Self-driving cars are at a fascinating juncture right now. We know they’re coming soon. We know they’re going to change things. But we don’t know how they’re going to change things — in what directions, to what effect, how quickly — so there’s no end of breathless speculation.
It stands to reason that vehicle automation could save energy and reduce emissions in some ways. Cars will be able to chain together more aerodynamically, drive at more consistent speeds, and perhaps serve as shared vehicles in lieu of individual vehicle ownership.
But it also stands to reason that automation could increase energy use and emissions in some ways. If driving is easier and more pleasant, people will do it more. Automation will open up car travel to populations (the young, the elderly, the visually or otherwise impaired) who did not previously have access. Self-driving cars could increase the overall amount of vehicle miles traveled.
So how will these factors balance out? What effect will self-driving cars have, overall, on energy use and carbon emissions from transportation?
Confident predictions about these matters are folly. Nonetheless, we do have some sense of the factors involved, enough to construct scenarios and get a sense of the possibilities.
That’s what researchers Zia Wadud (University of Leeds), Don MacKenzie (University of Washington), and Paul Leiby (Oak Ridge National Laboratory) have attempted in a new study: “Help or hindrance? The travel, energy and carbon impacts of highly automated vehicles,” in the journal Transportation Research Part A.
The study uses the ASIF model to assess emissions. That’s this equation:
Emissions = Activity Level * Modal Share * Energy Intensity * Fuel Carbon Content
It also considers how emissions effects differ at different automation levels. (In the US, automation levels run 1 through 4, where 1 is driver-assist stuff like adaptive cruise control and 4 is full, driverless automation.)
Anyway, I won’t bore you with all the calculations. I’ll just list the factors and let you know how they added up. Here they are in a quick chart:
Let’s break them out.
Eight mechanisms by which self-driving cars could reduce overall energy and emissions
The first six of these reduce energy intensity (the I in ASIF), while the seventh reduces driving activity (A) and the eighth reduces fuel carbon content (F).
1) Congestion mitigation: Self-driving cars can improve traffic flow, reducing congestion.
2) Automated eco-driving: This has to do with driving practices like avoiding sharp acceleration and deceleration and traveling at a consistent speed.
3) Platooning: This refers to cars linking together closely into vehicle trains, to reduce aerodynamic drag.
4) De-emphasized performance: If humans aren’t driving, they won’t demand the hyper-performance of today’s cars and will settle for slower acceleration.
5) Improved crash avoidance: Self-driving cars presumably won’t hit each other as often, also reducing congestion (and, y’know, death).
6) “Right-sizing” of vehicles: If there are fewer crashes, cars can be smaller and lighter.
7) Changes in mobility service models: Self-driving cars could reduce car ownership and increase car-sharing, reducing overall driving.
8) Fuel mix changes: There are three ways that self-driving cars could make alternative fuel vehicles (“electric vehicles, hydrogen fuel cell vehicles, or compressed natural gas vehicles”) more competitive. First, they could drive themselves to fueling stations (even if, like hydrogen stations, they are few and far between); second, they could reduce range anxiety (for, say, electric vehicles) by refueling themselves frequently; third, shared cars would be driven more frequently, which could create demand for cars that are more capital-intensive but last longer and use less fuel (like, say, electric vehicles).
Four mechanisms by which self-driving cars could increase overall energy and emissions
The first two increase energy intensity; the second two increase driving activity and modal share (more people switching from bikes, walking, and public transport to cars).
1) Higher highway speeds: Self-driving cars are safer and thus can drive at higher speeds on the highway, using more energy per mile.
2) Increased feature content: Passengers in self-driving cars will be spending longer in their vehicles and have more free time, which could lead to demand for additional features and amenities, increasing vehicle weight.
3) Increased travel from reduced cost of driver’s time: Right now, driving involves a cost, in time, attention, and stress. Automation could reduce or eliminate that cost, leaving “drivers” free to do whatever they want. When the cost of a service declines, demand rises (this is known as the “rebound effect”).
4) Increased travel due to new user groups: As mentioned before, populations previously unable to drive will now have access to personal vehicles, leading to an increase in vehicle miles traveled.
How do these mechanisms balance out?
Obviously, how all these mechanisms and factors balance out will depend on a number of things, including choices and policies we make today. Here is the study’s first approximation of the effects (the blue bars are ranges):
As you can see, the big swing factor here is travel cost reduction — in other words, how cheap and easy driving gets. If that stays at the low end, then the effects of self-driving cars on energy use are almost certain to be a substantial net positive.
If it reaches the high end, a 60 percent boost in energy consumption for transportation, all the energy-saving benefits could be wiped out, for a net increase in energy and emissions.
Here’s the key twist. Remember earlier we mentioned levels of automation, 1 through 4, with 4 being full automation?
It turns out that the energy-saving effects of vehicle automation are almost all captured at levels 1 through 3. You don’t need full automation to do platooning, car-sharing, and the like. You mainly just need cars to be able to communicate with one another better.
The energy-increasing effects of automated vehicles, on the other hand, mostly kick in at level 4: full automation. To put it simply, when driving is fully automated, it becomes super, super easy — the cost in time and attention falls to zero — so people are likely to do it way, way more.
Maybe it’s best to delay full automation
This leads to somewhat surprising policy implications. It may be that the socially optimal outcome, at least for now, is partial, not full, automation. That way the energy and emissions benefits of smarter driving practices can be fully captured, without allowing drivers to tune entirely out — without making it too easy.
The authors run four scenarios, involving various degrees of automation and changes in driving practices.
Scenario A involves all vehicles rising to level 3 automation. Scenario B involves stalling out at level 2. Scenario C involves higher-than-expected efficiency impacts from automation.
In scenario D — “dystopian nightmare” — everything lines up just wrong:
Policymaker and industry’s eagerness leads to broad adoption of Level 4 automation, which totally redefines what it means to travel by car. Drivers totally disengage from driving responsibilities, and the perceived cost of the their time plummets. On the highways, vehicles travel safely at higher speeds, creating continued demand for big, powerful engines. Platooning is forestalled by a regulatory and liability quagmire, and policy inaction. In the cities, congestion relief from operational improvements is swamped by the sheer increase in traffic volume. Automated eco-driving fails to catch on, as drivers value shorter travel times over energy savings. Vehicle designs and ownership models are largely unchanged from today, as consumers buy for their peak requirements.
Here are the impacts of the four scenarios:
As you can see, only in scenario D does net energy use rise.
In scenario D, we rush to full automation without getting the rules and regulations right first — to encourage platooning, eco-driving, car-sharing, reasonable highway speeds, and all the rest. We end up with tons more cars on the road, traveling much farther, with little gain in efficiency.
The future of vehicle automation is up to us
The larger message of this study is simple: The effects of vehicle automation are in the hands of today’s decision makers.
With some foresight and smart policy, we can maximize the energy and emissions benefits of automation while steering clear of, or at least minimizing, the rebound effect.
Perhaps when we get farther down the road (ahem) — when more vehicles are electrified, when car sharing is more firmly established, when the benefits of automation have proven out — we can move to full automation without the risk of carbon blowback.