“I’m still getting used to it,” says Cath Roberts as we sit in rush-hour traffic together. Three cars in front of hers is an empty electric vehicle emblazoned with Google logos. And it is driving very strangely.
“Google first rolled out their autonomous vehicles to support Google Maps back in 2005,” says Evan Johns, head of traffic management in San Francisco. “We trialled them extensively in 2013 to ensure they were safe and monitored them closely.”
Then they noticed something peculiar. Google was attempting to optimize the time the vehicles spent on the road and so the vehicles formed a mesh network. They would detail traffic build-ups and manage their speed to reduce energy consumption and time on the road. Yet that also had an effect on other drivers.
“Basically, where our vehicles were operating traffic tended to flow smoothly. We realized that we were actually breaking up traffic jams,” says Alison Rogers of Google Maps.
Traffic managers were excited about the potential and soon Google had been invited to run their robots during rush hour.
“It used to take me over an hour to get to work,” says Cath, “now it takes just 40 minutes. It’s awesome.”
ANALYSIS >> SYNTHESIS: How this scenario came to be
Rush-hour traffic is awful. Or, to put it another way: “Traffic congestion is a condition on road networks that occurs as use increases, and is characterized by slower speeds, longer trip times, and increased vehicular queuing.”
Mathematicians use the equations of turbulent flow to model the chaotic events that lead up to jams. “Phantom jams can form when there is a heavy volume of cars on the road. In that high density of traffic, small disturbances (a driver hitting the brake too hard, or getting too close to another car) can quickly become amplified into a full-blown, self-sustaining traffic jam… A team of MIT mathematicians has developed a model that describes how and under what conditions such jams form, which could help road designers minimize the odds of their formation. The researchers reported their findings 26 May 2009 in the online edition of Physical Review E. Key to the new study is the realization that the mathematics of such jams, which the researchers call ‘jamitons,’ are strikingly similar to the equations that describe detonation waves produced by explosions, says Aslan Kasimov, lecturer in MIT’s Department of Mathematics. That discovery enabled the team to solve traffic jam equations that were first theorized in the 1950s.”
These harmonics need to be broken up before they become entrenched. This leads to the theory of Active Traffic Management. In the UK, where it is at its most sophisticated, this includes computerized speed limit displays to reduce the speed limits as problems are identified. MIDAS sensor loops placed every 328 feet monitor traffic density. Variable displays indicate alerts to drivers. The hard shoulder can be used as an extra lane once traffic slows down.
Unfortunately, people don’t always take note of the instructions and so the system doesn’t work as well as it should. Something is needed to manually slow down traffic.
2005 – 2007: DARPA Grand Challenge
In 2005, the US Defense Advanced Research Projects Agency runs the second of its autonomous vehicle challenges. The one held in 2004 ended without a single vehicle completing the 150 mile course in the Mojave Desert. Carnegie Mellon University’s entry travels the furthest; 7.36 miles.
On 8 October, 23 contestants set out along the same 2004 course. Five vehicles complete the race including both Stanford University and Carnegie Mellon. Vehicles in the 2005 race pass through three narrow tunnels and negotiate more than 100 sharp left and right turns. The race concludes through Beer Bottle Pass, a winding mountain pass with a sheer drop-off on one side and a rock face on the other.
In the same year Google begins testing a small fleet of driverless vehicles in order to automate their Streetview service. Google’s vehicles are only partially automated since the routes themselves are carefully plotted and planned in advance.
In 2007, DARPA runs a third competition: the Urban Challenge. This takes place on 3 November at the closed George Air Force Base and involves a 60 mile urban course to be completed in six hours. Rules include obeying all traffic regulations while negotiating traffic, obstacles and merging into traffic. Carnegie come first, swapping results with old rivals Stanford who come second.
All in all, six teams finish the race.
2010: From Italy to China, to Google
On 26 July 2010 four vehicles set out from Parma in Italy, to drive the 8,100 miles to Shanghai, China. The VisLab Intercontinental Autonomous Challenge (VIAC) mission is designed to test existing technology in conditions that will be more extreme than most autonomous vehicle will be expected to operate in.
The vehicles are entirely driverless and their objective is to make it to the Shanghai World Expo. “We were preparing our demo, testing in a pedestrian area with our driverless vehicles,” one of the VisLab team members blogs in September. “A local policeman, who apparently was not aware of our presence, saw our vehicles moving in a restricted area and approached them to talk to the driver and stop them. Then he realized there was no driver! And the vehicles were speeding between obstacles. He then looked around and tried to find a clue of what was happening. He really seemed puzzled about that.”
They reach Shanghai after 92 days.
On 10 September, Sebastian Thrun, a Google Distinguished Software Engineer, blogs: “So we have developed technology for cars that can drive themselves. Our automated cars, manned by trained operators, just drove from our Mountain View campus to our Santa Monica office and on to Hollywood Boulevard. They’ve driven down Lombard Street, crossed the Golden Gate bridge, navigated the Pacific Coast Highway, and even made it all the way around Lake Tahoe. All in all, our self-driving cars have logged over 140,000 miles. We think this is a first in robotics research.”
“Our automated cars use video cameras, radar sensors and a laser range finder to “see” other traffic, as well as detailed maps (which we collect using manually driven vehicles) to navigate the road ahead. This is all made possible by Google’s data centers, which can process the enormous amounts of information gathered by our cars when mapping their terrain.”
Google starts actively lobbying for their vehicles to be granted rights to move about unimpeded.
2013: When ‘cruise-control’ means something
Audi and Volkswagen offer a feature pack with their new Golfs and TTs which permits vehicles to take over from the driver on regular journeys. In addition, sensors pick up if the driver is fatigued and ‘correct’ their driving to keep the vehicle at a safe following distance and on its route. It costs US$ 10,000 but prices are expected to drop as demand takes off.
“We’re very excited about this service. According to the World Health Organization, more than 1.2 million lives are lost every year owing to traffic accidents. We believe that our system will save lives,” says Jasper Luffs of Volkswagen.
By the end of the year, drivers are becoming less startled by seeing driverless vehicles with the passengers reading papers or dozing as Mercedes, BMW and Toyota all offer autonomous vehicles.
Google is offering its vehicles for extensive testing in major cities around the world. During tests in San Francisco traffic managers notice that, while Google vehicles are on the road, traffic management appears significantly easier. They call in top traffic analysts and partner with MIT to study whether this is an anomaly or can be replicated.
2015: Rush-40-minutes has less ring to it
“We’re satisfied that the system works and will be launching at the end of the month,” says New York’s mayor, Hilary Clinton, as she unveils new-look Google autonomous vehicles. They are to start work managing traffic into the city in the mornings and out again in the afternoons.
New York is only the fifth city to introduce the service.
“The trick is to deliberately slow traffic down and break up those ‘jamitons’ before they become entrenched,” says Aslan Kasimov. “I’m delighted with the system.”
At the end of the first day, Cath Roberts is delighted to discover that her usual one-hour journey took only 40 minutes.