A SoylentNews contributor writes:
This trade magazine, https://www.autonomousvehicleinternational.com/news/adas/mit-report-forecasts-global-emissions-impact-from-autonomous-vehicle-computers.html is reporting on a new study and modeling effort--
A new study conducted by MIT researchers has found that in the future, the amount of energy required to run computers on board an international fleet of AVs could generate the same amount of greenhouse gas emissions as all the world's current data centers. The study explored the potential energy consumption and related carbon emissions if autonomous vehicles were widely adopted.
At present, the data centers that house the physical computing infrastructure used for running applications account for approximately 0.3% of global greenhouse gas emissions (GHG). As there has been little focus on the potential footprint of AVs, MIT researchers developed a statistical model to study the potential issue.
The research team calculated that one billion AVs driving for an hour each day, with each vehicle's computer using 840W, would consume enough energy to generate roughly the same number of emissions as global data centers do currently. Researchers also found that in 90% of modeled scenarios, to keep AV emissions from surpassing present day data center emissions, the vehicle would have to use under 1.2kW of computing power. To achieve this target, more efficient hardware for AVs would be required.
In one test, the team modeled a scenario in 2050 where 95% of the global fleet is made up of AVs. During this scenario, computational workloads doubled every three years and the Earth continued to decarbonize at the current rate. Upon completion of this simulation, researchers found that 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 on board autonomous vehicles," said first author Soumya Sudhakar, a graduate student in aeronautics and astronautics. "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."
The source paper is here (abstract), https://ieeexplore.ieee.org/document/9942310 with the full text linked (may be behind an IEEE paywall?)
[Paper made available by MIT Open Access --hubie]
(Score: 4, Insightful) by SomeRandomGeek on Friday January 20 2023, @12:00AM (13 children)
So, the transportation sector currently generates 27% of all greenhouse gas emissions, but if we put big computers in all the cars it would generate 27.3% of all greenhouse gas emissions?
Sounds like someone doesn't understand Amdahl's Law. https://en.wikipedia.org/wiki/Amdahl%27s_law [wikipedia.org]
(Score: 2, Insightful) by Anonymous Coward on Friday January 20 2023, @01:31AM (11 children)
What I've read about self-driving cars says they will be used more, not less. For example, a fatcat could drive their AV into a big city (where there is no parking), send the AV out to drive around in a holding pattern while they attend a meeting, and then have it pick them up for the return to the 'burbs. Without self driving cars, they might take a taxi/uber that quickly picks up another fare.
Not sure what this does to Amdahl's Law--any thoughts?
(Score: 2, Interesting) by GloomMower on Friday January 20 2023, @02:14AM (8 children)
Could also be as likely that people will own less cars and you call your taxi/uber that is less expensive than now, as it is a self driving car that is near-by to take you were you need to go?
(Score: 1, Interesting) by Anonymous Coward on Friday January 20 2023, @03:05AM (7 children)
> you call your taxi/uber that is less expensive than now
Here's a lecture by an urban planner, traditionally (and logically) in favor of a variety of public transit (subways to Ubers). They noticed that even major efforts to get commuters out of personal cars were never very successful. So they decided to study the perceived value of owning a personal vehicle, https://www.youtube.com/watch?v=UN_eoAbKugk [youtube.com]
Quick summary, they tried to buy the participants away from having their own personal car. It's been awhile since I watched it, but from memory they offered free Uber and a variety of other enticements of fairly high value (as well as pointing out the savings from not owning a car). Results of the surveys are that, for people that currently have cars, it will be nearly impossible to take away (or "buy away") their personal car.
No matter how quickly the Uber (etc) gets to your door, it's not as convenient as a spur of the moment decision to get into your own car and go. Or as useful as keeping various personal items already stocked in your car. Or as pleasant as getting into your own clean or messy car (instead of someone else's mess). All these things add up to very high perceived value for personal cars. Logic doesn't seem to enter into this decision, the sense of personal mobility may be an emotional choice that can't be easily "bought".
(Score: 1) by khallow on Friday January 20 2023, @03:32AM (6 children)
The problem with assertions like this is that you're extrapolating from a starting point that's way back from the "no matter how" point. Doesn't make sense to speak of the supposed suspension of logic, when the alternative doesn't actually get to the point where logic would be challenged.
This reminds me of the occasional idiots arguing that I would irrationally hold to an opinion even if they were to provide a rational, good faith argument to counter it. How about you try that argument first?
(Score: 0) by Anonymous Coward on Friday January 20 2023, @03:39AM (3 children)
parent AC here.
Logic not my strong suit. It's not a bad lecture and describes nicely how they tried to configure surveys to answer their question (percevied value of owning your own car). Let us know what you think after reviewing the study.
(Score: 1) by khallow on Friday January 20 2023, @04:54AM (2 children)
(Score: 2) by SomeRandomGeek on Friday January 20 2023, @04:59PM (1 child)
There is the key element. There are lots of people who have been driving cars for their entire adult lives. Getting their license and owning their first car were rites of passage. It will be very difficult for these people to ever give up their personal cars. But they will eventually die off. The first generation after self driving cars go mainstream will be different. For them, getting a license and buying a car will be a huge hassle compared to just hailing a shared car.
