MIT researchers investigated the potential energy consumption and carbon emissions if autonomous vehicles became widely used. They calculated that 1 billion autonomous vehicles, each driving for one hour per day and powered by an 840-watt computer, would consume enough energy to produce roughly the same amount of emissions as data centers currently do. Data centers currently account for about 0.3 percent of global greenhouse gas emissions, or roughly the same amount of carbon as Argentina produces each year. The study was published today in the January-February issue of IEEE Micro.
The model is determined by the number of vehicles in the global fleet, the computing power of each computer, and the number of hours driven by each vehicle. To put that in context, each day, all of Facebook’s data centres make a few trillion inferences (1 quadrillion is 1,000 trillion). Some research suggests that the amount of time spent driving in self-driving vehicles may increase because people can multitask while driving and the young and elderly may drive more. He claims that because autonomous vehicles would be used to transport both goods and people, a massive amount of computing power could be distributed along global supply chains. Their model only considers computing; it ignores the energy used by vehicle sensors or the carbon emissions produced during manufacturing.
To keep emissions under control, the researchers discovered that each autonomous vehicle must consume less than 1.2 kilowatts of computing power. To achieve this, computing hardware must become significantly more efficient at a much faster rate, doubling in efficiency every 1.1 years. One way to improve efficiency is to use more specialised hardware designed to run specific driving algorithms.
After demonstrating this framework, the researchers intend to continue investigating hardware efficiency and algorithm improvements. Furthermore, they claim that characterising embodied carbon from autonomous vehicles — the carbon emissions generated when a car is manufactured — and emissions from a vehicle’s sensors can improve their model.
“We hope that people will consider emissions and carbon efficiency as important metrics in their designs. (…) An autonomous vehicle’s energy consumption is critical not only for extending battery life, but also for sustainability “explains Vivienne Sze, associate professor in the Department of Electrical Engineering and Computer Science (EECS)