Greater Greater Washington

Roads


Self-driving cars are coming, and they could change everything we know about cities

Autonomous, self-driving vehicles are getting more attention from the media, but little from transportation planners. Given the technology's potential impacts on our transportation network, it's time for planners to start thinking about it.


Photo by Dave Schumaker on Flickr.

As the technology advances, mainstream media now treat self-driving cars with seriousness and respect, as do business advisors like KPMG. Developers are designing self-driving car use into future retirement communities, while carmakers like Mercedes advertise passive "self-driving" safety features. Analysts predict that completely autonomous cars will be on sale by 2020.

Self-driving cars have the potential to reduce both car crashes and traffic congestion, and to use wasted time driving for work or entertainment. These are benefits usually attributed to transit; as a result, autonomous vehicles could strengthen arguments for designing for more cars in our cities and suburbs, instead of more pedestrians, cyclists, and placemaking.

Transportation planners aren't talking publicly about driverless cars

By contrast, online searches for "transportation planning" and "self-driving cars" turn up thoughtful, if skeptical reviews by urbanists Todd Litman and Jarrett Walker; a sober, academic summary of key issues by the Eno Center for Transportation; and a thorough debate from 2011 on the issue here on GGW; and an article from Governing Magazine that exemplifies the public preoccupation with regulating driverless cars rather than planning and policy issues.

There isn't a lot of evidence of transportation planners at public agencies giving serious attention to the matter, at least not publicly. A recent blog entry from Bacon's Rebellion also concludes that transportation planners are not paying attention. Though, to be fair, the topic was covered at a recent Florida Department of Transportation conference and the Transportation Research Board a few weeks ago.

Self-driving cars address many of the safety and travel efficiency objections that Smart Growth advocates often make about road expansion, or the use of limited street space. As a result, planners and placemaking advocates will need to step up their game.

They need to better define in what environments bike- and pedestrian-oriented designs are still appropriate even when we can solve our congestion problems with self-driving cars. They need to promote street and intersection that can work for bikes and pedestrians as well as for self-driving cars; and to make a strong cases for Smart Growth and TOD that are based on diverse benefits, not just on the ability to move people.

Capital planning decisions last for thirty years and beyond. The officials responsible for parking lot and garage building, transit system growth, bike lane construction, intersection expansions, sidewalk improvements, and road widenings need to analyze quantitatively how self-driving cars could affect their plans, and to prepare alternatives in case things change.

How could self-driving cars disrupt the planning process?

Here are two examples of situations where planners may need to adapt to self-driving cars:

Self-driving cars coupled with "smart intersections" that communicate with vehicles to let them pass without traditional stoplight timing could result in less congestion, but may speed up cars in places where cyclists and pedestrians are competing for space. The cars will be faster, but also safer to be around. The question is whether a more efficient auto network outweighs the negative impacts to other parts of the urban environment.

They may also make car use more competitive with bus transit in low-density settings and may erode the demand and need for transit (and paratransit). On the other hand, changed transit economics resulting from driverless buses could mean that extending transit into new areas will make more economic sense in the future than it makes today.

Ways to prepare for self-driving cars

So, what could the region's planners do now to anticipate the potentially sweeping changes that self-driving cars will cause? How can planners today insure that scarce infrastructure dollars are spent on things that might be less needed in the near future?

For example, if intersections can handle more vehicles per hour with self-driving cars than with human-driven cars, they may not need to be widened. Or if transit commuters can get to the station in a self-driving car, park-and-rides may not be necessary, because the car will just drive itself back home.

First, land use, highway, and transit planners should simply acknowledge the issue. They should begin to define how large different impacts may be, when those impacts are likely to occur, what the range of public responses will need to include, and when those public responses may have to start occurring.

Self-driving cars will change patterns of car ownership and travel. Planners need to examine how travel forecasting tools that are based on current patterns of car ownership and use will need to change to adapt to new statistical relationships between population, car ownership, trip-making, car-sharing, and travel patterns.

Because cars that can drive themselves won't stay parked all day, builders and regulators should think about how new parking structures should be designed for adaptive reuse if future parking demand declines.

State and local DOTs should measure how smart intersections could increase the number of vehicles that can use an intersection per hour, and how to design roads and intersections that work for self-driving cars, as well as pedestrians, bicyclists, and the creation of public spaces.

Finally, the region's transit agencies should study how driverless operations could affect operating costs for bus, rail, and paratransit services, and should update their long-range capital and operating needs forecasts to reflect what they learn.

Many aspects of the self-driving car world remain in doubt. That is not, however, a reason to avoid thinking about how to benefit from the capabilities that self-driving vehicles offer. Even if planners are only able to do general studies rather than detailed forecasts, that would still be a useful exercise. Understanding how to adapt our communities for the benefits and challenges of self-driving cars would be a huge step forward.

Nat Bottigheimer is a professional transportation planner and consultant with a background in public policy and real estate economics. Until 2012, he was an assistant general manager at WMATA, where he promoted bike and pedestrian access, sustainability, bus priority investment, TOD, data and information sharing, and more. He moved to Princeton, NJ in 2012, where his wife is an astrophysicist. 

Comments

Add a comment »

It's going to depend on what a "self-driving car" is. I assume that at first the cars will still require a licensed driver in the driver seat capable of taking over if the automated system fails. Eventually we may get to the point where you could put your frail grandmother or 12-year-old child in a self driving car and send them off.

I hope we figure out road tolling before then, because the increase in driving will quickly overwhelm the roads. If I could send our car to pick my wife up rather than drive there myself, I'd probably do it all the time. Of course, if everyone did that, it would take longer to pick her up than for her to just take the Metro.

by Michael Perkins on Feb 4, 2014 11:14 am • linkreport

Maybe a reason transportation folks aren't making a big deal about self driving cars is that walking isn't an inconvenience, it's the easyest way to stay healthy. That aside, if it's coming we should definatly plan for it. Is it an all in propositin, in other words, does everybody have to be in a self driven car for the system to work as advertised? Can they be hacked easily, and then who's to blame for crashes? What if you want to "take control" at certain points becasue driving can be fun, does it account for a sudden switch if you see a nice building by the side?

