I’m not sure how excited to be about the likelihood of autonomous vehicles.
From a climate perspective, it is hard to know whether vehicle autonomy will help. “Autonomous Connected Electric Shared Vehicles” (ACES) may help reduce carbon emissions by accelerating adoption of electric vehicles and possibly by encouraging ride sharing. Or, self-driving vehicles may lead to more driving, including driving of traditionally-powered vehicles.
Similarly, from an equity perspective, autonomous vehicles may give people new inexpensive mobility, but they may also destroy an major job category that does not require much education.
Those with a passion for the ACES vision foresee a compelling and rapid transition, like the single-decade transition from horse carriages to cars that happened in the last century: The elimination of driver labor will lower the cost of a chauffeured ride to the point where most people will suddenly find that using “Transportation as a service” (TaaS) is both cheaper and preferable to individual car ownership. Enormous value will be created for consumers because they will not have to spend money on buying and maintaining their own car. Shared vehicles will be marketable at a higher price point, facilitating the needed transition to electric vehicles. Fleets of self-driving vehicles that can be summonsed from a smartphone will move efficiently through the streets stopping only for a rapid recharge at well-timed moments. Both cities and suburbs will become greener as parking lots become unnecessary.
I wouldn’t bet against the technology. Major automakers, as well as Google and Amazon, are pouring money into it. The prototypes are working better and better. When a model is perfected, it could be licensed to multiple automakers.
I hope I am wrong, but my instinct is that owning a personal vehicle will remain a high priority for many people even when TaaS becomes less expensive because vehicles can drive themselves.
To the extent we continue to travel daily to work and school at around the same time, we will still have a rush hour. The total vehicle fleet will continue to need to be big enough to handle all of that simultaneous demand. In other words, it will have to be almost as big as our existing collection of individually owned vehicles. At full scale, TaaS would not have a big cost advantage over private vehicle ownership.
If many more people were willing to routinely share rides, that could increase the TaaS cost advantage. Shared rides have always been encouraged as a solution to congestion. In any region at rush hour, there are thousands of people making the same trip at the same time. If they could share rides we would see a dramatic reduction in vehicles on the road. Apps can match riders more efficiently than workplace bulletin boards.
But car pooling peaked in the energy crisis of the 70s and declined as gas prices went down and private vehicle ownership has increased. People are not necessarily looking for companionship on the road. People under time pressure do not want to detour to pick up or drop off another rider. Many people use their car to store convenience items.
The limited market penetration of Uber and Lyft offer an indication of the limits of Transportation as a Service. While they have wreaked havoc on the taxi business, they accounted for only 1.3% of all passenger trips in 2019 in Massachusetts. And those were shorter trips on average, heavily concentrated in urban areas, more than half starting in Boston or Cambridge.
Use of ride service vehicles could become even less attractive without drivers. Service drivers are screened for criminal records and serve as a third party to referee disputes and oversee the interior space of a vehicle. Sharing the back seat with a random stranger is even less appealing if there is no one else in the vehicle. Remote supervision by camera could offer some assurance of safety and deter some crime, but a remote supervisor could not immediately intervene within the vehicle. And remote supervisors cannot do much about the spilled coffee and worse things that humans leave behind.
Even if private vehicle ownership persists, that does not mean that autonomous vehicles will make no difference. Autonomous vehicles that communicate with each other may be able to drive closer to one another and increase the capacity of high speed roads. (See the Governor’s Commission on the Future of Transportation, Volume II, page 64 for this point and a good overview of the autonomous vehicles.) Autonomous taxis could offer less expensive mobility options in poverty areas that are currently inadequately served.
Autonomous vehicles could help offer “last mile” connections to mass transit (although at rush hour, the simultaneous station arrivals would likely be too numerous for the shared autonomous fleet to support). For a thoughtful exploration of the impact of shared autonomous vehicles in cities that combine auto use with transit see MIT’s Insights into Future Mobility (Chapter 6); the upshot is that unless carefully coordinated with transit, they will like result in less transit use and more congestion.
Autonomous vehicles could free up the attention of drivers and allow them to safely read or converse, perhaps combining the comfort of a train ride with the privacy of a personal vehicle. Unfortunately, this could draw more longer haul commuters off of rail and contribute to congestion.
Autonomous vehicles may have a larger role to play in trucking. Autonomous vehicles do best in simpler environments like limited access highways. They could reliably handle the long-haul freeway driving, waking up the truck driver to handle the final maneuvering for delivery.
