Autonomous taxis would deliver significant environmental and economic benefits, new study shows

Autonomous taxis would deliver significant environmental and economic benefits
Berkeley Lab researchers Jeff Greenblatt (left) and Samveg Saxena. Credit: Berkeley Lab

Imagine a fleet of driverless taxis roaming your city, ready to pick you up and take you to your destination at a moment's notice. While this may seem fantastical, it may be only a matter of time before it becomes reality. And according to a new study from Lawrence Berkeley National Laboratory (Berkeley Lab), such a system would both be cost-effective and greatly reduce per-mile emissions of greenhouse gases.

The analysis found that the per-mile emissions of an electric deployed as a self-driving, or autonomous, taxi in 2030 would be 63 to 82 percent lower than a projected 2030 hybrid vehicle driven as a privately owned car and 90 percent lower than a 2014 gasoline-powered private vehicle. Almost half of the savings is attributable to "right-sizing," where the size of the taxi deployed is tailored to each trip's occupancy needs.

The results were published online today by Nature Climate Change in an article titled, "Autonomous taxis could greatly reduce greenhouse gas emissions of U.S. light-duty vehicles," co-authored by Berkeley Lab scientists Jeffery Greenblatt and Samveg Saxena.

"When we first started looking at , we found that, of all the variables we could consider, the use of autonomous vehicles as part of a shared transit system seemed to be the biggest lever that pointed to lower energy use per mile," said Greenblatt.

Many automakers and other companies are working on autonomous cars. Right-sizing is cost-effective for both the fleet owner and for passengers, and small one- and two-seat vehicles are being explored by researchers and companies. To illustrate the concept, consider a single passenger with no luggage versus a party of four passengers with suitcases. The single passenger would require a much smaller taxi than the party of four, saving money for vehicle owners and passengers. Right-sizing, of course, assumes a fleet of taxis managed by a single entity.

"Most trips in the U.S. are taken singly, meaning one- or two-seat cars would satisfy most trips," Greenblatt said. "That gives us a factor of two savings, since smaller vehicles means reduced energy use and greenhouse gas emissions."

Would you get in a driverless car? Engineers think they could be safer and more efficient, but in this Nature Video we find out how they could be greener too. Great excuse to play with Lego!

Another factor contributing to lower emissions for autonomous taxis is a cleaner electric grid. By 2030 power plants are expected to be using more renewable energy and emitting less pollution, meaning the greenhouse gas intensity of electricity will be lower.

Self-driving cars have additional efficiencies that have been covered in other research, such as the ability to drive closely behind other autonomous cars to reduce wind resistance ("platooning"), optimally routing trips, and smoother acceleration and braking. "These are all incremental, but they do add up," Greenblatt said. "However, we didn't even include these effects in our baseline results, and we still get huge savings without them."

The researchers also conducted an economic analysis to determine how cost-effective autonomous taxis would be. At 12,000 miles per year, the average distance traveled in the U.S. for privately owned cars, electric vehicles in 2030 are still expected to be more expensive than owning and operating a gasoline-powered car, the study found.

But if the vehicle is driven 40,000 to 70,000 miles per year, typical for U.S. taxis, they found that an alternative-fuel vehicle (hydrogen fuel cell or electric battery) was the most cost-effective option. This was based on costs for maintenance, fuel, insurance, and the actual cost of the vehicle (assuming a five-year loan). The reason is that despite the higher cost of a more efficient vehicle, the per-mile cost of fuel is lower, so the savings can pay for the extra investment.

"You don't often find that, where the cheapest is also the greenest," Greenblatt said.

While autonomous technology is currently estimated to add as much as $150,000 to the cost of a vehicle, an autonomous taxi using today's technology would still be cheaper than an ordinary taxi not simply due to its greater energy efficiency, but also due to the fact that no operator would be required. By 2030, autonomous taxis could be far cheaper than their driven counterparts.

The scientists did not try to estimate how widespread this technology would be in 2030. However, they did calculate that if five percent of 2030 vehicle sales (about 800,000 vehicles) were shifted to autonomous taxis, it would save about 7 million barrels of oil per year and reduce annual by between 2.1 and 2.4 million metric tons of CO2 per year, equal to the emissions savings from more than 1,000 two-megawatt wind turbines.