It sounds cool on the surface, but it breezes by such fundamental questions like, should we be encouraging this ever increasingly isolation that the car enables? Can we sustain the refuse of so many more disposable cars vs. the sustainability of upgrading fixed transit? I wouldn't bet on it, like google glasses, but I'm probably in the minority here. People love new gadgets.

http://bettercities.net/news-opinion/blogs/jay-walljasper/20873/wonder-drug-walking

by Thayer-D on Feb 4, 2014 11:15 am • linkreport

Self driving cars and computer assisted technologies could indeed change the way cities, roads, and our concept of cars drastically. The issue though is time. It will take a lot of time before self driving cars to fully arrive, tested and trusted. Longer still for them to be in sufficient numbers to actually make much of a difference and a generation for all of the gains to be realized.

Once they start to arrive in great numbers, local authorities could start restricting roads to their use to increase capacity, safety, and incentivize holdback users to convert. For instance make 66, self drive only during rush hour.

Still we are a long, long way away.

by Richard on Feb 4, 2014 11:18 am • linkreport

This is thoughtful. Especially contrasted with the way that most people get glassy eyed when they start talking about it. Then it turns out that people are so in love with the thought of autonomous cars that they demand all current transportation plans halt immediately.

The thing is, we still need to make our cities and communities pedestrian (and bike) friendly. Self driving cars can help but that won't erase the need for designing things in a way that considers all modes of transportation and prioritizes the ones that are safer and healthier.

So plan away, but lets continue to stick with what we know works.

by drumz on Feb 4, 2014 11:19 am • linkreport

"Self-driving cars have the potential to reduce both car crashes and traffic congestion"

No, they don't. Not unless every car on the road is 100% autonomous. Study after study of traffic congestion has concluded it takes only a single driver braking erratically to cause lengthy backups. Self driving cars may be able to help clear these backups slightly faster by moderating acceleration/braking. What all these rainbows and unicorns stories seem to forget is that automatic cars will be a luxury toy for rich people for many decades to come. And even when they become accessible to the majority of people, there will always been those who either don't trust or simply refuse to the use the technology.

I just don't see what problems autonomous vehicles can solve unless jurisdictions make certain routes/highways available to self-driving cars. And that will never happen.

by dcmike on Feb 4, 2014 11:20 am • linkreport

*last sentence should say "exclusively available to self-driving cars".

by dcmike on Feb 4, 2014 11:22 am • linkreport

well, the thing about ITS, and a driverless car is an element, is improving capacity. driverless cars + ITS may well just be a form of PRT. The point of "mass transit" is scale and capacity. If you double the capacity of a road lane because of driverless cars--something that will take some decades to happen--it will still be less efficient in a capacity sense, compared to mass transit.

And yes, the point of walking, biking, and transit is about sustainability, not just personal health, but has to do with how we use resources.

More cars, even if connected in an ITS grid, isn't better for cities.

You mention planning and driverless cars, and while I agree completely and note that the DC Transportation Element of the Comp. Plan doesn't even use the word "taxi" once, I think more importantly the point is to plan, land use wise and transportation wise, for people.

That would be more transformational. Right now, most of our planning is to enable the car at the expense of people, even in cities, where pedestrians (+ bikers + transit users) should be privileged and prioritized.

2. Similarly, wrt "transit" most agencies are not thinking of their planning function in terms of enabling competitive advantage and economic development, it's more about getting people from point a to point b.

by Richard Layman on Feb 4, 2014 11:25 am • linkreport

"Self-driving cars have the potential to reduce both car crashes and traffic congestion"
No, they don't. Not unless every car on the road is 100% autonomous.

Reduce, not eliminate. If you had 30-50% autonomous cars, they would start to reduce crashes and congestion. At 80-90% you would see even steeper reductions. Presumably at 100% they would nearly eliminate crashes and drastically reduce congestion.

As I said, if you regulate roads and limit or eliminate the non-autonomous cars on them you can get the benefits. After you get to 30% of the population of cars equipped, make all interstates autonomous drive only and drastically improve their safety and capacity.

by Richard on Feb 4, 2014 11:28 am • linkreport

I think there are a lot of good points made all around here. It's really an unexplored subject still because we don't know until it happens. That said, 75% of the workers are single drive alone commuters. About 10% of the population has severe disability that may affect their ability to easily transport themselves. On the otherhand, something like 90% of the day cars are parked and unused. Overarching land use patterns aren't going to change significantly in the next 20 years except in a few urban areas maybe. Even if the vast majority of the country isn't affected by self driving cars I could see a huge impact in Metro areas where millions of people only need cars a few hours a week at most. I use zipcar and car2go but there are severe limitations because you're restricted by supply and where you can leave a vehicle. Self driving vehicles could open that market up to people who don't benefit from the current paradigm but don't want to own and store a car or who don't have the ability to drive a rented car themselves legally or physically. Plus you could drastically reduce parking requirements in places like stadiums and city centers because cars can be stored in decentralized locations. With increasing fuel efficiency standards and alternative fuels you could possibly make the case for environmental gains without getting rid of cars. I think it can even be complimentary to heavy transit like rail because it solves the last mile problem pretty elegantly. Also if its sophisticated enough to drive itself I'm sure they could be set up as group taxi options with on demand routing.

by BTA on Feb 4, 2014 11:31 am • linkreport

While the points of this post are vague, it's a far better piece than previous posts on the topic. It does acknowledge that there would be upsides and downsides should robotic cars ever become a mass market product.

I commend the author for taking a more realistic view on the matter rather than the usual techno-triumphalist puff piece that claims that the new tech will be wonderful without the need for management and mitigation of unintended consequences.

by Cavan on Feb 4, 2014 11:33 am • linkreport

Self driving cars will presumably be limited to the posted speed. That will definitely change things on the Beltway.

by Crickey7 on Feb 4, 2014 11:34 am • linkreport

Self driving cars will very likely increase traffic for the following reasons:

1) More taxis running as self driving taxis become much cheaper and more in use.
2) People who cannot drive now (think kids and elderly, disabled) start using self driving cars
3) People drive to work downtown then send their self drive cars to park in cheaper locations for the day.
4) People will drive more as they can sleep , read, play on idevice while being transported.

by David J on Feb 4, 2014 11:34 am • linkreport

No, they don't.

