February 26, 2016

Researching, planning, and preparing for the automated vehicle revolution

Zia Wadud, Paul Leiby, and I have just published a new article on the potential impacts of widespread vehicle automation on travel demand, energy, and greenhouse gas emissions. Bottom line: we can plausibly see automation leading to anywhere from a factor of 2 decrease in road vehicle energy consumption to a factor of 2 increase. We need more dispassionate analysis and fewer rose-colored glasses if we are to properly plan and prepare for automated vehicles on our roads.

The opportunities for energy savings are manifold. As shown below, we could reduce energy consumption per mile by up to 20% through platooning (vehicles driving in tight formations to improve aerodynamics), and another 20% by automating eco-driving (driving habits that reduce in-use energy consumption). These are both significant near-term opportunities.

In the longer term, a shared, automated, on-demand mobility service provider could dispatch a vehicle that meets the needs of a specific trip: a one-person pod when traveling alone, a minivan for a whole family traveling together. Compared with today, where we generally own one or two vehicles that we use for all trips, this “right-sizing” of vehicles could cut energy consumption per mile by nearly half.

But, there are serious risks as well, primarily due to the potential for a lot more miles of vehicle travel. This is basic economics: you make something easier, cheaper, and more convenient, and people will use more of it. The value of the driver’s time is the single largest cost of operating a vehicle. Cut that cost, and you’ll see more travel – potentially up to 60% more, by our calculations. Don’t believe me? Read this customer’s testimonial on Tesla’s Autopilot, a Level 2 automation system:

“It was the same distance, but the commute felt like it took half the time”

Just imagine how travelers’ valuation of driving time will change once we get to Level 3 and Level 4 automation!

Let’s all remember that automation is a means to an end. The energy impacts come not from automation per se, but from the changes it causes in how, and how much, we use our vehicles. Maximizing the benefits, and mitigating the downsides, of vehicle automation will require smart public policies at the federal, state, and local levels. Federally, a mandate for low-latency vehicle-to-vehicle communication capability could be a key enabler for platooning, while “off-cycle” credits could provide auto manufacturers with an incentive to deploy eco-driving algorithms in automated vehicles, even though such algorithms would not save energy on EPA’s standardized test cycles. Locally, it may not be popular, but road pricing will likely become more important in an era of self-driving cars – and also easier to implement from a technical standpoint. Finally, some of the largest benefits of automation appear to lie in enabling the widespread use of shared, on-demand mobility services. Local governments should consider how they can integrate such services into their cities in a way that complements public transportation and lays the groundwork for future expansion.