To estimate the number of trips taken by different numbers of occupants, the researchers analyzed National Household Travel Survey data from the Federal Highway Administration. The scientists then modeled hypothetical one- and two-seat vehicles based on Nissan Leaf parameters driving three test-drive cycles as defined by the Environmental Protection Agency (EPA) using Autonomie, a vehicle-modeling tool developed by Argonne National Laboratory that simulates energy consumption on a second-by-second basis.

Furthermore, the scientists explored the net energy effect of combining ride-sharing with right-sized autonomous taxis. For example, if 10 percent of one-person rides were shifted to two-person rides, the total miles traveled would decrease 3.1 percent while average energy consumption (due to the larger vehicle) would increase 0.6 percent, resulting in a net energy decrease of 2.5 percent.

Greenblatt said plans for further study on this topic are in the works, including exploring the effect of battery degradation, looking at optimal vehicle designs, and making a more realistic simulation of how a fleet of autonomous taxis would actually operate in a metropolitan area.


Explore further

CEO: Nissan will be ready with autonomous driving by 2020

More information: Nature Climate Change, DOI: 10.1038/nclimate2685
Journal information: Nature Climate Change

Citation: Autonomous taxis would deliver significant environmental and economic benefits, new study shows (2015, July 6) retrieved 16 October 2019 from https://phys.org/news/2015-07-autonomous-taxis-significant-environmental-economic.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.
833 shares

Feedback to editors

User comments

Jul 06, 2015
With unemployment at unacceptable levels for many years and with no relief in sight, I wonder at the propensity for the construction of fanciful scenarios of self driving electric cars having a cost advantage by putting more people out of work.

The desire to induce poverty in the general population is troubling.

Jul 06, 2015
Yeah, it reminds me of the pay at the pump no-fault car insurance idea. Great plan, very efficient. Only it puts about a million people out of work.

The taxi industry has been a stepping stone profession for generations, a place to get a foothold in our economy, while your children take the next step. Wiping that out is not a benefit to society, you have to weigh the consequences of these ideas.

Jul 06, 2015
Autonomous taxis would deliver significant environmental and economic benefits


And put lost of Pakis and other third worlders out of work..

Jul 06, 2015
So... what, we should turn down technologies to employ people for jobs they aren't needed for?

The reality is that there is a bigger, broader economic change we're in the middle of. Like the change that came with the industrial revolution. This economy will eliminate broad sectors of employment, no doubt. Anything that is repetitive and automatable probably will, within a generation or two, be automated.

The question is, as a society, if we're willing to accept that an economy can support more people than are needed to do work. Ie, if our production, P is a product of workers, w, and some productivity factor, F (P=wF), what happens when F grows at a fast rate? If F grows faster than w diminishes (ie, the productivity per worker improves more quickly than we lose workers), then our economy still continues to grow.

//cont. below.

Jul 06, 2015
The change that will be needed is a general update to 18th century economic models (capitalism) that were developed for industrial revolution era production.

I'd say the most likely change will be something along the lines of a guaranteed basic income. If our production is significantly greater than a basic level of income for all families in the country, we can distribute that money so that everyone has a (simple) place to live and food to eat. Then, if you work, you're working for luxuries and conveniences. Or you're creating art or doing something in your community socially, but still able to live.

That, in a way, is the promise of the automated workforce. That humans should be free from the need to labor, free to explore alternatives, without suffering in that exploration.

(Whether this is realistic to see implemented or not is... as always, another matter entirely)

Jul 06, 2015
Well it may be overly optimistic to assume that people who lose their jobs in one industry will be able to easily transition to another. Without significant reform to job training and our post-secondary education system, I doubt a cab driver can simply "move" to an in-demand field like health care or engineering.

Jul 06, 2015
A bad idea from well meaning tech lovers. The notion that robot driven cars are safer is a logical fallacy. No new tech is ever as good as advertised until its been through numerous upgrade/revisions that come from feed back of real world use that typically includes many failures. Even if the robots could be made perfect the software they run on is programmed by humans who do make mistakes.

Jul 07, 2015
A bad idea from well meaning tech lovers. The notion that robot driven cars are safer is a logical fallacy. No new tech is ever as good as advertised until its been through numerous upgrade/revisions that come from feed back of real world use that typically includes many failures. Even if the robots could be made perfect the software they run on is programmed by humans who do make mistakes.

But robots don't get drunk, don't smoke, don't divert their attention from their duty, don't miss a stop or a red light... Driving is not a creative job (at normal speed, anyway) robots can do it much better than humans.