They do. You either have potential or you don't. Even an elaborate and unlikely chain of events still qualifies as "potential."

by Another Nick on Feb 4, 2014 11:35 am • linkreport

@Richard, there is currently no projections that I'm aware of indicating self-diriving car penetration. Based on the expected high cost of purchasing one, the 30-50% fleet size you talk about is, as I said earlier, many decades off (if ever). Not to mention the average the US fleet now stands at 11.4 years and has been on a steady increase since 1995 (http://bit.ly/1bqT3P4).

It will never be politically popular enough to dedicate exclusive rights of way for autonomous vehicles when the barrier to entry is so high.

by dcmike on Feb 4, 2014 11:39 am • linkreport

I'm imagining all the trips and congestion being made today, and adding in to the mix additional driverless cars, empty, on their way to or from dropping someone off. It sounds like a nightmare.

We are trying to INCREASE the average auto occupancy. Driverless cars provide the opportunity for it to actually drop below 1. Bad idea jeans.

by recyclist on Feb 4, 2014 11:41 am • linkreport

Flying cars are coming too. Don't believe everything you read.

And given that WMATA is years off from driverless trains -- much less buses - much less cars -- can we focus on what we can do in the next 25 years?

I remember cars from the 1990s. They had computers are well. In terms of driver automation, we've got -- adaptive cruise control, automatic city braking, and some lane control.

I do agree that we may be living in a silver age of the streetcar/walking city, and that increased automation might disrupt that.

by charlie on Feb 4, 2014 11:41 am • linkreport

Has the liability issue on self-driving cars been settled yet? This seems to be a huge (yet kind of boring) issue that often gets left out of these articles.

If a self-driving car injures or kills someone, who is liable for the damages? Or is it going to take a lawsuit and a test case to figure that out?

by RP on Feb 4, 2014 11:44 am • linkreport

Probably be the final death of public transportation in the US at any rate. Especially fixed-guideway stuff like subways and commuter trains. Bye-bye, Metro and Amtrak.

by R. Stanford on Feb 4, 2014 11:45 am • linkreport

Don't know why people think this will hurt transit. How much more convenient will transit be when you can drive to the station, get out and your car goes back home instead of parking there?

by Crickey7 on Feb 4, 2014 11:49 am • linkreport

What percentage of our media coverage of automobilia talk makes sure it's fussed and gee-whizzed over like it's the Futurama exhibit at the 1939 World's Fair?

by KadeKo on Feb 4, 2014 11:51 am • linkreport

Transit still wins in terms of capacity. There's still a need for it in our densest places. Self driving cars are still cars that take up a specific amount of space and are bound by the laws of physics.

Moreover, if we're going to have pedestrian friendly spaces then it won't behoove a city to do things that would speed up traffic because the cars are all self driving all of a sudden.

by drumz on Feb 4, 2014 11:54 am • linkreport

From a bike/ped standpoint, I foresee these being a disaster. Streets will be designed by transportation (read "auto traffic") engineers to accommodate these cars and the greater amount of auto traffic throughput they'll bring, and that will inevitably mean minimizing any accommodations for those of us who'd rather not own a car at all. It could be argued this would not necessarily be the case, but come on…we all know this is how it would go down.

Adoption of self-driving cars will mean the current status quo of low-density greenfield development and the high priority on auto infrastructure (new highways, higher speed limits, more parking lots, more asphalt, more maintenance for roads) will remain. Just one more excuse not to revisit a 60 year policy of promoting suburban development over urban life.

by Simon Barnes on Feb 4, 2014 11:55 am • linkreport

There's no reason that driverless tech will mean the end of public transit - transit vehicles will likely have this technology well before enough of the US auto fleet has it to make a difference.

There are really two separate issues here:
1. Do we need to/how can we incorporate ITS tech into the transportation system to improve it?
2. What to do about driverless cars?

Transportation planners are not really considering #2 because it is probably WAY further off than most of the driverless techheads seem to think it is. Personally I think multiple decades until there is something mass-produced that can drive itself somewhere without a human inside.

The other issue with #2 is that driverless cars can't actually accomplish all the things people are touting. For example, how can you claim to both reduce traffic congestion while at the same time massively reducing parking requirements? You can't really massively reduce congestion on city streets - interstates sure - but if you add tons of VMT having cars drive all over the place empty to find parking then you are going to cancel that out.

#1 will be more interesting, especially now that USDOT has announced they are going to be requiring more ITS tech in cars. But again, on city streets you aren't going to see a huge improvement because you still have to rely on humans to operate the vehicles and not be doing other stuff. Just think how many times per week you see traffic held up because some dumb-dumb is looking at their cell phone when the light turns green. Now imagine your scenario where we do away with "traditional" light timing and the light can turn green at any moment - making this efficient relies on people to actually be paying attention.

by MLD on Feb 4, 2014 11:58 am • linkreport

I also think if done right self driving cars could be paired with pedestrian/bicycle robust infrastructure. You could potentially ban cars in the CBD because it would not be a serious constraint if your car could just drop you off on the edge and pick you up there later. Imagine if downtown became ped /bike/ transit only? It would also help with denser development because it would nullify a lot of the NIMBY arguments about street parking space needs. And garages would be more efficient when combined with fully automated parking systems which could save ~30% percent of the space (think the way rental companies have learned to pack cars in head to tail) that is used now for the traveling lanes.

by BTA on Feb 4, 2014 11:59 am • linkreport

Crickey7: "Don't know why people think this will hurt transit. How much more convenient will transit be when you can drive to the station, get out and your car goes back home instead of parking there?"

But how much more convenient would it be to just take your own car for the whole way rather than using transit?

When Britain cut many of its local trains in the 1950s in favor of a centralized hub-and-spoke model, the result was people decided it was easier to just drive the whole way, since the new trips would now involve 1). driving to the station, 2). taking a train, 3). and then having to secure other transportation on the other end. Before the cuts one could simply walk to the local station and take the train system all the way to the other end.

by 1DCampbell on Feb 4, 2014 12:04 pm • linkreport

R. Stanford:
"Probably be the final death of public transportation in the US at any rate. Especially fixed-guideway stuff like subways and commuter trains. Bye-bye, Metro and Amtrak."

Perhaps but not necessarily. There is considerable debate on this issue and nobody knows at this point what ownership of autonomous cars will look like.

On the one hand, some people claim that automated cars will serve as virtual chauffeurs, dropping the vehicle's owner off at work in the morning and then continuing on to go do errands throughout the day. In this case, vehicle ownership likely will not be significantly reduced.

