- Innovators are developing the systems necessary for cars to drive themselves. Some, including Google, have already staged impressive demonstrations.
- However, synergies with alternative fuels appear modest, and the largest efficiency gains from self-driving cars are likely to be deferred until they dominate the market.
Although the entire concept of a self-driving car might seem science-fictional, it shouldn't greatly surprise anyone who has reflected on the implications of drone aircraft, GPS, smartphones, and the increasing electronification of average cars for the last several decades. From that perspective, the most important constraints on their emergence probably depend less on technology than on social and regulatory factors.
The development of self-driving cars and their precursors has been embraced by some of the biggest names in the global automotive industry, including GM, Toyota, Audi, BMW, Volvo, and Nissan, which announced plans to make the technology available across its entire product line sometime in the next decade. (Nissan also recently reported that its EV sales are lagging years behind plan.)
Suppliers to the OEMs are also making important contributions. I vividly recall driving a car equipped with radar adaptive-cruise control and other then-cutting-edge safety features in city traffic at the 2009 D.C. Auto show, courtesy of Robert Bosch, LLC. All I had to do was tap the gas pedal to engage the system and then steer, while the car did the rest. Systems like this are already appearing in production models.
The two main ways in which self-driving cars could affect future transportation energy usage involve making the operation of vehicles more efficient and enabling bigger changes in vehicle design than would otherwise be feasible. Some of these benefits would start to accrue from the day the first autonomous car left a dealership, but most would require either a critical mass of such cars in the fleet, or overwhelming dominance of the fleet. That could happen sooner in fast-growing developing countries, where legacy fleets are smaller, than in the developed world.
Consider operational changes first. Highway fuel economy could be improved by 20% by means of "drafting"--one car using the car ahead to reduce wind resistance--in automated , self-organized "platoons" of multiple cars. This, together with the avoidance of collisions, would also reduce traffic congestion, variously estimated at costing up to 2.9 billion gallons of fuel each year in the US, or up to 2% of US gasoline demand. The combined potential of these savings, assuming 100% market penetration of autonomous cars, might reach 10 billion gallons per year, a quantity larger than the gasoline displaced by corn ethanol in the US. Of course achieving such savings depends on having large numbers of self-driving cars on the road; imagine the risks if a daring driver in a conventional car attempted to join a platoon of tightly packed autonomous cars.
The efficiency gains from unattended autonomous parking don't require critical mass, and they might be significant, especially in congested urban areas, where one study suggested parking consumes up to 40% of gasoline used. However, most of these potential fuel savings could also be achieved through simpler and more easily implemented means, such as parking-space sensors and smartphone apps. And while self-driving cars might make car-sharing more popular, fewer vehicles wouldn't automatically translate into less fuel consumption if the same or more miles are driven.
The second major category of energy savings is associated with structural changes made possible by self-driving cars, mainly resulting in smaller and lighter vehicles. If cars no longer collided with each other or with inanimate objects, they wouldn't need to be nearly as robust. Saving weight saves lots of fuel. Yet it's hard to see how this process could begin before autonomous cars reached nearly 100% market penetration, since for many years they must share the road with millions of cars driven by fallible humans.
Nor is it obvious that self-driving cars would be infallible. We've already seen ordinary models exhibit random self-starting, due to malfunctioning of remote starter systems that would make up just one small subsystem of an extraordinarily complex self-driving architecture.
Some have suggested that the downsizing and weight savings facilitated by autonomous cars would hasten the adoption of battery-electric cars. The cost of today's EVs is driven largely by battery size, which is in turn a function of the vehicle's weight and its desired performance. A smaller, lighter car could make do with a smaller, cheaper battery pack. Cheaper EVs might well sell faster. However, if that must wait until enough self-driving cars are on the road for downsizing and radical lightening to become safe, it's a reasonable bet that improvements in battery technology in the intervening decades will have largely bypassed this potential benefit.
In the interim, while there might be some less-significant synergies between EVs and autonomous vehicles, neither technology is likely to depend on the other for its attraction to potential buyers. Nor do I see any obvious benefits from self-driving cars for helping alternative fuels like CNG, LNG or biofuels to gain market share.
On balance, if the average medium-term unique fuel savings of self-driving cars are limited to the 10-15% that I calculate--impressive but not game-changing--then the opportunities to improve safety and driver productivity seem like much more important motivators for this technology, for now. I also discovered a fair amount of skepticism about how soon fully autonomous cars would be widely acceptable to both consumers and regulators. Today's energy concerns might look quaint by the time such cars arrive in sufficient numbers to have a meaningful impact on them.
A different version of this posting was previously published on the website of Pacific Energy Development Corporation.