Comment The amount of energy required in the future to run just the computers on a global fleet of autonomous electric vehicles could result in the emission of as many greenhouse gases as all of the data centers in the world right now.
Based on a statistical model that calculated the energy outputs that a fleet of one billion autonomous electric vehicles would generate if they ran for one hour per day, researchers from MIT announced this finding on Friday. The number that came out was roughly equivalent to 0.3 percent of the world’s emissions of greenhouse gases.
In order for self-driving cars to safely navigate traffic, they need a lot of computing power to run sophisticated algorithms and onboard camera systems.
The weirdest thing about driving a self-driving car Despite the fact that the findings are only projections, lead researcher Soumya Sudhakar of MIT stated that the findings should make self-driving car researchers and manufacturers realize that “business-as-usual” is not sufficient, and that computing efficiency should be at the forefront of their considerations.
She stated in a statement, This has the potential to become an enormous problem.
However, if we get ahead of it, we could design autonomous vehicles that are more energy efficient and have a lower initial carbon footprint.
From Tesla to GE, major automakers have staked their fortunes on the future’s self-driving vehicles. However, they have experienced numerous delays, technological difficulties, and safety concerns. Market statistics nonetheless indicate that the sector was roughly valued at $22 billion in 2022 and could reach nearly $76 billion by 2027.
Sudhakar said that even though that equation seems “deceptively simple,” it’s hard to figure out because it’s still unclear how self-driving cars would change how people drive.
For instance, some studies claim that because people can multitask, self-driving electric vehicles would make driving longer.
Others assert that driving time would decrease as a result of computers discovering the most efficient routes to destinations.
In addition, the researchers discovered that in order to prevent computer-generated emissions from spiraling out of control in the upcoming decades, each autonomous vehicle would need to consume less than 1.2 kilowatts of energy for computing.
For a global fleet to not exceed these emissions estimates by 2050, scientists also found that, in some scenarios, self-driving companies would need to make their computer hardware roughly double in efficiency every year.
According to the researchers, specialized hardware that is made to run specific driving algorithms and power navigation and perception tasks could be developed by scientists to improve efficiency. However, Sudhakar stated that this comes with a challenge due to the fact that vehicles typically have life spans ranging from 10 to 20 years. As a result, any hardware upgrades that are currently in place might not be “future-proof” enough to run brand-new algorithms.
Automakers are rushing to produce self-driving vehicles. Why then?
According to the authors of the study, researchers could also try to make algorithms more efficient and use less computing power. However, they might have to give up accuracy for efficiency, which could have an impact on vehicle safety.
A study author and associate professor in the Department of Electrical Engineering and Computer Science at MIT, Vivienne Sze expressed the hope that automakers will consider the technology’s carbon efficiency and output of emissions.
Public by world news spot live