Autonomous cars are coming.  Exactly when is anyone's guess, but ten years seems like a safe bet for mass-market availability.

The way we use cars is going to change.  Those changes will be wide-ranging, from car culture to the economy to the global death rate -- but, one area that hasn't been investigated in as much detail: the impact on the environment.

Cars are one of the biggest contributors to CO2 production in the developed world, and any change in the way we use cars should be considered in that light.  Can self driving cars beat conventional public transit for efficiency?  How much of an improvement do they offer over regular car ownership?  Today, we'll be investigating the issue of the eco-friendliness of autonomous vehicles, based on a lot of research and a little speculation.

Efficient Driving Strategies

flockingbirds

However good a human driver might be, they do make mistakes, and those mistakes have an impact on fuel efficiency.

Humans get lost, get stuck in traffic, and struggle to find parking.  Autonomous vehicles would drive better, and save fuel doing it. Autonomous cars can "flock", drafting behind one another to reduce drag, and there's the possibility of automatic carpooling in a fleet scenario, which could significantly reduce the number of vehicles needed, and the total fuel consumption.

The drafting concept alone could save 30% of car fuel usage (according to Rocky Mountain Tech), and MIT suggests that by eliminating the need to find parking, more than 40% of fuel consumption could be eliminated in the cause of autonomous vehicles.  They could also avoid traffic jams in town, which would have a huge impact on travel efficiency.  According to the KPMG institute,

In an autonomous vehicle transportation system, vehicles will navigate far more efficiently than current human operators do. The inefficiency of human-driven vehicles leads to considerable congestion at high traffic volumes and frequent traffic jams. In its 2011 Urban Mobility Report, the Texas Transportation Institute estimated that traffic congestion costs Americans 4.8 billion hours of time, 1.9 billion gallons of wasted fuel (equivalent to two months’ operation of the Alaska Pipeline), and $101 billion in combined delay and fuel costs.

Ecological Car Design

The more self-driving vehicles that are on the road, the fewer accidents there will be.  Getting most drunk drivers (the cause of 26 out of the 91 average daily car fatalities in the US) to use autonomous vehicles would have a huge impact on driving safety.

Aside from the obviously benefit of people not dying, this has other unexpected perks.  Notably, if the risk of accident can be made low enough, car companies could build much lighter cars by discarding many traditional safety features (like air bags, seat belts, and a crumple zone). Car insurance costs would also go down.

If driving becomes as safe as flying (so safe that individual accidents make the news the same way that air disasters do), people may be much more comfortable driving in much lighter, less robust cars, which can (in turn) save a tremendous amount of fuel.

truckframe

Furthermore, since you may be able to just summon as many cars as you need, there's no real need for most of the cars in the fleet to be standard five-seater sedans.  If you're willing to build two-seater vehicles and minimize weight and drag, you can build gas prototype cars that get 285 MPG, almost an 80% decrease in fuel consumption compared to the best consumer cars now available.  According to a journalist for Autos.ca who spoke to the Volkswagen team that designed the prototype:

Despite its small size and light weight, Volkswagen says the 1-Litre-Car is as safe as a GT sports car registered for racing. With the aid of computer crash simulations, the car was designed with built-in crash tubes, pressure sensors for airbag control, and front crumple zones. Its aluminum fuel tank is located in a protected area behind the passengers.

Electric Car Adoption

Right now, electric cars are in the awkward position of being almost perfect -- but the 'almost' really matters.  Even with traditional vehicle engineering (and a lot of incredibly heavy lithium-ion batteries), modern electric cars get the equivalent of 120 MPG, and the electricity is produced in substantially cleaner ways.

In terms of overall fossil fuel usage, this is about a 50% savings compared to top of the line traditional automobiles, given the same build weight.

The Tesla Model S, an all-electric car, got the highest score ever awarded by Consumer Reports, a 99 out of 100.  Jake Fisher, the head of automotive testing for Consumer Reports, said that his only remaining concern was charging limitations.:

If it could recharge in any gas station in three minutes, this car would score about 110.

Sadly, electric cars take much longer to recharge than gas cars take to fill up, and have a limited range on each charge.  This leads to the unfortunate truth that electric cars are not currently a viable option for most Americans.

Even though they're totally adequate for 95% of the trips people ever take (running errands in town), they fail entirely on the last five percent of long-distance cross-country trips. Finding places to charge your electric car can be difficult, and downright impossible in some places.

Autonomous cars offer a few solutions to this problem.  Autonomous cars would be most useful in a fleet scenario, in which all of the cars are owned by a central agency, and can be dispatched as required by users.

In this case, rather than the user having to wait for the car to recharge, the user could just switch cars if they over-ran the range on a particular vehicle.  If the passenger part of the car and the robot/motor/battery part of the car were separate modules, these swaps could be done in a few seconds without the riders ever having to get out.

This has all the advantages of battery swaps, with the additional perk that the battery can come to you.  Furthermore, by using a diverse fleet that includes a small fraction of gas-powered vehicles, you could allow for trips to areas that wouldn't otherwise be accessible to self-driving cars, particularly in the early days when the charging infrastructure is less available.

Comparison to Public Transit

oldtrolley

The nice thing about all of these different potential savings is that they're mostly independent.  In other words, all of these benefits stack.

If you work it out, between the 80% savings given by designing light-weight microcars, the 30% savings from smart drafting, and the 50% savings you get from using electric vehicles, you wind up with a total ecological impact that's something like 14 times less than a modern 60 MPG gasoline car, a benefit that holds up even for most long-distance trips (assuming there's a fleet of cars you can swap with in each city along the way).  That's not counting numerous other small benefits like car pooling, spending less time lost and in traffic, and more efficient autonomous parking lots.

This blows away every other form of transit by a landslide.  Traditional buses are merely five times more efficient than cars when fully loaded with passengers (which is unusual).  Trains and planes may actually be worse than cars under real-world usage conditions.  A fleet of autonomous vehicles would probably be the most efficient option for moving people around, to say nothing of the convenience of being able to summon a robot to carry you around anywhere at any time.  Public transport takes work to use efficiently, and autonomous vehicles could blow that away.

How Much Driving?

This is all really good news for the environment.  Unfortunately, there is a major caveat, which is that self driving cars are probably going to have a dramatic impact on how much people are driving.  Young, old, and disabled are all going to be able to get around much more easily, and, while it's hard to suggest that this is a bad thing, it does also have a measurable environmental footprint.

The situation is worsened because savings in fuel efficiency is also savings in dollars, and it's not just the price of gas.  The cost of operating a traditional gas car works out to about sixty cents per mile including maintenance, the cost of the vehicle over its lifespan, and the price of fuel.  In contrast, because electric cars are mechanically simpler and don't have as many high-temperature moving parts, the real cost of operating electric cars like the Nissan Leaf are closer to 3.5 cents per mile, of which about three cents are fuel consumption.

Electric vehicles require less maintenance, and last longer.  There isn't much you can do about the remaining maintenance costs, but the fuel costs can be reduced to negligible levels compared to the cost of wear and tear on the vehicle.  At slightly over half a cent a mile, crossing the United States (three thousand miles) would cost about fifteen dollars per person.  Even accounting for a hefty profit margin, that's still absurdly cheap.

Those numbers hold up under experimentation, too.  One electric vehicle owner (of the ActiveE, a car considerably less efficient than the Leaf), kept track of his energy usage and maintenance costs for several years, and, at consumer electrical costs:

...it would have cost me about $3,000 to drive the Active E the past 20 months and 57,000 miles,

That kind of reduction in real cost is going to have a measurable impact on how much people drive.  People would be much more willing to accept long commutes if they're incredibly cheap and they can watch Netflix while  traveling.

Also, in terms of road trips, imagine "sleeper cars" -- autonomous vehicles, designed essentially to be a queen-sized bed on wheels. These would allow users to depart Friday evening, sleep the entire journey to their destination, enjoy a day practically anywhere in the country, and then sleep the trip back and be home, well rested, on Sunday morning -- all for the cost of a couple of dollars.  Clearly, that's going to have a profound impact on the number of people day tripping.

It's not clear how much of an impact this is going to have on total driving rates -- but there is some good news.  People already spend well over an hour a day driving.  Even if every single person in the US spent twelve hours a day in a car, they still wouldn't be able to make up the difference in efficiency between a self-driving electric car fleet and traditional gas vehicles.

Overall Efficiency

Looking at the numbers, it seems clear that autonomous cars will be an overwhelming force for environmental good.

Granted, many of these optimizations (most notably the fuel savings of removing safety systems from cars) depend upon human-driven cars being absent from the road.  Or, at the very least, rare.  An autonomous vehicle can protect you from your own potential mistakes very effectively, but it can do less to protect you from every other idiot on the road.

We're probably a few decades out from the point at which we'll begin to see the traditional, privately-owned gas-powered car really start to vanish from the streets.  So, for the near future, performance gains will be a little less dramatic.

Still, it seems clear that (in the long run) self-driving vehicles are going to go a long way towards making travel both cheaper and more convenient, and less burdensome on the biosphere and on our limited natural resources.

Image Credits: "'Driving' the frame of dad's truck", by JD Hancock, "In Flight", by Don McCullough, "Safety First" by Toronto History