Engineering news
That was the warning from researchers at the Massachusetts Institute of Technology (MIT), following a new study that explored the potential energy consumption and related carbon emissions that could arise if autonomous vehicles are widely adopted.
Data centres currently account for about 0.3% of global greenhouse gas emissions, according to the International Energy Agency – equivalent to Argentina’s annual emissions.
Realising that less attention had been paid to the potential footprint of autonomous vehicles, the MIT team built a statistical model to study the problem. They determined that one billion autonomous vehicles, each driving for one hour per day with a computer consuming 840W, would consume enough energy to generate about the same amount of emissions as data centres currently do.
The researchers also found that in over 90% of modelled scenarios, each vehicle would have to use less than 1,200W of power for computing to stop emissions outpacing those from data centres, a challenge that would require more efficient hardware.
In one scenario – where 95% of the global fleet of vehicles is autonomous in 2050, computational workloads double every three years, and the world continues to decarbonise at the current rate – they found that hardware efficiency would need to double faster than every 13 months to keep emissions under those levels.
“If we just keep the business-as-usual trends in decarbonisation 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,” said graduate student Soumya Sudhakar, first author of the research.
“But if we get ahead of it, we could design more efficient autonomous vehicles that have a smaller carbon footprint from the start.”
The team’s model is a function of 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.
To tackle uncertainties, the researchers modelled the workload of a popular algorithm for autonomous vehicles, known as a multitask deep neural network because it can perform many tasks at once. They explored how much energy the network would consume if it were simultaneously processing many high-resolution inputs from many cameras with high frame rates.
One way for carmakers to boost efficiency could be more specialised hardware, designed to run specific driving algorithms. Vehicles tend to have 10- to 20-year lifespans, so one challenge would be future-proofing the hardware to run new algorithms.
In the future, researchers could also make the algorithms more efficient, so they need less computing power. This would also be challenging, however, because trading off some accuracy for more efficiency could hamper vehicle safety.
The model only considered computing, and did not take manufacturing emissions or the energy used by vehicle sensors into account.
This work was funded in part by the US National Science Foundation and the MIT-Accenture Fellowship. It was published in IEEE Micro.
Want the best engineering stories delivered straight to your inbox? The Professional Engineering newsletter gives you vital updates on the most cutting-edge engineering and exciting new job opportunities. To sign up, click here.
Content published by Professional Engineering does not necessarily represent the views of the Institution of Mechanical Engineers.