Jul 07, 2015
If F grows faster than w diminishes (ie, the productivity per worker improves more quickly than we lose workers), then our economy still continues to grow.


You forget that the population also increases.

In the US history for example, the number of industrial manufacturing workers remained almost constant between circa 1950-2000 while the output per worker roughly doubled thanks to automation. Meanwhile, the population increased 3x, which means tangible productivity per capita actually decreased by 33%

We actually became poorer despite the automation.

We're accumulating a class of people who only consume because income has been decoupled from actual productive labor. The problem then becomes that this class of people grows faster - both in proportion and in absolute numbers - than the one that actually works to produce the goods that everyone depends on.


Jul 07, 2015
And it's not just the rich bankers. It's everyone.

Currently this class of idle people employs themselves by doing something entirely trivial like serving expensive coffee, which ultimately benefits no-one. On the contrary: people are "working" to grab money off of each other without realizing that money isn't wealth.

All the stuff that they're not producing is wealth.

Tempting as it would be, the worst thing you could do is to say you no longer need to work to make a living - that you get some sort of "citizen's wage" just for existing - because with free money the idle class would simply explode in numbers and the problem would get worse.

There wouldn't be anyone left even to operate the machines. Why do anything as boring as that when you can simply sit back and browse the internet and the government will pay you for existing?

Jul 07, 2015
If F grows faster than w diminishes (ie, the productivity per worker improves more quickly than we lose workers), then our economy still continues to grow.

True. However we need to add an additional factor into this: The benefits of that increased economy only goes to the employed (or rather - the employers. The employees haven't been seeing much of this increase in economy for some decades). So there needs to be an ancillary change in society from "you have to work to make a living" to "you may work if you want to but will get an income anyways".
Or we just start widespread/mandatory job-sharing.

The transition period will be difficult, because there will be regions and sectors of employment where this will go at a different pace. This will be quite a challenge - but one we must eventually face.

Jul 07, 2015
So there needs to be an ancillary change in society from "you have to work to make a living" to "you may work if you want to but will get an income anyways".


What people are really saying when they say we no longer need everyone working in order to produce goods for everyone, is rather: "We don't need -me- working to make -me- a living".

Even if machines could do all the work, we would still need some non-political means to distribute wealth among the people to prevent an explosion of an idle class that would gobble up all the resources and pull everyone down to poverty. Non-political, because when people can vote themselves wealth, the democratic system will simply collapse.

In that sense, it would be more rational not to have robotic taxi drivers in favor of having something to do for real taxi drivers.

Jul 07, 2015
Or we just start widespread/mandatory job-sharing.


Oh, like forcing someone to give up some of their hours when they're already making minimum wage?

In general, it would be wasteful to have two people work a one-man job, due to the overhead of training both for the job and swapping them in and out. Imagine the amount of traffic generated by a factory running on six shifts instead of three.

Jul 07, 2015

I'd say the most likely change will be something along the lines of a guaranteed basic income. If our production is significantly greater than a basic level of income for all families in the country, we can distribute that money so that everyone has a (simple) place to live and food to eat. Then, if you work, you're working for luxuries and conveniences. Or you're creating art or doing something in your community socially, but still able to live.

That, in a way, is the promise of the automated workforce. That humans should be free from the need to labor, free to explore alternatives, without suffering in that exploration.


I love your optimism, sounds like something the venus project is working towards

Jul 07, 2015
Even if machines could do all the work, we would still need some non-political means to distribute wealth among the people to prevent an explosion of an idle class that would gobble up all the resources

False dichotomy. Just because people can be idle does not mean that they can therefore demand access to unlimited resources. The limit of resources is determined by what the economy can provide. You may just have to decide how to spend the 'resource points' that are allotted to you (and if you want more then you should find a job that creates this additional wealth or others willing to donate their points to you)


Oh, like forcing someone to give up some of their hours when they're already making minimum wage?

Straw man. That would mean reducing the income per person. We're talking here if enough products/resources are available per person thoruhgg increased production by robots. (I.e. raising minimum wages and THEN job sharing).

Jul 08, 2015
False dichotomy. Just because people can be idle does not mean that they can therefore demand access to unlimited resources.


Isn't - they can because I'm assuming we're talking about a democracy or a republic instead of a despotic dictatoriship.

It's an inevitable conclusion that every politician will promise their constituents more resources in exchange for votes, directly or indirectly, and the whole political system collapses into special interest groups that are fighting to corner the most resources for themselves using whatever means possible.

ou may just have to decide how to spend the 'resource points' that are allotted to you


Allotted by whom?