The physical measures the public sector might need to take to support autonomous vehicles are mostly the same kinds of measures we are already taking to support electric vehicle expansion (charging stations) and to support human-driven transportation network companies like Uber and Lyft (pickup locations, coordination with transit, etc.).
There are likely some regulatory changes needed to clarify rules and liability for autonomous vehicles. In 2016, the Governor signed an executive order creating an initial framework for autonomous vehicle testing and a MassDOT working group has produced a report suggesting areas for legislative adjustment.
For now, my take is that we should make the necessary adjustments to facilitate progress, but we should not expect autonomous vehicles to solve our clean energy challenges in transportation.
Very thoughtful comments. for me the real issue is housing, that is where people will live, and how that will impact commutes, which i think account for much of the negatives around individual car use. That is if work can be easily accessed by public transportation, whether or not using autonomous vehicles, then the discussion becomes much different.
And for me, the issue of housing is the most important issue, period. That is successfully cultural/race/demographic housing is the key to our future as a civilization. And here in MA we have a real chance to address, if not solve that problem by making it top of mind in all public policy areas, and being aggressive and creative as we think and plan for housing in the future. I believe that the future is in more dense than suburban type housing. As demographics change, more people waiting to have families and more options for family living in urban areas, transportation options will change as well. And of course schools, shopping and more.
When the research on pandemic living is presented, I do not think there will be a great cry for more commute time, rather a change in the work dynamic, fewer days in person, modified daily schedules – 11am -2pm for example. I could go on, but my instinct is that the experience of the 22-30’s in the time of covid will dramatically change their outlook and desires. And as someone who raised 2 kids from ages 2-20+ in the city, going to city schools, playing and shopping many years without a car, just using day rentals when needed to go skiing, etc. , it’s a great way to live.
NB What about converting empty offices in cities and industrial parks on T lines to mixed use housing? even high rises?
Agreed. Housing will remain a central part of the conversation. I’m less certain what the future of housing policy looks like post-COVID, but will be paying close attention.
Using public transportation plus occasional taxi runs is cheaper than owning a car. Uber is a bit cheaper than a taxi. But car ownership, like home ownership, is not mainly about the money. We often hear about the networks of trolley lines that pre-dated our automotive culture, they disappeared pretty fast when people could buy a car.
I think the question of refactoring our domestic environment is at the heart of this issue. To the extent that we can, we should be supporting walkable access to resources rather than centralized resources. Reverse the trend of consolidating schools and districts, encourage small neighborhood businesses, post offices. Encourage mixed use zoning, so that businesses are placed in walking distance. Encourage small business. Discourage factory farming to diversify the food supply chain. Discourage business merger and acquisition, or maybe tax M&A. We need to scale down and localize so that we can reduce reliance on oil for transportation. Really at all times we have to become aware of the conception that people will drive to our businesses and facilities, could they walk instead?
Also, I definitely take your point on encouraging the growth of charging networks. And it would be ideal if we supported homeowners and business more strongly to install solar and wind generation. What can be done to engineer the power network to distribute excess capacity more efficiently? Can we encourage businesses to install Tesla type “powerwall” batteries for the low generation hours (and for outages.) Can we encourage solar roofs? Could we support a DARPA like grant program to support development of alternative battery technologies, that require less destructive mining activity?
Thanks, Rich. It may well be that we can down scale travel and live more locally as more work from home. I agree that is an attractive vision. Small businesses outside the urban core may become more viable. It’s a tough transition for the downtown service businesses though.
The primary reason to have autonomous vehicles on the market is safety of pedestrians and other motorists. I am deeply concerned about any visual deficiencies that cue these vehicles to operate safely. Many towns, including mine, cannot or don’t keep up with striping, curbing, and other visual and physical cues that ensure pedestrian and motor safety.
I also think control for demand and use is more easily facilitated and taxed with automation, similar to time of use demands for energy and smart meters. That should yield true costs and those costs should be appropriately shouldered.
As a user of the roadways, we use one car, multiple bikes, our sneakers, the MBTA buses, MBTA trains, Peter Pan, Go Bus, Amtrak, taxis, Zipcar, Lyft, and Uber. When Uber and Lyft became popular, I was supportive, but a couple of bad experiences with drivers and I no longer use them unless I am traveling in a group and with a man, though my husband still does. If the autonomous vehicles were single rider vehicles, I may use them in lieu of taxis.
