Technology NewsThe computers powering self-driving cars could be a huge...

The computers powering self-driving cars could be a huge driver of global carbon emissions

-



In the future, the energy needed to run the high-powered computers in the global fleet of autonomous vehicles could cause as many greenhouse gas emissions as all the data centers in the world today.

In the case of widespread adoption of autonomous vehicles, a rapid development of hardware efficiency will be required in order to emissions related to computer technology can be kept under control. This is one of the most important findings of a new study by MIT researchers, which examined the potential energy consumption and associated carbon dioxide emissions in the event of widespread adoption of autonomous vehicles.

Data centers, which house the physical computing infrastructure used to run applications, are widely known for their high carbon footprint. According to the International Energy Agency, they are currently responsible for about 0.3 percent of global greenhouse gas emissions, which is about as much carbon dioxide as Argentina produces annually. Recognizing that less attention has been paid to the potential footprint of autonomous vehicles, MIT researchers built a statistical model to study the problem. They found that 1 billion autonomous vehicles – each spending an hour a day in traffic with an 840-watt computer – would consume enough energy to generate roughly the same emissions as data centers currently generate.

The researchers also found that in more than 90 percent of the scenarios modeled, in order for autonomous vehicle emissions to not exceed current data center emissions, each vehicle would need to use less than 1.2 kilowatts of energy to compute, requiring more efficient hardware. need. In one scenario (in which 95 percent of the global vehicle fleet is autonomous by 2050, computing workloads double every three years, and the world continues to decarbonize at its current rate), they found that hardware efficiency would need to be at least every 1.1 should be doubled in order to keep emissions below the mentioned levels – the university’s news page reported.

“Just maintaining business-as-usual trends in decarbonization and the current rate of improvement in hardware efficiency doesn’t seem like enough to limit the on-board computing emissions of autonomous vehicles. This could become a huge problem. But if we prevent it, we can design more efficient autonomous vehicles that have less CO2 to begin with.” emissions,” said the study’s first author, Soumya Sudhakar.

Sudhakar wrote the study – which appears in the January-February issue of IEEE Micro – in the company of MIT researchers Vivienne Sze and Sertac Karaman.

The researchers built a framework to study operational emissions caused by computers with a global fleet of fully autonomous electric vehicles without a backup human driver. The model depends on the number of vehicles in the global fleet, the power of the computers on each vehicle, the number of hours driven by each vehicle, and the carbon intensity of the electricity that powers each computer.

“On its own, this seems like a deceptively simple equation. But all of these variables have a lot of uncertainty because we’re looking at an emerging application that doesn’t exist yet,” says Sudhakar.

According to some research, for example, the time spent with autonomous vehicles may increase because people can multitask while driving, and young and old people can drive more. Other research, however, suggests that driving time may be reduced because algorithms can find optimal routes that get people to their destinations faster. In addition to accounting for these uncertainties, the researchers also had to model advanced computing hardware and software that did not yet exist.

To this end, they modeled the workload of a popular algorithm for autonomous vehicles, called a multitasking deep neural network, because it can perform multiple tasks at once. They looked at how much power this deep neural network would consume if it were simultaneously processing high-resolution input from many cameras at high frame rates.

When testing different scenarios with the probabilistic model, Sudhakar was surprised how quickly the algorithm’s workload added up. For example, if an autonomous vehicle uses 10 deep neural networks that process images from 10 cameras and the vehicle drives for an hour a day, it will make 21.6 million conclusions per day. One billion vehicles would make 21.6 trillion (thousand trillion) conclusions. To put this in perspective, all of Facebook’s data centers around the world make a few trillion inferences per day (remember: 1 trillion = 1000 trillion).

According to Karaman, autonomous vehicles would also be used to move goods and people, allowing massive amounts of computing power to be distributed along global supply chains. And their model only considers computing, not energy used by vehicle sensors or emissions from manufacturing.

To keep emissions from spiraling out of control, the researchers determined that each autonomous vehicle would need to use less than 1.2 kilowatts of energy for computing. For this to be feasible, the efficiency of hardware must increase at a significantly faster rate, doubling approximately every 1.1 years.

One way to increase efficiency is to use more specialized hardware designed to run specialized driving algorithms. Since researchers are familiar with the navigation and perception tasks required for autonomous driving, Sudhakar says it would be easier to design specialized hardware for these tasks. However, vehicles typically have a lifespan of 10 or 20 years, so one of the challenges in developing specialized hardware would be to “future-proof” it to run new algorithms.

In the future, researchers could also make the algorithms more efficient, so they would need less computing power. However, this is also a challenge because if some accuracy is lost in exchange for greater efficiency, it can also jeopardize the safety of the vehicles.

Now that this framework has been introduced, researchers want to continue improving hardware efficiency and algorithms. In addition, according to their statement, their model can be further developed by characterizing the embodied carbon dioxide emissions of autonomous vehicles – the carbon dioxide emissions generated during the manufacture of the car – and the emissions related to the vehicle’s sensors.

While there are still many scenarios that need to be explored, the researchers hope that this work sheds light on a potential problem that people may not have considered. “We hope that people will think about emissions and carbon efficiency as important metrics to consider when designing. The energy consumption of an autonomous vehicle is really critical, not only for extending battery life, but also for sustainability” – said Wed.

Hardware, software, tests, interesting and colorful news from the world of IT by clicking here!

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Latest news

Google Pixel Watch receives February 2023 update featuring latest security patches

A new update has been pushed to Google Pixel Watch devices, giving users access to important security patches....

Is it a tablet or a laptop? This is the new Xiaomi Book that the firm is preparing

Xiaomi will present a renewal of its convertible and economic concept with Windows 11 very soon in China....

காலநிலை மாற்றம் ஒரு பெருங்கடலை “பேரழிவை” ஏற்படுத்தலாம்

மோசமான வெப்பமயமாதலின் கீழ், தெற்கு மெரிடியனல் ஓவர்டர்னிங் சர்குலேஷன் 2300 இல் முற்றிலும் நிறுத்தப்படலாம் என்று உருவகப்படுத்துதல்கள் தெரிவிக்கின்றன.வலுவான வெப்பமயமாதல் ஆழமான கவிழ்ப்பு சுழற்சியை...

The Last of Us 3 may already be in development against previous forecasts. Plans are to include even PlayStation 6

The series created by Naughty Dog has been success after success, recently going far beyond its native medium....

Is it a tablet or a laptop? This is the new Xiaomi Book that the firm is preparing

Xiaomi will present a renewal of its convertible and economic concept with Windows 11 very soon in China....

The Last of Us 3 may already be in development against previous forecasts. Plans are to include even PlayStation 6

The series created by Naughty Dog has been success after success, recently going far beyond its native medium....

Must read