Straw man. That would mean reducing the income per person.


It does mean reducing the income per person. We're not talking about increased production here, but two people doing the job of one: more workers, same output. That's job sharing.

The robot argument is extra.

Jul 08, 2015
We're talking here if enough products/resources are available per person thoruhgg increased production by robots. (I.e. raising minimum wages and THEN job sharing).


The point I was making earlier is that the amount of resources available per person is not going up with increased productivity because the idle class keeps growing faster than the robots manage to increase productivity.

So you're talking of a hypothetical situation where we've already managed to end the explosion of the idle class, and then somehow allotted the working hours to the people, which actually just reduces back to the politically determined income distribution.

The job sharing scheme was supposed to be a solution to the problem of the idle class, but it can't be done until the problem of the idle class is solved.

Jul 08, 2015
The point I was making earlier is that the amount of resources available per person is not going up with increased productivity because the idle class keeps growing faster than the robots manage to increase productivity.

Whut? Robots increase productivity (that's the whole point of having them). Populations in the developed world are already stagnating or even declining insome cases. How does that work mathematically that resources per person drop under those conditions?

Because it doesn't matter if the idle class explodes or not. The total number of people in a nation matter. The ratio of idle to non-idle is completely irrelevant.

Jul 08, 2015
Whut? Robots increase productivity (that's the whole point of having them). Populations in the developed world are already stagnating or even declining insome cases. How does that work mathematically that resources per person drop under those conditions?


I gave a specific example. The number of US manufacturing jobs have remained constant, while the population has tripled, while the productivity in manufacturing has doubled. The production per capita has therefore decreased by a third.

The stagnation in populations today is largely due to the baby boomers who are increasingly dying at the same time and not contributing to new births.

Because it doesn't matter if the idle class explodes or not. The total number of people in a nation matter. The ratio of idle to non-idle is completely irrelevant.


What nonsense. Of course it matters.

Jul 08, 2015
The total number of people in a nation matter. The ratio of idle to non-idle is completely irrelevant.


In order to keep everyone's living standard the same the idle class still must not grow any faster than the increase in productivity of the producing class.

If productivity doubles, but the size of the producing class more than halves, then the per capita productivity of the society declines regardless of the total population size. It doesn't matter whether you do this comparison in absolute or relative terms.


Jul 08, 2015
Furthermore, the dominant strategy for a group that is looking to increase its own wealth and political power in an economy where each individual gets a set amount, is to increase the size of the group.

Recruit more people, and when you run out of willing recruits to your side, start making more people.

That's why the idea that we'll just start paying people for existing is going to backfire terribly. We don't have enough to give everyone a decent living standard, and once we start divvying things up, people will respond by making more people in a bid to catch a larger share to themselves.

You'll end up policing people like the communist China does.

Jul 08, 2015
How does that work mathematically that resources per person drop under those conditions?


If A is the size of the idle class and B is the productive class, and P is productivity, and C is "consumptivity" as a measure of living standard, then:

C(A + B) = BP

P = CA/B + C

If we hold a constant C = 1, which is to say we don't change the living standards, then P = A/B + 1 or in other words, changes in productivity have to match the changes in proportion of idle to productive people regardless of population size. If dP < d(A/B) then C must decrease to compensate: living standards must drop.


Jul 08, 2015
To sum the argument up:

Because we can't get rid of the real people who currently drives taxis, having a robotic taxi driver is probably going to be a net loss to the society because the robot has a cost, and the recently unemployed taxi driver likely won't find himself anything productive to do and therefore simply increases the size of the idle class who either does no work, or does unproductive and wasteful work.

The productivity increase of the robot would need to be higher than the cost of having another idle worker. It's not enough that it simply replaces the taxi driver.

KBK
Jul 11, 2015
i don't have a problem with eliminated the 250,000 cab driving jobs in the USA alone.

I'd agree, ONLY of we also end the existence of those who wish to squeeze the citizenry down into stupidity and death, at the same time.

We're also looking at eliminating the trucking industry, in total, at the same time. The trucking industry would go like a domino.....

the drivers ( 3,500,000 jobs) in the trucking industry would go first (in their system)..and... since they are not around to complain, some of the freight would move to rail transport. Freight handling jobs might temporarily increase slightly, but them go down as that gets automated as well.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more