On the other hand, Alain Kornhauser from Princeton U. argued that autonomous cars can serve as virtual taxis (http://www.advancedtransit.org/wp-content/uploads/2013/08/APO20-Alain-Kornhauser-TRB_AutomationWorkshop_Stanford_071513a.pdf), greatly reducing vehicle ownership rates and auto congestion. Kornhauser noted that currently the average vehicle occupancy is about 1.1 people per vehicle. With autonomous cars continuously picking up and dropping passengers off, Kornhauser suggested that it is possible to double the vehicle occupancy rate.

Personally, I think that the autonomous vehicle as taxi model can compliment and improve traditional transit in at least three ways:

1) These can serve as a replacement for very expensive para-transit (estimates of para-transit costs are $5 per mile per passenger) which are draining money from transit agencies.
2) Replacement for traditional buses on late-night, low-ridership routes (I've been on buses many times during off-peak hours when there are just 3-4 riders).
3) if vehicle occupancy doubles, the need to invest in more highway capacity will decline. This money can be redirected to more traditional transit (rail and bus).

by 202_cyclist on Feb 4, 2014 12:05 pm • linkreport

if your car could just drop you off on the edge and pick you up there later.
Pretty sure Metrorail does this already. Leave your car at the park and ride lot and take the train in to the city.

by dcmike on Feb 4, 2014 12:07 pm • linkreport

@Simon Barnes:
"From a bike/ped standpoint, I foresee these being a disaster. Streets will be designed by transportation (read "auto traffic") engineers to accommodate these cars and the greater amount of auto traffic throughput they'll bring, and that will inevitably mean minimizing any accommodations for those of us who'd rather not own a car at all."

There are more than 3000 annual pedestrian and bicyclist fatalities. The current vehicle technologies are currently not perfect either.

by 202_cyclist on Feb 4, 2014 12:10 pm • linkreport

Pretty sure Metrorail does this already. Leave your car at the park and ride lot and take the train in to the city.

I agree and that's how I use Metorrail but I'd say ~80% of even the regional core area isnt within a one mile walk of a station so it's a complimentary option for people that are already going to drive. Even DC proper i'd say only about 50-60% is sufficiently covered by Metro access.

by BTA on Feb 4, 2014 12:19 pm • linkreport

I think his point was that it's more that self-driving cars will give yet another reason to not reconsider the wisdom of the single occupant vehicle model.

On the other hand, self driving cars are unlikely to have little effect on rising fuel prices (any fuel they save from their ability to reduce traffic jams is likely to be offset by their greater use because of their reduced opportunity costs), so at least that pressure still remains.

by 1DCampbell on Feb 4, 2014 12:21 pm • linkreport

What about the sunk cost theory? Currently people are spending upwards of $500 a month just to own a car (including avg insurance and gas). I think a lot of people would rather pay $200 a month for periodic access. I know I would do that and a usage fee would actually make people consider each trip more thoughtfully. Do I pay $5 to drive to Silver Spring or spend $3 to take the Metro?

by BTA on Feb 4, 2014 12:28 pm • linkreport

"[...] even when we can solve our congestion problems with self-driving cars."

When you think about it, this is actually a totally outrageous statement, completely unsupported by any facts.

Even if we assume all the benefits of self-driving cars actually acrrue, they will never be able to solve the basic problem of too many drivers occupying too little road space. Self-driving cannot overcome the basic laws of physics and won't overcome induced demand either.

For example: How many drivers cross the 14th St Bridge into and out of the District everyday? Even if you assume that all such cars will at one point be self-driving and that results in, say, a 50% increase in the number of vehicles able to cross in the same time period, it will never "solve" congestion on the 14th St Bridge.

What happens when the bridge becomes faster to commute on, and more drivers who otherwise would have avoided the bridge now choose it instead of an alternate route or taking transit? What happens in 20 or 30 years when the DC metro has hundreds of thousands of new residents - population growth and migrants - trying to commute on the same infrastructure?

Self-driving cars will almost assuredly help with reducing parking demand and may allow more frequent, faster and possibly cheaper paratransit services like Metro Access. But they will never "solve" congestion, especially in large metro areas like DC. If anything, they may even exacerbate some congestion in that a self-driving car returning to its depot or making another trip adds an additional vehicle to the road network where before it would have remained parked. They may help parking congestion, but will never solve road congestion.

It is all well and good to look ahead and plan for semmingly inevitable technological change. But in so doing, we must not become enraptured by the latest technological fad and surrender our planning to dubious claims arising therefrom.

We should remember the example set by planners of the mid 20th century, who were also enamored of dubiuos claims of the benefits of automobile travel. We are still living with the huge mistakes they made. Let's hope we don't make a huge set of new mistakes that saddles the next generation with similar problems.

by ndw_dc on Feb 4, 2014 12:29 pm • linkreport

I anticipate driverless cars being paired with car sharing systems to reduce the overall number of cars, and commensurate need for car storage. It'll kill the taxi industry, for sure.

by Crickey7 on Feb 4, 2014 12:31 pm • linkreport

I'm a big fan of the potential of self-driving cars, but I don't see them reducing congestion until WAY late in the game.

And how that will happen is that, AFTER driverless cars are legal, and we can get a Zipcar to come to our door from a lot a few miles away, we will get to a point where we have and use only as much car as we really need.

And that will trigger the development (and eventual wide use) of cars that are only one person wide.

Cutting the width of cars in half will be the point at which self-driving cars substantially ease congestion, because they'll turn what are now 4-lane roads into 8-lane roads without any widening.

But due to the longevity of cars if nothing else (says the guy driving a 2000 Accord with 214K miles on it that shows no signs of giving up), I can't see this step happening any earlier than the 2040s, even if (as I devoutly hope) my 6 year old gets a self-driving car for his 16th birthday.

by low-tech cyclist on Feb 4, 2014 12:37 pm • linkreport

@Richard, there is currently no projections that I'm aware of indicating self-diriving car penetration. Based on the expected high cost of purchasing one, the 30-50% fleet size you talk about is, as I said earlier, many decades off (if ever). Not to mention the average the US fleet now stands at 11.4 years and has been on a steady increase since 1995

My previous post did mention generations. I dont know if we should change our infrastructure investment now on things that may come to pass closer to the 22nd century than the 20th.

by Richard on Feb 4, 2014 12:47 pm • linkreport

Some of these gloom and doom comments are hilarious. First of all, legally, I believe there will be some requirement to have a human being of age in the car at all times in transit. What happens if you let your car drive your kids to school and it breaks down or malfunctions while it’s looking for that cheap parking space? You’ll get an alert but it doesn’t help that you don’t have a way of getting to your car unless you driverless-Uber it or catch an available ride-share.

