According to researchers at Carnegie Mellon University's College of Engineering, actually changing the way the EPA treats autonomous cars could have an impact on efficiency. The report, Fuel economy testing of autonomous vehicles, published in the journal Transportation Research Part C: Emerging Technologies, finds that if the automakers take efficiency into account of their autonomous driving algorithms, vehicles could yield efficiency benefits of up to 10 percent.
Autonomous driving algorithms could yield efficiency benefits of up to 10 percent.
As vehicles become smarter and more connected, they can predict and even communicate behavior to one another. This impacts how smoothly a car accelerates or decelerates, which has a big influence on fuel use. And the more driving duty the technology takes over in coming years, the bigger impact that programming will have. If the EPA takes early autonomous vehicle technology into account for its fuel economy and emissions testing, its influence could help convince automakers to unlock the fuel efficiency afforded by smart use of autonomous driving programming.
On the other hand, if the EPA does nothing to account for vehicle autonomy, and automakers don't take it upon themselves to incorporate fuel-saving methods into their self-driving algorithms, fuel economy could degrade as much as three percent. "Because existing standardized tests don't consider AV technologies, there are limited incentives for car manufacturers to design cars for optimum fuel efficiency," says assistant professor of civil & environmental engineering Constantine Samaras.
Or fuel economy could degrade as much as three percent.
The researchers recommend incorporating vehicle autonomy into driving cycles when performing tests, and they offer their own test results as a "starting point" for revamping EPA tests. While any given driver will use autonomous features differently, the fact is that they are becoming a part of real-world driving, and tests should try to reflect the fact that human drivers are and will be handing over more and more driving duties to autonomous vehicles. And if automakers know they can improve their EPA ratings through optimized software, that's bound to help steer autonomy in a cleaner direction. Learn more in the press release below.
PITTSBURGH, March 22, 2016 /PRNewswire/ -- Researchers in the College of Engineering at Carnegie Mellon University discovered that fuel efficiency for self-driving cars could improve by up to 10% if fuel economy and emissions standards tests were revamped to account for the early autonomous vehicle (AV) technologies likely to be offered in the coming years. Further, without changes to the current tests to include AV technologies, the research found fuel economy could degrade by 3%, depending on how manufacturers program AVs.
Research conducted by Assistant Professor of Civil & Environmental Engineering Constantine Samaras and Ph.D. student Avi Chaim Mersky reveals that current Environmental Protection Agency (EPA) standardized tests do not take into account how current AV technologies, like adaptive cruise control, and future AV technologies will impact vehicle fuel use. As AV technologies are gradually adopted, more aspects of driving will shift from human drivers to the car. Connected self-driving cars will have the capacity to predict how other cars will behave on the road. These predictive capabilities will govern how smoothly self-driving cars accelerate and decelerate when following other cars, which significantly affects fuel use.
"Because existing standardized tests don't consider AV technologies, there are limited incentives for car manufacturers to design cars for optimum fuel efficiency. The EPA can use our research as a starting point in redesigning fuel economy testing for autonomous vehicles," says Samaras.
In the study, researchers developed simulations to incorporate AV technologies within the bounds of current fuel economy testing, and they simulated a range of AV testing scenarios to estimate changes in fuel economy. They examined how driverless cars will perform when sharing the road with primarily conventional vehicles, and they looked at connected vehicle scenarios in which information about a lead car's travel behavior was communicated to an AV following this lead car. The study found that more advanced connectivity could enhance a vehicle's performance by providing the vehicle with more time to plan future actions. The longer the vehicle plans into the future, the greater the fuel economy benefits.
"What we have quantified is that fuel economy testing will need to account for AV technologies in the not-so-distant future," says Mersky. To start these discussions, the study provided suggestions on how current EPA fuel economy tests could be modified to address AV technologies.
The study, "Fuel economy testing of autonomous vehicles," was published in Transportation Research Part C: Emerging Technologies.