A study by the Massachusetts Institute of Technology (MIT) has found that an en-masse switch to autonomous cars could actually drive carbon emissions up, not down.
Car and technology companies have touted autonomous driving as a panacea to several ills, from road traffic accidents to commuting to emissions. In theory, computers can automatically drive a car in the most efficient manner possible, including “platooning” cars close together on motorways which helps to save fuel or electricity consumption through better aerodynamics.
However, MIT has found that there is a series of unintended consequences from a potential switch to autonomous motoring. One such is the need for more energy to handle the computing power necessary to help run autonomous vehicle systems. According to MIT, data centres already account for about 0.3 per cent of global carbon dioxide emissions, which doesn’t sound like a lot, but it’s about the same amount as Argentina produces.
The MIT study found that: “One billion autonomous vehicles, each driving for one hour per day with a computer consuming 840 watts, would consume enough energy to generate about the same amount of emissions as data centres currently do.”
While it will take quite some time for a billion autonomous cars to be put on the road – if it ever happens – the ones that do make it to our streets will certainly be driving for more than an hour per day. Indeed, it’s an essential part of autonomous vehicle design – do one journey, dropping a customer off at the kerb, and then head off immediately on the next journey.
MIT’s research backs up claims made by Irish academic Brian Caulfield that chasing autonomous tech is actually the wrong thing to be doing, and that the money and effort involved would be far better directed elsewhere
MIT also modelled scenarios that showed the computing hardware fitted to autonomous vehicles would have to become more efficient at breakneck speed if it’s to keep ahead of the need to reduce carbon emissions. “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. But if we get ahead of it, we could design more efficient autonomous vehicles that have a smaller carbon footprint from the start,” says first author Soumya Sudhakar, a graduate student in aeronautics and astronautics and one of the authors of the MIT study.
There’s a difficult set of equations for autonomous cars to juggle if they’re actually to be more efficient than human-driven vehicles. According to MIT: “Some research suggests that the amount of time driven in autonomous vehicles might increase because people can multitask while driving, and the young and the elderly could drive more. But other research suggests that time spent driving might decrease because algorithms could find optimal routes that get people to their destinations faster.”
[ How will we get around in 2030?Opens in new window ]
Part of the problem will be the sheer workload for an autonomous car’s on-board computer, which will have to monitor and calculate its relationship with the real world via a gaggle of cameras, Lidar sensors and radar every second that it’s on the road. The task is so complex that MIT found that if an autonomous vehicle has 10 deep neural networks processing images from 10 cameras, and that vehicle drives for one hour a day, it will make 21.6 million inferences each day. One billion vehicles would make 21.6 quadrillion inferences. To put that into perspective, all of Facebook’s data centres worldwide make a few trillion inferences each day (one quadrillion is 1,000 trillion).
“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. 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,” says Karaman.
The idea of just sitting into a couch which effectively drives itself around? I don’t think that’s a great idea and I do think people would end up leaving public transport and opting for one of these pods
— Brian Caulfield of the Centre for Transport Research, Trinity College Dublin
MIT’s research backs up claims made by Irish academic Brian Caulfield that chasing autonomous tech is actually the wrong thing to be doing, and that the money and effort involved would be far better directed elsewhere.
Dr Caulfield is an associate professor, and head of department at the Centre for Transport Research, part of Trinity College Dublin. “Companies such as Volkswagen and Tesla are spending huge amounts of money, developing technology such as this and all it’s basically doing is making it easier to drive a car,” Dr Caulfield told The Irish Times.
“From a carbon perspective, from a congestion perspective, these supposed long-haul routes across Europe and the US, it really would be much more efficient if they were done by rail, especially overnight rail. There is an Australian professor who recently said that autonomous cars are a solution in search of a problem. I think that it might seem like a cool new way of doing things, but that overnight rail or an overnight bus would be much more efficient from a carbon perspective. You’re moving more people.”
[ Mission impossible to cut car use? Reducing dependency core to the debateOpens in new window ]
Beyond that, he worries that autonomous vehicles in general are leading us into problems with worsening congestion, and drawing expertise and investment away from public transport. “I think it’s definitely the case that, with autonomous cars, we’d just end up with worse congestion,” says Caulfield. “If people thought that they could live in, say, Kilkenny and commute to Dublin in their sleep, well, then we’d just end up with ever-more urban sprawl, much worse than what we have now. Also, the benefits of autonomous cars in terms of congestion and emissions only really come into play when every single car on the road is autonomous. Before that happens, the benefits just don’t accrue.
“I think autonomous tech is great in the sense of having more sensors on vehicles to improve safety and so on, but the idea of just sitting into a couch which effectively drives itself around? I don’t think that’s a great idea and I do think people would end up leaving public transport and opting for one of these pods. I think that the effort being put into these autonomous cars would be better expended on making electric car batteries better, or on making roads safer. I just don’t think they’re a good idea. It’s a bit of a Pandora’s Box, and we’ve bigger problems in transport than just making it easier to drive around.”