I do believe congestion will be reduced (even if there are more cars) due to the fact there will be significantly less accidents. I also believe transit can still be attractive through dedicated lanes and potentially comfort amenities that rival sitting in your car. If you prove to people transit is faster, cheaper, and still comfortable, they will take it.

by Lane on Feb 4, 2014 1:19 pm • linkreport

@Lane:
I do believe congestion will be reduced (even if there are more cars) due to the fact there will be significantly less accidents.
Do you really think that accidents are the primary cause of congestion in the DC area? Most days, rush hour is accident-free on I-66 inside the Beltway, or I-395, or on 16th Street. Yet somehow, these come to a grinding halt every single day.

by Gray on Feb 4, 2014 1:34 pm • linkreport

I haven't read all the comments so apologies if this was already mentioned. Perhaps we can start transitioning to self-driving public transportation. Imagine a Metro that runs automatically like the train in Denver's airport. Decreasing the need for a human to drive every bus and train in the DC-metro area could save WMATA substantial expenses.

by Scott on Feb 4, 2014 1:44 pm • linkreport

@ndw_dc: "Even if we assume all the benefits of self-driving cars actually acrrue, they will never be able to solve the basic problem of too many drivers occupying too little road space. Self-driving cannot overcome the basic laws of physics and won't overcome induced demand either."

One solution to this is dynamic pricing. If you get charged tolls depending on how far you go, the toll price could vary depending on how busy the road is. You always price it so that the road is at capacity, and not beyond it.

This is imperfect, though - now people who are less wealthy aren't going to be able to travel at peak times. Or they take the bus, etc.

But, we already put a price on health and education so why shouldnt transport be any different?

by Steve on Feb 4, 2014 1:50 pm • linkreport

And given that WMATA is years off from driverless trains -- much less buses - much less cars -- can we focus on what we can do in the next 25 years?

WMATA is years from automatic control of trains with a driver in the seat, much less a driverless train.

While I'm skeptical that the legal and psychological issues of driverless cars can easily be surmounted (I don't question the technical feasibility), I also think the impact of true driverless cars will be greater than many realize. Some impacts:

* Driverless cars won't be the end of public transit. They will be public transit. Few people will own cars and certainly not more than one. When you're ready to go somewhere, you call upon a car stored at a nearby depot to come pick you up and take you wherever you need to go. You only use your personal car when a public one is unavailable.

* Lots of trips won't have any passengers at all. Why go to the grocery store for milk when you can send your car to the grocery depot, a worker will put groceries into the car for you, and the car will drop them back off to you.

by Falls Church on Feb 4, 2014 2:00 pm • linkreport

I think we'll see driverless cars on highways first, as a natural extension of passive automation for cruise control, etc.

By the time the cars are smart enough and we're habituated to them, I think they'll be an extension of a carsharing or taxi service.

by Neil Flanagan on Feb 4, 2014 2:08 pm • linkreport

@Gray

Traffic accidents may not be the primary cause of congestion but definitely up there in regards to the DC area. All of the roads you mentioned either reduces in the number of lanes and/or drops traffic abruptly into downtown. I'm not sure if you're going to see much efficiency on those roads during rush hour (esp. with more cars). However, roads like the beltway, 95, 295, GW Pkwy, etc. would most likely see reduced congestion with minimal to no accidents.

by Lane on Feb 4, 2014 2:20 pm • linkreport

Has the liability issue on self-driving cars been settled yet? This seems to be a huge (yet kind of boring) issue that often gets left out of these articles.
If a self-driving car injures or kills someone, who is liable for the damages? Or is it going to take a lawsuit and a test case to figure that out?

Liability will be dealt with the way it is now: insurance.

If everyone is required to have insurance, it doesn't really matter where the liability lies, at the end of the day it's the insurance companies that pay. Presumably, self-driving cars won't be allowed (or won't be insurable) until they are significantly safer than human-driven cars. At that point they will be a boon for insurance companies -- significantly safer means significantly lower payouts. Even insurers of human-driven cars will benefit.

There will be a good opportunity to reset the whole industry, eliminate two inequities in the current system. The first is that many of the costs of automobile usage are passed onto health insurance companies. Since the largest health insurer is the federal government these costs are passed onto the taxpayer, creating a negative externality for car users. The second is that there are some really, really bad drivers out there, and as a society we have made the decision that everyone can drive. If those really, really bad drivers were forced to pay the expected cost of their driving through insurance it would be unaffordable for them, so society subsidizes their insurance through rate regulation and assigned risk pools. Once self-driving cars are available such policies would no longer make sense. I would expect the worst drivers to be among the earliest adopters of self-driving technology.

by contrarian on Feb 4, 2014 2:35 pm • linkreport

you can send your car to the grocery depot, a worker will put groceries into the car for you, and the car will drop them back off to you.
Jeff Bezos will have his drones bombing groceries on your porch years before this ever becomes reality. Besides, we already have Peapod + company doing this today with far greater efficiency.

Autonomous cars are a solution looking for a problem. Investments in people-moving infrastructure are best spent on (public) mass transit. Leave the delivery of goods and services to private industry where economies of scale and competition will drive efficiency.

by dcmike on Feb 4, 2014 2:38 pm • linkreport

I might expect the worst drivers to be the ones that enjoy the risk-taking and agressiveness that comes from driving, so they may be late adopters.

Perhaps the earliest will be the elderly drivers we should be taking the keys away from under advice of their doctors or family members.

by Michael Perkins on Feb 4, 2014 2:44 pm • linkreport

Falls Church and low-tech cyclist are thinking about this in the right way. When driverless vehicles arrive they may start out as the same vehicles we have now but they will quickly transition to novel vehicle types geared toward specific uses.