I like transit that allows me to talk with others and not create a safety hazard, and ideally provide some movement options. I prefer the bus over taxi because I can stand and move around, and often meet neighbors and connect with them. I like trains for the same reason. I wish they were all more frequent, more convenient, and faster.
I support Seth Mouton’s call for high speed rail.
I think your comment about only wanting to use ACES if they were single occupancy is very important — you speak for many and that is, I think one of the real barriers to the ACES vision.
I think the key disagreement (or point of lack of agreement) I have with your analysis is around the effect of autonomous vehicles on car ownership. My understanding is that there are a lot of young people who are choosing not to have cars in the current climate, and I expect that that would increase a lot in an autonomous vehicle context. If so, the carbon costs of production of cars could be substantially decreased, because many fewer cars are needed.
I also wonder if more miles are driven by autonomous vehicles, if that will make electrification and other carbon-saving measures easier because there’s one or a small number of fleets to upgrade.
I do agree that ride sharing is unlikely to increase, and more miles may be traveled because of the increased convenience.
I do agree that within cities, on-demand vehicles support a lifestyle without car ownership. Not sure how much of a factor changing attitudes are. The MIT mobility study reaches this conclusion: “Although anecdotal evidence has been cited to suggest that millennials in the U.S. are less interested in car ownership, our econometric analysis concludes that this generation’s appetite for cars and car travel is very similar to that of older generations after accounting for relevant socio-economic factors, including income and urbanization.” Insights into Future Mobility, page 54.
I suspect that autonomous transport will change the look and feel of our streets in two ways. The first is via vehicles that carry humans, which is the topic discussed here. The other will be in delivery vehicles. For instance, Domino’s has just started delivering pizzas in Houston autonomously. (https://www.motorauthority.com/news/1131910_domino-s-launches-autonomous-pizza-delivery-with-self-driving-robot-car)
Exactly what proportion of the street/sidewalk traffic they will make up will depend on fees and taxes. But if the price is right — a very big if — I don’t expect this — I would use them for everything.
What matters to me as a patriotic freedom loving American; the free market should determine what we buy. Not our power hungry over powering government. We must not be dictated to buy panic peddlers who influence people by puffing. Government is incapable of solving any problem without creating other problems.
Yes but those who have market power in the free market have a big influence on what choices of freedom you are offered in the market. At least when the public, through government works through what the choices might be offered, some good and some less good choices are likely to be considered. Not just the choices that ensure that the interests with market power keep their market power. I believe the reason that current dominate private industry players are so interested in Autonomous Vehicles, is that the decision made about the form and function of transportation will determine the dominate interest’s profits and economic power into the future. This should not be confused with decisions that are in the best interest of the public, the environment, where and how we build our communities or even transportation. Our dominate economic interest’s goal is to continue to profit off of the design of how society is structured and its current infrastructure and the built in need for transportation. Other choices may offer other types of freedom, like not needing a vehicle or a transportation service to function day to day.
“The total vehicle fleet will continue to need to be big enough to handle all of that simultaneous [rush hour] demand. In other words, it will have to be almost as big as our existing collection of individually owned vehicles.”
The latter would follow logically if a large fraction of all individually owned vehicles are active during rush hour… But I see that this suggests otherwise:
I’d be curious if anyone knows of other analysis of that subject.
FWIW, while looking for info on that matter, I ran across this in regard to Mass DOT’s expectations for post vaccine rebounds in commuter traffic:
Steve, thank you for sharing the Rethinkx link. And here is the link to the study they cite 2009 National Household Travel Survey (see page 52). There is a 2017 version of the survey too.
It is definitely true that the rush hours account for only a fraction of the travel. That is even true on our major commuting highways.
But I think it is a different question as to how what share of the fleet is deployed at rush hour. I read the data as indicating that a very high share of the fleet is deployed at rush hour and that is the challenge for a shared model.
Senator B.: “I read the data as indicating that a very high share of the fleet is deployed at rush hour and that is the challenge for a shared model.”