(Score: 1) by khallow on Saturday January 21 2023, @06:10AM
(Score: 2, Interesting) by Anonymous Coward on Friday January 20 2023, @03:55AM (1 child)
(Score: 1) by khallow on Saturday January 21 2023, @06:55AM
(Score: 4, Interesting) by SomeRandomGeek on Friday January 20 2023, @05:16PM (1 child)
I find it very difficult to predict what the net effect of self-driving cars will be, because they will affect traffic patterns in so many different ways. A few of the more obvious ways:
1. An increase in traffic due to repositioning the car between rides from a place where people want to go to a place where people want to leave.
2. Smaller vehicles. People will stop using light duty trucks that they chose for their worst case cargo scenario to commute to work every day.
3. Higher mileage vehicles. Currently, cargo trucks have a life expectancy of millions of miles and passenger cars of a few hundred thousand miles. That is because passenger cars are used for only about an hour a day, and they reach the end of their lifespan due to age rather than mileage. Ride share vehicles see more action each day, they will see higher mileage over their shorter lifespan.
4. Electric vehicles. Rideshare vehicles rarely go on long trips. Therefore, EVs with short ranges but lower costs per mile are a great fit. So a transition to self driving vehicles is likely to speed the transition to EVs.
(Score: 0) by Anonymous Coward on Friday January 20 2023, @10:51PM
> People will stop using light duty trucks that they chose for their worst case cargo scenario to commute to work every day.
Okaaay. But imo many light duty trucks are bought for completely different reasons -- they are "force multipliers" for some who hate looking up in traffic, compensation for short dicks for others (often found outside bars/pubs), a way to tow a house/horse/toy trailer (ie, snowmobiles), and many other reasons including construction/trades.
I don't own a pickup (or SUV), but I work from home. If I had to fight an urban commute every day I might think about getting some sort of urban assault vehicle too. Maybe something like ths (my friend restored one)... https://www.thehenryford.org/collections-and-research/digital-collections/artifact/154693/ [thehenryford.org] From that seat height, we looked down into every pickup and SUV that went by us (grin).
(Score: 2, Interesting) by Anonymous Coward on Friday January 20 2023, @06:04AM
Well human drivers might use 100W-200W while driving BUT the self driving systems are likely to use more (computing, LIDAR etc). The 840W figure by MIT sounds plausible to me ( previous was 2.5kW https://archive.is/038BZ [archive.is] )
In contrast house flies, dragonflies and crows consume less than 13W. So our tech has a long way to go.
Even if dragonflies are dumber[1] than self-driving cars in terms of navigation and "driving abilities" I'm pretty sure ravens/crows are better.
http://people.eku.edu/ritchisong/birdmetabolism.html [eku.edu]
Ravens with 108kcal per kg/day are about 4.4W on average and that's including flying energy expenditures. So if you want to be fair you'd only count the "brain" energy consumption during flying and exclude the other consumption e.g. flight muscles, lungs, heart, etc.
[1] But are they really dumber than our current tech?
https://www.researchgate.net/publication/12609048_Prey_pursuit_and_interception_in_dragonflies [researchgate.net]
(Score: 2) by ElizabethGreene on Friday January 20 2023, @03:23AM (5 children)
There are, roughly, 1.4 billion cars, trucks including commercial trucks, and SUVs globally. Commercial vehicles make up about 10% of those. In the US, a very vehicle heavy country, the average person drives 46 minutes per day and commercial vehicles comprise roughly 10 hours per day of driving time. (The math is fuzzy on that with looser requirements in e.g. Alaska and rest break requirements depending on if your carrier operates 7 days a week or not. I'm going with 10/day.)
Doing the math, is it a reasonable assumption that 1 billion/1.4 billion = 70% of the entire world's vehicle population will adopt both AV systems and US-style driving habits? This seems a bit high.
More broadly, I don't think this is a problem we need to solve yet. When we do have functional systems then the cost of batteries will be a large enough economic incentive for automakers to keep a sharp eye on efficiency.
(Score: 1, Interesting) by Anonymous Coward on Friday January 20 2023, @07:42AM (1 child)
If there are 1 billion EVs their other electrical consumption will swamp that 800W per vehicle number anyway. e.g. a Tesla will use 7.5kW to 16kW while cruising at 35-55mph. https://insideevs.com/news/317340/heres-how-speed-impacts-range-of-the-tesla-model-s/ [insideevs.com]
Just air-conditioning or heating alone will use about 500W to 2.5kW ( https://www.reddit.com/r/teslamotors/comments/i2j71z/model_3_factfinding_air_conditioning_power_draw/ [reddit.com] ) .
If you want very warm temperatures in very cold winters while wearing light clothing you might use more than 2.5kW for heating especially if the vehicle is traveling fast enough for the air to take the heat away quickly.