As it stands, we use our vehicles for a multitude of tasks (commuting, shopping, transporting kids, road trips, etc.). So we choose vehicles that can fulfill all those needs. Once vehicles can drive themselves and they can be rented on demand like Uber then it makes sense to have different vehicles for different tasks. A single occupancy vehicle for commuting, vehicles with storage for shopping, vehicles with entertainment systems and beds for road trips and intercity travel, etc.

Once the systems for driverless vehicles are created there will be few barriers to this kind of specialized vehicle paradigm. Driverless vehicles are the invention of a new class of transportation like the introduction of trains, automobiles, or planes not just an iteration of what we already have.

by meegles on Feb 4, 2014 3:00 pm • linkreport

Great. After nearly a century of work engineering the landscape for the sole benefit of cars, we can finally let them have the run of the place. Cue "Twilight Zone" clip of the robot that replaces Mr. Whipple.

by tdballo on Feb 4, 2014 3:03 pm • linkreport

You won't buy a car. You will buy a car service with a combination of up-front fee and per-use charges. You will order car of specified type to be at your door at a given time. When you are done, it will go into a storage space calculated as most convenient to its next expected use. When it gets low on gas, it will take itself to be filled and when it needs service, it will go to a service depot. Anything that breaks will register on a central computer and produce an appropriate response, including dispatching another car if one breaks down.

Upper middle class families will keep one car "for emergencies", but most won't bother.

by Crickey7 on Feb 4, 2014 3:17 pm • linkreport

I have no coherent thoughts on the matter, but a couple of - probably not very novel - observations:

1. Carshare/driverless taxi/rental: The terms become somewhat interchangeable, but seems me that whatever you call it, it would be a big and obvious phenomenon once practical vehicles are available.

After all, aside from storing old fast food condiment packets in the center console, the big advantage of your *own* car is mostly just that it's theoretically nearby and relatively hassle free when you need it. No need to go track down a car share, return it, contend with other users, surly taxi drivers, etc. But individually-owned cars are obviously really inefficient: most of the day and night, most cars are just expensively sitting somewhere idle, taking up space. Assuming there's a fleet of self-driving cars autonomously patrolling the streets for clients, than you'd just be an app tap away from a car anywhere you go - possibly even out into lower density areas that aren't currently easily served by taxis or car sharing. Why would most people own a car at all?

2. Parking: Car sharing as above could help reclaim a lot of wasteful parking space - in principle, if the management software is pretty good, then many such vehicles might never really park. Just get routed from pickup to dropoff and immediately to pickup again.

Privately owned cars could be routed back to their home garage during the day, or a parking location not necessarily right next to the owner's destination. Obviously that has some implications for energy use/occupancy ratio/congestion.

Either way, it affects place design: maybe the 'convenience' ideal is no longer a Walmart type box with a gigantic desert of pavement in front (or below) to store your car, it's maybe more along the lines of a fancy hotel with a small circular driveway to facilitate drop offs from chauffeured clients - but such facilities might still not necessarily be ideal for pedestrians without deliberate thought.

3. Congestion: Obviously a roadway of 100% automated vehicles could be, at least theoretically, very efficient, but there might be tipping points where some benefits could be seen even at adoption rates of less than 100%. What happens to grid lock if just, say, 10% of the cars on the road resolutely obey traffic signals, and refuse to box block? Or what happens to the weird wave-like patterns of spontaneous freeway congestion if the road is salted with vehicles which have a different reaction profile - or are even programmed to deliberately break up harmful patterns? (By, e.g., driving smoothly at the average speed of the road, rather than stop/go/stop/go.)

Some of these interact: autonomous vehicle sharing services could accelerate the use/adoption/vehicles-on-the-road rates far above what you might project if you assume a mostly operator-owned fleet that needs to turn over a couple times. So the urban design/transportation planning issues might sneak up on us rapidly. Low hanging tipping points in terms of road behavior might be reached by just one or two successful services.

by jack lecou on Feb 4, 2014 3:33 pm • linkreport

I think ultimately public transit will evolve as much as cars - if not more so. I imagine automated vans, electric quadrupeds or other light taxi-type vehicles picking you up from your home or origin point, transferring you to a node on a corridor effectively solving the "last mile problem." Once at the node, you are then you are picked up by a high capacity, high speed transit system. On one end of the spectrum, this would be a future automated version of Bus Rapid Transit or Light Rail on dedicated corridors. At the other end of the spectrum would be a high speed rail network, MAGLEV or Hyperloop operating on the reclaimed right of way from the existing interstate highways. Private automobiles are the real dinosaurs in this scenario. Or, we could just have a rebirth of 1950's sprawl with epic commutes from the exurbs enabled by automated land yachts, but I think enough trends outside of transportation are pushing against that outcome.

by Ryan on Feb 4, 2014 3:34 pm • linkreport

With the possibility of a mixed fleet consisting of connected, semi-autonomous, and human driven cars perceived to exist in less than 10 years, it is time the planners reevaluate the assumptions on how much more volume can be added to existing capacity. This could mean that the 10 year and 20 year forecasts may need to evaluate how much more we can do with existing capacity as adjusted for mode splits.

Another piece of this puzzle is that on the fully autonomous vehicle side of things, the test vehicles today need a high degree of infrastructure "fidelity" to identify signs, pavement markings, traffic signals, etc. in addition to potential hazards potholes, and conflicts w/ pedestrians, bicycles, etc. --- in this context "fidelity" means good repair...and we are a long way from that in most urban areas, sadly

by Some Ideas on Feb 4, 2014 3:46 pm • linkreport

You won't buy a car. You will buy a car service with a combination of up-front fee and per-use charges. You will order car of specified type to be at your door at a given time. When you are done, it will go into a storage space calculated as most convenient to its next expected use. When it gets low on gas, it will take itself to be filled and when it needs service, it will go to a service depot. Anything that breaks will register on a central computer and produce an appropriate response, including dispatching another car if one breaks down.

This seems like pie-in-the-sky. I would love for it to be the case, because it would probably mean more people want to live closer to good transit service so they can get to work easily. But if someone drives to work every day they're still just going to "rent" a car every day? I don't see it.

by MLD on Feb 4, 2014 3:51 pm • linkreport

We'll see Super Bowl ads from car service companies, not car companies.

by Crickey7 on Feb 4, 2014 3:54 pm • linkreport

But if someone drives to work every day they're still just going to "rent" a car every day? I don't see it.

