The environmental impact is potentially more damaging than imagined.
Self-driving cars have been touted by many in the industry as the future, but a study conducted by researchers at the Massachusetts Institute of Technology (MIT) found that this future could be a massive driver of global carbon emissions.
Quite simply, the energy required to power the advanced computers for a global fleet of autonomous vehicles would emit the same amount of greenhouse gas emissions as all the world's existing data centers.
The fact that rising emissions from these technologies would coincide with the rise of electric vehicles, the very cars designed to be kinder to the environment, feels like one step forward and two steps back.
Cruise
Data centers housing physical computing infrastructure used to run applications typically have a large carbon footprint, currently emitting around 0.3% of global greenhouse gas emissions. That's roughly equivalent to the annual carbon emissions of a country such as Argentina, as per the International Energy Agency.
MIT researchers decided to build a statistical model to analyze the possible carbon emissions of autonomous cars since little attention has been paid to it. These findings show that a billion autonomous cars would consume enough energy and thereby generate similar emissions as the world's data centers. That is based on each vehicle driving for one hour per day and with a computer using 840 watts.
“If we just keep the business-as-usual trends in decarbonization and the current rate of hardware efficiency improvements, it doesn't seem like it is going to be enough to constrain the emissions from computing onboard autonomous vehicles,” said first author Soumya Sudhakar, a graduate student in aeronautics and astronautics.
GM
The research team admitted that there are many uncertain variables since autonomous vehicles are still emerging. In America, for instance, GM's Cruise has a small fleet of self-driving robotaxis using modified Chevrolet Bolt EV hatchbacks, but that won't come close to the number of autonomous vehicles on the road in a future scenario where consumers can actually buy and own them.
The MIT team further said that in a scenario where autonomous vehicles become more widespread, the time that people spend in cars could actually increase. This is because people can multitask if a car is self-driving, and both younger and older individuals could travel more since they won't need to drive themselves. The Hyundai Seven concept, which will inspire the production Ioniq 7, is an example of this; designed for autonomous mobility, it has a retractable driver control stick, a flat floor, and a rear bench that looks more like a lounge.
However, alternative research suggests that driving times could drop since algorithms could guide self-driving cars to their destinations quickly. It's simply too early to be certain of how commuters' habits will change.
Stephen Brashear for Cruise
All that being said, the MIT researchers have exposed a potentially concerning aspect of a widespread autonomous fleet that hasn't been considered up until now.
“These vehicles could actually be using a ton of computer power,” said Sertac Karaman, associate professor of aeronautics and astronautics. “They have a 360-degree view of the world, so while we have two eyes, they may have 20 eyes, looking all over the place and trying to understand all the things that are happening at the same time.”
An example was given where one autonomous vehicle has 10 deep neural networks. These process images from 10 cameras, and over one hour of commuting, that's 21.6 million inferences in a day. That rises to 21.6 quadrillion inferences (1 quadrillion is 1,000 trillion if you can wrap your head around such numbers) for one billion cars. Consider that Facebook's global data centers make a few trillion inferences each day, and that puts into perspective the emissions price that we may have to pay for a large autonomous fleet.
Baidu
Making matters worse is that the model used by MIT researchers didn't even consider the energy used by vehicle sensors or emissions that come from the production process.
So, what is the solution, and is there even a viable one at all? According to researchers, a single autonomous vehicle should consume below 1.2 kilowatts of energy for its computing needs. That will only be possible if the efficiency of computing hardware increases much faster, with a doubling in efficiency every 1.1 years or so. That could come about by incorporating specialized hardware, a challenge on its own considering the lifespan of the average vehicle.
With consumer readiness for autonomous vehicles still very low, the findings by these MIT researchers will hardly bolster confidence in self-driving cars. Much in the way that the widespread adoption of EVs has created unexpected problems like lithium shortages and transportation challenges, this is unlikely to be the last stumbling block for self-driving cars.
Hopefully, it's one that industry experts can work together to circumvent.
Hyundai CarBuzz
Source Credits: MIT News
Keyword: Widespread Adoption Of Autonomous Cars Could Lead To Major Emissions Issue