So in the big picture this computing wattage stuff is a small problem compared to the bigger EV problem...
(Score: 2) by ChrisMaple on Sunday January 22 2023, @06:23AM
The key to better economy is and always has been multiple small improvements in every aspect of the vehicle. Reduce weight in hundreds of parts. Make tires with lower rolling resistance. Improve aerodynamics in numerous ways. Use less viscous lubricants. Improve engine efficiency, improve transmission efficiency. Every little bit helps, and the absurdly wasteful 800 W of computing power in a modern vehicle needs to be sharply cut.
For cars not running on batteries, the electricity to run computers comes from an alternator. From what I've read, the efficiency of automotive alternators is abysmal. That 800 W is sapping much more than 800 W from the engine.
(Score: 1) by khallow on Saturday January 21 2023, @07:03AM (2 children)
Keep in mind that with AV systems, you no longer are limited by the constraint to have a human driver. There's no reason that a country couldn't have a massively larger number of vehicles than it has people. If the vehicles drive themselves, then why couldn't the US have tens of billions of AVs? (just bumping up the number of vehicles by a bit more than an order of magnitude) Sure, you would need more road, cargo handling, and energy infrastructure for that situation, but we already know we can build infrastructure.
(Score: 2) by ElizabethGreene on Monday January 23 2023, @03:36PM (1 child)
My knee-jerk reaction is that's unrealistic as it exceeds the country's population by 1.5 orders of magnitude.
Rolling the idea around, I wonder if I might be missing something. We kicked off a technology revolution when we stopped treating compute as "special" and started "wasting" it on frivolities like GUIs. Is there an application for ubiquitous physical transport services?
If we pushed that capability into the logistics world we could create virtual warehouses by extending the just-in-time delivery idea so "trailers" of goods marshall until they are needed and then deliver to the customer site at the time of need. Energy costs money, so I'd guess they'd queue in parking lots instead of rolling around on the highway.
Ubiquitous personal transport would, I'd guess, reduce the need (but not necessarily the desire) for dedicated personal transport.
Package delivery: Less drivers, many more trucks, many more runs per day, but do you need a human in the loop for the last hundred feet of delivery?
I see less application for living "on the road". First, it's limited by the energy cost again, but there's another non-intuitive consideration. I lived in and traveled in an RV for a while, so I have some experience in this space. They don't advertise it on the tin but being on the road in one sucks. "Travel days" are equivalent (or more severe than) a significant magnitude earthquake. The chassis flexes and things break. Anything not tied down falls, rolls, bangs, and smashes, e.g. I broke several plates when the stack bounced from hitting a pothole. Put on the brakes and/or take a left turn too gingerly and now the contents of your fridge are all packed up against the door and ready to fall on the floor. Anything rigidly fixed to the structure has to be built like a tank or super flexible or it will be shaken to pieces. It's a problem.
Derailing the thread entirely, do you have any applications for ubiquitous transport?
(Score: 1) by khallow on Monday January 23 2023, @04:24PM
There already is application for ubiquitous physical transport services. The limitation of requiring a human driver doesn't limit it much at present. But an existing limitation of much of human infrastructure is that it requires massive human labor just to keep operating. For a historical example, one of the contributing factors to the abandonment of Angkor, a Cambodian city that was abandoned in the mid-15th century, was that the population had declined to the point that they could no longer had the labor maintain the extensive network of canals and other waterworks that the city needed to survive. A network of automated transport of materials wouldn't magically eliminate the need for human labor, but it allows for support of larger projects than one limited by presence of human labor.
(Score: 4, Funny) by Rosco P. Coltrane on Friday January 20 2023, @04:29AM (1 child)
the individual private vehicle is a lot more inefficient than well-run public transport solutions. Who would have thought...
(Score: 3, Touché) by DadaDoofy on Friday January 20 2023, @02:41PM
Only if public transportation goes to your destination.
(Score: 2) by Freeman on Friday January 20 2023, @02:50PM
Sure, if you live in one of the few big cities with a passable public transit system. You might be able to get away without having a car. Otherwise, you're going to be leaving yourself beholden to someone you're ride-sharing with or own your own car. Most of the time that means just owning your own car. Due to the extra complexity of planning your work life around someone else's schedule. The COVID-19 pandemic pushed some businesses to let people work from home. Now that it's essentially over. Most places have been back in the building for over a year, if not more. With very few holdovers that actually like the concept of having the majority of their workforce, work from home, so corporate doesn't have to rent/own as much office space. Or maybe they can lend their office space instead, which may make them even happier.
Joshua 1:9 "Be strong and of a good courage; be not afraid, neither be thou dismayed: for the Lord thy God is with thee"
(Score: 2) by DannyB on Friday January 20 2023, @03:37PM
Someday in the future, scientists are going to invent a way to build automobiles that do not require microprocessors.
It is believed among some that this may actually be possible.
The lower I set my standards the more accomplishments I have.