Maybe.

Urban real estate will still be expensive (read: parking will still be expensive). Automation will lower the operating costs for taxis; taxi rates should fall.

Consider it a mix between Uber and Car2go, driven by a computer.

It certainly has the potential to lower car ownership within the city (where not all trips rely on a car). Maybe not for a completely car-dependent place, but you never know.

by Alex B. on Feb 4, 2014 3:59 pm • linkreport

Everyone is thinking of driverless cars but the same technology can be used for other sorts of vehicles. How about driverless electric bikes or segways? That could be the future of Cabi because getting a ride in a driverless car would be a lot more expensive than a ride on a driverless electric bike/segway.

by Falls Church on Feb 4, 2014 4:01 pm • linkreport

But if someone drives to work every day they're still just going to "rent" a car every day? I don't see it.

I think many people aspire to be chauffeur driven to work every day.

by Falls Church on Feb 4, 2014 4:04 pm • linkreport

@Falls Church
I think many people aspire to be chauffeur driven to work every day.

Well yeah, but if your use is every day at peak period I don't see how owning is going to be that much more expensive than a peak-period rental every day.

@alex B
It certainly has the potential to lower car ownership within the city (where not all trips rely on a car). Maybe not for a completely car-dependent place, but you never know.

Agreed.

@Some Ideas
Another piece of this puzzle is that on the fully autonomous vehicle side of things, the test vehicles today need a high degree of infrastructure "fidelity" to identify signs, pavement markings, traffic signals, etc. in addition to potential hazards potholes, and conflicts w/ pedestrians, bicycles, etc. --- in this context "fidelity" means good repair...and we are a long way from that in most urban areas, sadly

For many of the predictions to pan out, it's going to require cars to be connected to a central system that at minimum doles out information about light phases, speed limits (though that could be stored) etc. The Google car model isn't going to cut it.

by MLD on Feb 4, 2014 4:17 pm • linkreport

It could even be as simple as dictating when a car picks you up whether or not you are willing to split the ride and presumably therefore the cost.

by BTA on Feb 4, 2014 4:30 pm • linkreport

@ BTA:

I like that. An option when you rent to share and split costs, at some modest additional time.

by Crickey7 on Feb 4, 2014 4:52 pm • linkreport

"They may also make car use more competitive with bus transit in low-density settings and may erode the demand and need for transit (and paratransit)."

Wont happen until self-driving cars no longer require a person who can drive in the drivers seat. Until you can stick a blind person in the drivers seat with no other person in the car all you have is auto-drive and cruise-control for cars

When that day truly arrives which wont be in the next 10 years it better not be like Johnny Cabs; it should be a vehicle with no drivers seat at all and the first use should be on public transit.

by kk on Feb 4, 2014 5:37 pm • linkreport

Great article - and a timely reminder of how transformational automated vehicle technology will be.

Some of us have been seeking to raise awareness of the need for planners to take much more notice of automated vehicles. Here is one example of an article that we did for Ontario: http://ogra.uberflip.com/i/193788 (From p.25)

I personally have been trying to shake up the Institute of Transportation Engineers and they have now formed a committee that includes automated vehicles, although it is unclear if any policies have been formulated yet.

This article from October 2012 elicited a single response form the planning profession: http://clearlightpr.com/wp-content/uploads/2013/04/Driverless-article.pdf

Even now all levels of government are making decisions involving billions of dollars around transportation and infrastructure, with no cognisance of the potential impacts of automated vehicles.

I appreciate articles like this that draw our attention to this very important subject - in which all taxpayers are stakeholders

by Paul Godsmark on Feb 4, 2014 6:46 pm • linkreport

There still will be accidents. There still will be road repair. Snow removal. Vehicles with human drivers making unpredictable moves. A thousand monkey wrenches for software to deal with. Unless, as some articles I've read are suggesting, making driving illegal. So everything besides cars, from delivery trucks to snow plows to tow trucks to ambulances/fire/police, construction vehicles...all self driving? Self driving cars are a utopia being hyped up mainly by people looking to become rich, cloaked in "environmental responsibility" and "safety".

by Bob See on Feb 4, 2014 7:55 pm • linkreport

I agree with the primary premise of this article: that planners need to incorporate autonomous vehicle scenarios into their work -- certainly for medium and long range planning.
Big issues include the transition. It will likely take some time before fully autonomous vehicles that function on all roadways and that are affordable appear. But partially autonomous vehicles appear likely quite soon, although they may limit themselves to expressways and major arterials at first.
At some point there will be a significant positive impact on congestion, in part due to fewer accidents. but mostly to reduced headways between vehicles. Reduced congestion will improve overall urban accessibility, with important economic and social impacts.

by Richard Mudge on Feb 4, 2014 9:40 pm • linkreport

The problem with self-driving cars is that it is mainly discussed by tech enthusiasts who care more for their theoretical limits than for the practical considerations. Computer-driven cars are virtually guaranteed at this point, and they seem likely to be commercially available within a decade or so. Driverless cars are a different matter--and there is a distinction between the two.

You can already derive some benefits (like light timing) with today's technology: detecting presence and speed of cars is not a difficult task; "smart" cars aren't necessary. More fundamental changes are however predicated on the notion that driverless cars are ubiquitous. That won't happen for 25 years at least, which puts it beyond any useful planning horizon at the moment.

by Joshua Cranmer on Feb 4, 2014 9:59 pm • linkreport

There still will be accidents. There still will be road repair. Snow removal. Vehicles with human drivers making unpredictable moves. A thousand monkey wrenches for software to deal with. Unless, as some articles I've read are suggesting, making driving illegal. So everything besides cars, from delivery trucks to snow plows to tow trucks to ambulances/fire/police, construction vehicles...all self driving? Self driving cars are a utopia being hyped up mainly by people looking to become rich, cloaked in "environmental responsibility" and "safety".

They don't have to be perfect, just better than the average human. It isn't that high a bar, about 40,000 people are killed by drivers every year. There are lots of tasks that computers do better than people.

by contrarian on Feb 4, 2014 10:00 pm • linkreport

Perhaps a decent comparison for self driving cars is positive train control. PTC deals with trains, and there are fewer trains than cars, the trains have to stay on the tracks and not change lanes and what not, and are generally a whole lot simpler of an intellectual problem than cars. PTC is really expensive to install. Why should self driving cars be cheaper?

by anonomouse on Feb 4, 2014 10:05 pm • linkreport

@MLD "For many of the predictions to pan out, it's going to require cars to be connected to a central system that at minimum doles out information about light phases, speed limits (though that could be stored) etc. The Google car model isn't going to cut it."