I would be curious to understand more precisely how the data lead you to that interpretation. The graph on page 52 doesn’t show more than about 20% variation from 11:30 to 5:30. And ride trips at the daily peak don’t seem much more than double the average or median… While lots of sources suggest that private cars typically spend about 95% of their time idle (e.g., https://senseable.mit.edu/unparking/). Hard for me to square all that with the notion that throughout the quarter of the day starting a little before noon, a majority of cars are in active use. Though perhaps much shorter average trip lengths outside commuting hours could help explain…
Surprisingly hard to locate estimates of the percent of total fleet that is active at rush hour…
However, in regard to potential reductions in the total vehicle fleet, the rethinkx folks cite a study from the International Transport Forum
asserting that a fully autonomous fleet could meet demand at all times in Lisbon, Portugal with only 3% of the current vehicle fleet. While the rethinkx model “sees just over 11% – a more conservative figure.”
They write that “there is a growing body of academic research (for example https://www.caee.utexas.edu/prof/kockelman/public_html/TRB15SAVsinAustin.pdf and
http://www.caee.utexas.edu/prof/kockelman/public_html/TRB16SAEVs100mi.pdf) that has modeled the potential reduction in vehicle numbers due to TaaS, with figures largely falling between 5 and 11 vehicles taken out of the fleet for every TaaS vehicle. Our model equates to 8.6 to 1 (232m individual vehicles replaced by 26m TaaS vehicles). However these estimate tend not to include any smoothing of peak demand through demand-based pricing, which could increase these figures.”
The ITF’s Lisbon model depends heavily on use of six-seat shared taxies in conjunction with “a fleet of eight-person and 16-person mini-buses (‘Taxi-Buses’) that serve pop-up stops on demand and provide transfer-free rides. Rail and subway services keep operating in the current pattern.” I would agree with your skepticism about any similar degree of shared occupancy here in Boston… but remain optimistic that there can be a great many fewer vehicles in operation once vehicle autonomy is pervasive.
I would be most concerned with the loss of being able to travel when you want. Undoubtedly, this would end up with a few oligarchs controlling a few transport companies, with the power to control peoples movements. Probably will tie in with your Covid passport. We have seen quite a bit of these kind of restrictions over the past year. I’m not a fan of centralized power like this. We should rethink this obsession with Co2 reduction. It is wrong on many levels. There is even a movement of enviros who want to install methane collectors on cows backs to capture their flatulence. Total insanity.
Privacy is definitely an issue in an ACES model.
Perhaps we should think more broadly about vehicle categories. For example a commuter could have a single passenger autonomous vehicle that could significantly reduce congestion. Also if Uber had single passenger autonomous vehicles at reasonable cost, it would reduce congestion within cities. Just a thought
I am going to ignore the climate deniers and instead focus on what I think is not being talked about enough – the safety, or lack of safety, of autonomous vehicles. I work in technology and as unsafe as human drivers are, once you cut through industry propaganda, autonomous vehicles are nowhere near safe enough to be on our streets. Even in limited on-street tests they are hitting and killing people.
We already do next to nothing to ensure the vehicles on our streets are safe (see European safety standards, which require the safety of pedestrians and bikers to be factored into a vehicles safety rating, to American “safety” standards, which only factor in the safety of the driver, whoever they hit be damned). (Also see the new EV Hummer, which is physically wider than an average urban street!).
Given our track record in America, it’s reasonable to predict that pedestrian and bike safety in relation to autonomous vehicles will likely ignored, or even demonized. We need truly safer cars and safer streets before we can even consider autonomous vehicles.
Completely agree. Would love to see that the cars made to fit the roads we have.
Safety is definitely a continuing top legislative priority.
I think now is a great time to look at how we can integrate high speed trains to move people distances joined with local bike paths, autonomous vehicles, buses, and other modes of transportation. One of the big issues today is that cars are by far the quickest way to get to many locations but too much time over mass transit. If the time was significantly closer many would leave their cars home. This is the role for planning to envision how this can be done over the next 10 years. Autonomous vehicles will take some time before they will be able to safely operate in our curves and snow covered roads so we have time to do this thoughtfully.
Agreed that coordination with the transit system to meet last mile demand is a goal for any ride-sharing service, whether autonomous or human driven.
If our incredibly destructive income distribution, driven by relaxed rules and imbalance of economic power (high corporate person versus low human person bargaining power) and a host of other biased conditions (See Joe Stiglitz “Rules”), if all were redressed, there could be: higher use of transit, lower use of personal car, reduced demand for ride companies, more walking and biking, better ‘climate’, and superfluous self driving cars.
