A global fleet of about a billion autonomous vehicles, each driving just for one hour per day, may generate about the same amount of carbon emissions as data centres currently do, according to a new study.
The research, published recently in IEEE Micro, modelled the potential energy consumption and related carbon emissions if autonomous vehicles are widely adopted.
Data centres across the globe that house the physical computing infrastructure used for running applications currently account for about 0.3 per cent of global greenhouse gas emissions – about as much carbon as the country of Argentina produces annually – scientists, including those from Massachusetts Institute of Technology in the US, said.
In the new study, they found that to keep autonomous vehicle emissions from zooming past current data centre emissions, each vehicle must use less than 1.2 kilowatts of power for computing in over 90 per cent of modeled scenarios.
Researchers call for more efficient hardware to cut emissions from computer operations in self-driving cars.
Under one of the modelled scenarios where 95 per cent of the global fleet of vehicles is autonomous in 2050, scientists say the hardware efficiency would need to double faster than every 1.1 years to keep emissions under those levels.
“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. This has the potential to become an enormous problem,” study co-author Soumya Sudhakar said in a statement.
Scientists say there is a need to design more efficient autonomous vehicles that have a smaller carbon footprint from the start to overcome the hurdle.
In the research, they developed a framework to explore the emissions from the operation of computers on board a global fleet of electric vehicles that are fully autonomous.
The emissions were calculated based on a number of factors including the number of vehicles in the global fleet, the power of each computer on each vehicle, the hours driven by each vehicle, and the carbon intensity of the electricity powering each computer.
Researchers also modelled the emissions based on the advanced computing hardware and software that will be used in such vehicles.
Citing an example, they say if an autonomous vehicle has 10 deep neural network AI that processes images from 10 cameras, and that vehicle drives for one hour a day, it would make 21.6 million inferences each day.
For a billion vehicles, they say this would make 21.6 quadrillion inferences.
In comparison, all of Facebook’s data centres worldwide make a few trillion inferences each day (1 quadrillion is 1,000 trillion), researchers explained.
“After seeing the results, this makes a lot of sense, but it is not something that is on a lot of people’s radar. These vehicles could actually be using a ton of computer power,” Sertac Karaman, another author of the study, said.
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,” Dr Karaman added.
The numbers calculated in the study are only considering the computing operations in self-driving cars, and do not take into account the energy consumed by the vehicle sensors or the emissions generated during manufacturing, scientists say.
They say each autonomous vehicle needs to consume less than 1.2 kilowatts of energy for computing to prevent emissions from spiraling out of control.
To achieve this, researchers say the computing hardware must become more efficient at a significantly faster pace – doubling in efficiency about every 1.1 years.
“We are hoping that people will think of emissions and carbon efficiency as important metrics to consider in their designs. The energy consumption of an autonomous vehicle is really critical, not just for extending the battery life, but also for sustainability,” Vivienne Sze, another study co-author, added.