, believes that such computer-operated cars, which are also being developed by GM and a team of Chinese researchers, could have a substantial impact on congestion.
Because they rely on sensors, not sight, driverless cars don't need to see brake lights to understand that another vehicle is slowing down, he said, but would brake as soon as they receive the message that the car in front has decreased its speed. This ability would prevent the extreme, last-minute decisions that drivers make that often lead to accidents.
"We have this imperfect decision-making system," Mr. Vanderbilt said. "Each of us having that small moment of indecision amplifies into larger shock waves causing traffic. Just through pure performance, [driverless cars]can drive closer to one another, know exactly what's going on, know when to stop and how to stop."
Driverless cars could also theoretically create smoother traffic flow, he added, through more even speed distribution. Studies have found that the ideal highway speed is about 100 kilometres per hour, and that problems are caused when drivers go faster or slower.
Raj Rajkumar, co-director of the General Motors autonomous driving collaborative lab at Carnegie Mellon University, predicts that driverless vehicles will be on the roads in the next 10 years and drastically improve the use of road infrastructure.
Right now, drivers' manuals recommend maintaining a distance of about 30 metres from the car in front when driving at highway speeds. And while most drivers do not stay that far away, he said that cars run by computers could cut the distance to an average of less than 10 metres.
"You'd be using more of the infrastructure that exists today without having to build new roads," he said.
Already, most modern cars incorporate computer systems, he said. Adaptive cruise-control systems include radar under the car's hood that slows down automatically when it detects an obstacle ahead.
But seeing driverless cars on the road will require an evolution of attitude as well as technology.
People will have to wrap their minds around the social, legal, regulatory and liability issues associated with hands-free driving before fully commercial autonomous vehicles are allowed on the road.
But Mr. Rajkumar believes that for drivers, the shift will not be that difficult.
"When we fly, the autopilot is in control for much of the flight," he said. "We already trust our lives to computers in these safety-critical situations."