I believe that autonomous vehicles COULD be a great service for many groups of people including the elderly and disabled individuals who can’t or don’t have access to their own vehicle.
Having said that, I would not count on autonomous technology maturing anytime soon. As a computer scientist, I like to say that “artificial intelligence” (the technology behind vehicle autonomy) has a lot of “artificial” and very little “intelligence.” Believe me, it’s not a technology that is anywhere near ready for prime time despite what any automakers or software companies would have you believe.
In terms of environmental costs, the number of autonomous vehicles would need to match current peak capacity (i.e., rush hour) as was mentioned. So I don’t see a huge reduction in the absolute number of vehicles, certainly not an order of magnitude, for instance. Also, to be available for constant use, these vehicles will need to be charged continually throughout the day, so power consumption and demands on the grid are not likely to decline. So, bottom line: autonomous vehicles are not likely to be THE technology that contributes greatly to carbon reduction within the next decade.
I’d like to point out that bicycles would have greater potential if there weren’t so many roadblocks to their use. Los Angeles permits bikes on all public transit vehicles (subway, light rail and buses), every day, throughout the day; see: https://www.metro.net/riding/go-bike/bike-transit/.
Contrast that with the T’s policy at https://www.mbta.com/bikes. So if one could bike to a bus or T stop, ride public transport, and then bike the rest of the way, public transport use would be more attractive to more people.
If we could integrate the Blue Bikes system with the T, so that you don’t have to have separate memberships, that would be really great.
Well taken points.
If crowding on the T is permanently reduced we would have a good case for a more liberal bike policy. It’s tough to bring a bike onto a car that is loaded to crush capacity, but maybe we won’t be doing that anymore.
So many people missing the point. Driving is a sport. Commuting should be done on foot or by bicycle. Autonomous cars are the work of the devil. I am not sure about autonomous motorcycles.
I’m not too excited about autonomous vehicles. Too many unknowns.
Technology will make autonomous vehicles a reality, but they should be introduced slowly. Perhaps limit autonomous vehicles to intercity traffic, at the same time excluding manned vehicles. Then let’s see what happens.
I do think autonomous vehicles will allow for more people to opt-out of not owning an automobile, in part as you have suggested by making services like Uber and Lyft not require a driver, which will make them cheaper to use and likely more readily available, as self-driving fleets can be deployed quite easily. You are also correct that the low capacity nature of these vehicles could easily result in worse congestion as more people opt to use them.
I would like us to also recognize that automated operation will also benefit transit vehicles. Transit agencies may eventually be able to run driverless trains, buses, or vans on fixed or on-demand routes. If these routes can be run in a dedicated reservation or bus lanes, automated operation would be even easier.
I find this Jan 2021 presentation extraordinarily helpful in understanding the technical path towards autonomous driving… Prof. Amnon Shashua, CEO and president of Intel’s Mobileye
I am curious to know the degree to which Massachusetts has embraced and implemented the recommendations of:
“How Governments Can Promote Automated Driving”
That was published in 2017 but includes timeless wisdom such as: “Key developers of automated systems… are generally wary of both process (legislative and potentially administrative efforts in multiple states) and product (disparate legal regimes that create confusion, inconsistency, and unintended impediments to innovation). A legislator who introduces a bill without consulting these developers may get their attention—but probably not their affection.” Starting on page 113, the article includes a great many practical recommendations for government action.
The author, Bryant Walker Smith,
was also a co-author of the “taxonomy with detailed definitions for six levels of driving automation, ranging from no driving automation (level 0) to full driving automation (level 5), in the context of motor vehicles”
I think he also may have helped develop the 2018 model law proposed by the Self-Driving Coalition:
“Argo AI, Aurora, Cruise, Embark, Ford, Kodiak, Lyft, Motional, Nuro, TuSimple, Uber, Volvo Cars and Waymo are part of the Self-Driving Coalition to work with lawmakers, regulators, and the public to realize the safety and societal benefits of fully self-driving vehicles.”
“State and local policymakers: If you are thinking about taking action in your area to develop policies for self-driving cars, we’d like to hear from you. Please take a moment to email us at email@example.com and let us know how we can help.”
Steve, thanks so much for these helpful citations.
I’ve added above in the body of the post a mention of the following document REPORT OF THE MASSACHUSETTS AUTONOMOUS VEHICLES WORKING GROUP which does identify a number of specific policy measures to support autonomous vehicle development in Massachusetts.
Comments are closed.