By Emily Badger, October 25, 2013
Image taken from the back seat of a driverless car at Volkswagen Automotive Innovation Laboratory at Stanford University.
We've heard a lot about the projected benefits of driverless cars
(whenever we get them, that is). The roads will become safer, as we
remove distracted, flawed drivers (and human error) from behind the
wheel. Congestion will decrease, as cars that drive themselves and
communicate with each other are able to more efficiently share roadways.
Fuel economy will go up as a result, and emissions will go down. We'll
need to devote less space to parking, as automated vehicles come to
function more like public transit, remaining perpetually in motion. And
all kinds of people who can't currently drive – the young and old, as
well as the disabled – will become more mobile.
Which is all well and good. But exactly how many fewer crashes are we talking? How much money and gas and time would we save? Can we get some numbers, please?
The Eno Center for Transportation this week released a helpful paper
that corrals many of these estimates. The benefits of autonomous cars
will expand as there are more of them on the road. Ten driverless cars
on the 405 in Los Angeles, for instance, won't do much to improve
congestion for everyone around them. But if 10 percent of vehicles on
the highway were autonomous – with the capacity to communicate with each
other – that might start to change things.
Below, Eno has estimated several impacts of AVs in a graduated future
in the United States where 10 percent, then 50 percent, then 90 percent
the market shares this capability:
"Preparing a Nation for Autonomous Vehicles" by the Eno Center for Transportation
Currently, the United States has about 5.5 million vehicle crashes a
year, about 32,000 of them fatal. And transportation researchers believe
that about 93 percent of those crashes are caused by human error.
Remove that element from the roadway, and we might approach a future
where the fatality rates on roadways come to look a lot more like they
do with airplane and rail travel. (Random fact: the only crash that
Google has so far reported from its AV experiments occurred after a
human driver took over for the computer.)
The right column above is pretty far into the future. But with 10
percent penetration, we're still talking about 200,000 fewer crashes,
100 million gallons of gas saved, and $37 billion in economic savings.
Eno estimates that all drivers will experience congestion savings early
on, whether you're in an automated vehicle or not. Crash benefits, on
the other hand, will go primarily to the people riding in AVs. So those
numbers will rack up more over time. The report also projects that the
total number of vehicles on the road will decline, as the average number
of miles driven per vehicle goes up.
Eno acknowledge that its analysis is "inherently imprecise," and of
course a whole lot of obstacles – of the technical, legal and cultural
varieties – need to be overcome before we even get to a 1 percent market
share. But it's helpful to see some numbers in the meantime.
For the curious, these are the assumptions baked into the above data: