Simple Machine: a mechanical device that changes the direction or magnitude of a force. In general, a simple machine can be defined as one of the simplest mechanisms that provide mechanical advantage (also called leverage). (Wikipedia definition for Simple Machine)
The reason autonomous, or driverless, cars are important is the same reason reducing greenhouse gas emissions from the transportation emissions is really hard – leverage (given the tomes already written about autonomous cars, I’m skipping the background tutorial). Electricity emissions reductions are RELATIVELY easier – there are a few large powerplants, a few large utilities, etc. Not so in the transportation section. In the US alone we have more than 250 million vehicles and more than 100,000 filling stations. And let’s not start on land use – with decisions being made by homeowners, and every jurisdiction from federal down to municipality influencing these decisions. Leverage is why fuel efficiency standards are so important, and why we spend so much time obsessing about transit routing.
Autonomous cars, if deployed en masse, under the right conditions, could provide significant leverage in reducing transportation emissions. This is what makes them exciting, in principle; significant opportunities follow. To be clear up front – none of these benefits are guaranteed to occur, they are simply potential benefits that could follow assuming a lot of details get worked out.
- Alternative fuel adoption: The big challenge with adoption of alternative fuels (whether we’re talking about biofuels, compressed natural gas, electricity, or hydrogen) has always been chickens and eggs: Consumers don’t want to buy something they can’t reliably refuel, filling stations don’t want to pay to install refueling if there are no customers. For those of us working on alternative fuels, a common mantra is: start with fleets. Fleet vehicles refuel at a central location. Plus they are driven many miles over their life. Alternative fuel vehicles often have high upfront costs, and low fuel costs. This is a problem for your typical consumer, but not for a fleet operator. Autonomous cars would essentially transfer maintenance of personal vehicles from individual car owners, to a central fleet service. That fleet operator would have a lot more motivation to adopt lower carbon fuels and vehicles (and could perhaps be more effectively motivated by policy and regulation). Leverage.
2. Smaller or shared cars: It is often assumed that transit is by definition preferable to driving with respect to emissions. It tends to be, but this advantage is not inherent. A fully occupied car (4 person carpool), actually competes quite well with most modes of transit in “average occupancy” mode (SEE ABOVE). Of course that is an unfair comparison – most car trips are pretty close to single occupancy, and a typical transit trip wins out even without full occupancy. But autonomous cars enable two scenarios that would encourage lower emissions for personal vehicles: the opportunity for shared trips, and the opportunity for much smaller, lighter, purpose-designed vehicles. Our current habit of leaving our cars unused 95% of the time and driving around in them at ¼ of their capacity is wasteful. But it would financially untenable for a company operating a fleet of cars. So in all likelihood, you’d see two things happen. Small 1-2 person vehicles would predominate (think Smart cars, but smaller) – they would cost less to operate, fuel efficient, and rival transit modes for GHG emissions. And larger vehicles would be reserved for picking up multiple passengers.
3. Operational efficiencies. Peloton Technology uses vehicle-to-vehicle communication to coordinate chained “pelotons” of over the road trucks. They claim aerodynamic benefits sufficient to reduce emissions by 10%.There is a whole subculture of “hypermilers”, who maximize gas mileage without the use of efficiency devices – only by altering driving techniques (e.g. slow acceleration, coasting, etc). Another term is “ecodriving”. According to Wikipedia, someone once got over 200 mpg in a Honda Insight. There is clearly huge potential for vehicles under autonomous control, coordinating with each other and programmed for fuel efficiency, to achieve dramatic reductions in fuel use, independent of other technological innovation.
4. Manufacturing emissions. Brendon Slotterback covers this topic nicely here, so I will only touch on it. Personal vehicles, I repeat, typically sit unused 95% of the time. If everyone used autonomous cars for all the trips now taken in personal vehicles, we would need 95% fewer vehicles. Now, most of the emissions from vehicles over their lifetime come from burning fuel. But still, this in not insignificant.
5. Motivation to drive less, or more? You can make the argument that autonomous cars could transform driving into an entirely variable cost (where you pay per mile), from mostly a fixed cost (i.e. most of the cost of driving is car ownership). That might motivate people to drive less, since every additional mile costs more. Some have made the induced demand argument – by making driving much cheaper and more convenient, people might drive more. Who knows the answer? Not me. A lot of it probably comes down to the pricing model, like a cell phone (unlimited minutes? pre-paid autonomous car cards?).
6. Reinforcing urban living? This is the big question. There are, of course, potential pluses. Imagine the space in the city that is devoted to parking that could be freed up for other uses if we needed 95% fewer cars, and the remaining 5-10% were in nearly constant motion (not parked). It could lower the cost of denser housing by eliminating the demand for large underground parking structures. It would eliminate some negatives of living in the core (parking is a hassle, traffic noise, congestion, etc). It could be transformative for urban air quality, etc. All that curbside parking could be turned into bike lanes, who knows? The possibilities are endless. Worth at least 5 more posts (but I’ll spare you).
On the other hand, if you don’t need to drive yourself on your long commute – maybe long commutes become more appealing. Also, what do autonomous cars do to the major transit investments we’re making? What does TOD mean anymore?
In closing, I think some of these issues will come down to careful policy design. People I talk to in the industry tell me to expect commercial autonomous vehicles entering the market in 3-5 years. Let’s be pessimistic and say 10 (ignoring the products already on the market). Makes sense to start thinking it through…
Nice article. I would take issue with the “95% of the time” comment. While it is true the average car is in use 60 – 90 minutes per day, unfortunately many cars are in use at the same time. (Hence congestion).
To be able to reduce car ownership 95% (assuming the same vehicle occupancies) would require shifting trips so they were spread uniformly throughout the 24 hour cycle. This is unreasonable for a variety of reasons: human circadian rhythms for one, a desire to have people at work at the same time for another.
Some gains could come from dynamic carpooling, but I think greater gains could come from reducing vehicle size so that most cars were skinny cars serving 1 or 2 people (see the GM Lean Machine for an early prototype). People would be a lot more comfortable in a “motorcycle” if it were safe, enclosed, and driven autonomously.
Keep in mind that CO2 emissions per mile matter less than per trip. Cars enable (and land-use patters orienting themselves around autos require) longer distance trips. This includes everything from the daily commute to the grocery run. The same land-use patterns (and poor transit service) also drive average occupancy of public transit to be much lower than it could be. (a quick search on Google didn’t give me data to support this theory, but I would wager occupancy rate by transit technology is higher in European cities than the US average). It also ignores the number of trips that require no transit if you live in a place where transit would support your daily needs (how many trips do you make by foot or bike).
The figure also ignores the emissions and loss of vegetation that has gone toward building the road and parking network to support our current vehicle fleet. Obviously autonomous cars could represent a huge downward shift in lane number and width over the next 30-40 years, slightly remedying this issue.
Another thought (from an environmental perspective) is that transit agencies should more aggressively adopt fuel source changes for its fleet of vehicles. They have the capability to manage the fuel source/filling location just as you describe.
I’m also very skeptical regarding regarding practical implementation of the benefits autonomous vehicles bring:
– Adoption rate. The benefits people cite regarding this type of vehicle (outside the driver freeing up personal time) can only be truly maximized when every vehicle on the road is part of a shared network to reduce congestion, minimize emissions, etc. Many techno-optimists love the idea of a self-driving car, but what about the vast amount of people who ‘prefer to drive’ (feel in control, etc)
– Sharing. How many people lease out unused bedrooms, yard space, the ability for 2 devices to stream Netflix at once, their 3rd vehicle used infrequently for big trips or loads, etc? How many people will be comfortable allowing their vehicle to become a personal taxi when they’re not using it, even if it does bring them personal revenue? (ie what if I want it suddenly but someone else is in it!?). How many people deride transit because you’re sharing space with strangers. Etc
Good catch David. There’s two important percentages – average capacity utilization (5-10%), and peak utilization. I may have glossed over that detail in the post.
What really surprised me is how low PEAK utilization is (e.g. rush hour). The authors of the Earth Institute study looked at 2009 NHTS data for Ann Arbor, MI. They found that peak utilization for an average weekday was a bit less than 20%. Meaning that even during rush hour, less than 20% of vehicles are in use. So if everyone in Ann Arbor was part of an autonomous car sharing system, you would need 80% less vehicles (let’s be conservative and say 70%).
My initial response upon starting to read the study was that they must be excluding rush hour, and assuming autonomous car use for other trips. But not so, they assume ALL trips.
I’m not sure Ann Arbor is a fair representation of the rest of the US in that realm.. As a major college town with a pretty high walk, bike, and transit mode share relative to other metros (according to Wikipedia the combined share is over 30%, compare to roughly 8.5% nation-wide).
Census data from 2009 shows that commute (defined as trips for work, what a slippery word, though: http://www.humantransit.org/2013/07/slippery-word-watch-commute.html ) says that 76.1% of trips were SOV, representing 105.4 million vehicles. It was estimated in 2009 that there were 254.2 passenger vehicles total in the US. This would lead me to believe that vehicle utilization across the US during rush hour was ~41%.
I don’t think it’s as easy as eliminating the remaining vehicles from the road network assuming an autonomous sharing system. That assumes land-use patterns could support a near-entire shift to shared trips (how many people would share a taxi with someone needing to be picked up 5 miles away), destinations are close enough to be reasonable, personal preferences would allow it (would they want to even if it were convenient), and a host of other factors.
I would guess that peak utilization periods would see a reduction in vehicles on the road, but by perhaps 10-20%?