Electric cars will likely have to sacrifice range if they’re also self-driving, but not so much that automakers will need to include gas engines as an inevitable part of autonomous models, according to a new study.
Published in Nature Energy this week, the study from Carnegie Mellon University engineers examined whether self-driving systems could be designed in ways that save battery life and preserve the range of all-electric platforms.
"We wanted to see whether autonomy will be an enabler or something that impedes the growth of electric vehicles," said Venkat Viswanathan, lead author and an associate professor at CMU.
The question may have implications for CO2 emissions, particularly if self-driving cars ever experience widespread adoption. Lawmakers on both sides of the aisle tend to see electrification and autonomy as naturally complementary. But Trump administration officials have made clear they do not, and at least one automaker, Ford, says it’s not yet ready to ditch gas engines as part of self-driving prototypes.
In an email to E&E News, Ford spokesperson Daniel Pierce said yesterday the company planned to launch its self-driving models as hybrids because pure EVs suffer faster battery degradation and up to a 50% reduction in range. The pandemic has pushed back Ford’s initial rollouts of self-driving models into 2022, he added.
Using physics-based modeling, however, the CMU team found that automakers could find workarounds for their self-driving platforms that would limit battery range losses to 5% to 10% in the suburbs and on highways. In cities those losses could be limited to 10%-15%, according to the study.
Yet those reductions could turn off many consumers, particularly if they pulled a car’s range beneath 200 miles, said Viswanathan.
But as batteries grow more powerful, a compromised range might seem less like a non-starter to consumers and more like opting for last year’s iPhone, according to researchers. High-level automation will have "the same effect on range as a 1-2 [year] time lag on improvements" in a battery’s specific energy, the study found.
The team also found that a typical self-driving platform would have about a 5% loss in battery longevity, although that was likely to become "negligible" in the future as battery technologies advance.
The entire problem of range stems from two areas: the hardware that protrudes from the vehicle and gathers information on its surroundings and the extra onboard software that processes information while sucking power from the battery.
Automakers and their component-making partners have "a lot of options" for hardware and software alternatives, said Viswanathan, such as more energy-efficient computing tools, custom hardware that doesn’t protrude as much from the vehicle or just coordinating with aerodynamic specialists.
One less preferred option: lidar, a laser-imaging technology that serves as the vehicle’s "eyes" for most automakers’ early AVs, but which Tesla CEO Elon Musk has dismissed as a "fool’s errand."
Typically, those systems are mounted on a car’s roof, cutting range by an additional 5% in many cases, the study found, suggesting developers should contemplate eliminating it.
"We want to be able to give a firm basis" for auto designers to build platforms that are electric-compatible, said Viswanathan.
"It’s providing a direction for many of these efforts to align with electrification. If you don’t do many of these things, then it won’t be compatible."