Audi already is doing that...test location in Las Vegas and they have already asked Pasadena, CA to receive the real-time traffic signal controller information...This stuff is real.

by Some Ideas on Feb 4, 2014 10:16 pm • linkreport

I'm always curious: do the current self-driving cars obey the posted speed limit?

I'm not sure that I can see widely-deployed robot cars doing more than the speed limit, not even if instructed to by their human occupants. The vehicle knows the posted limit, and allowing speeding becomes a liability problem for the manufacturer.

by David R. on Feb 4, 2014 10:48 pm • linkreport

It's true, the transportation and urban planners have not been thinking about this very much. While nobody can tell them with certainty what the future will be, I am quite certain that they will be wrong if they assume it's going to look very much like it looks today.

Several years ago I put out the challenge to planners to examine the potential new rules created by the cars. This can be found at http://robocars.com/urban-plan.html

In addition, a bit of an answer, doing what I call "Robocar Oriented Development" can be found at http://robocars.com/rod.html

by Brad Templeton on Feb 5, 2014 4:55 am • linkreport

Great to see this discussion continue!

A note about how we'll have the rapid acceptance of autonomous vehicles that I, and others predict. Once a safer alternative to poor drivers becomes available, we simply won't accept the kinds of tragedies we've been trained to view as just a matter of course for use of the roadways. It will start with the chronically disabled as a means of giving them mobility they've not enjoyed previously, then it will quickly move to seniors who really shouldn't be driving, but do anyway (my grandfather kept driving even after he lost his night vision ability, he just would make a point of getting off the road before dark). Judges will get more cavalier about taking away people's driver's licenses, especially if alcohol or repeated poor judgement is involved, knowing that removing a license isn't simultaneously removing all reasonable mobility options. The wealthy will of course adopt these immediately as a toy, but it will quickly be adopted by the upper middle class as a way to squeeze more efficiency out of their limited time.

Taking up the issue of taxis, I can't help but note that the professional corps of taxi drivers appears to be the group I most frequently see committing the worst driving infractions. While we can anticipate some protests, I have no doubt that companies offering AV livery service will quickly overwhelm "manned" taxis.

I understand the skepticism about how long these will take to saturate the transportation network, but you have to anticipate how popular and efficient this technology will be. Add in that driving truly is a matter of life and death, and if AVs can significantly beat the performance of humans - not a high bar - and we'll see fleet turnover faster than most imagine.

by Will Handsfield on Feb 5, 2014 5:15 am • linkreport

The only thing worse than self-driving cars is flying cars and the only thing worse than that is self-flying cars. Why does our kind do these outrageous, wasteful, and very dangerous things simply because we can? How can we hope to move forward on all the issues that face us when we still do huge things like this with no thought whatsoever about the outcomes. This is not green, not safe, and not needed by anyone including billionaires. Cars are defined as lethal weapons as they have been used as such for generations. Robots running lethal weapons at very high speeds is simply unconscionable.

by AndrewJ on Feb 5, 2014 6:44 am • linkreport

Please note that some of the posters who have recently joined this discussion thread are promoters and advocates. Good for this audience to hear what they say, but... They may be few but they are loud and insistent that there world view will come to pass in short order.

...take with a grain of salt or Alka Seltzer as is your preference.

by Some Ideas on Feb 5, 2014 9:10 am • linkreport

"Why does our kind do these outrageous, wasteful, and very dangerous things simply because we can?"

That question seems to be poping up more and more from people who can't be simply dismissed as luddites. From an evolutionary perspective, our craving for novelty is strongly associated with survival as new tools and landscapes improved our chances, so we do these things becasue we're wired to.

But when technology begins to outstrips our biology and ecology, which has been happening on and off for the last 50 years, we need to zoom out and look at the larger picture to limit the law of unintended consequenses. In this case, as neat as it sounds, there are clearly too many negatives as noted by others here that this technology dosen't take into account. Sure there will be applications where elements of this are useful, but our ability to split the atom ad infinitum quicky becomes myopic if not balanced with societie's larger goals.

Fortunatley we are coming to realize that just becasue we can do something technologically, dosen't necessarily mean our quality of life will improve. Would this cultural shoft signify the end of progress? Not if we allow the definition of progress to encompass a broader view of the sciences and humanities.

by Thayer-D on Feb 5, 2014 9:10 am • linkreport

And we haven't even touched on the privacy issues.

by Crickey7 on Feb 5, 2014 9:15 am • linkreport

@Thayer-D

You are starting to get at why planners "aren't thinking about" the future world filled with driverless vehicles.

Forward-thinking planners are thinking about issues like global warming, CO2 and other emissions and their health consequences, the improved health outcomes of alternative transportation (bike/walk/transit) and more efficiently moving people in the restricted city environment. They are thinking about raising fuel taxes to levels closer to those in Europe, both because we actually need more money to pay for the transportation infrastructure we have, and to pay for the damaging effects of emissions.

The world envisioned by driverless car advocates is the antithesis of the city environment. It is "suburbs everywhere" though maybe with less parking.

by MLD on Feb 5, 2014 9:24 am • linkreport

One thing I want to comment on from the original article:

Self-driving cars will change patterns of car ownership and travel. Planners need to examine how travel forecasting tools that are based on current patterns of car ownership and use will need to change to adapt to new statistical relationships between population, car ownership, trip-making, car-sharing, and travel patterns.

We should note that our current methods for forecasting travel are crap. They're bad pseudo-science, the kind of forecasting that lacks precision and accuracy, lacks any after-the-fact corrections, and worst of all, gets codified into our regulations (via rules like minimum parking requirements in zoning codes) that make the perdicted traffic more of a self-fulfilling prophecy than anything else.

This is not robust forecasting, it never has been. Yes, we need to adjust our thinking for a future with autonomous cars, but we also need to be realistic about what our current forecasts can and cannot tell us; what the margin of error is in those forecasts (hint: it's huge).

Things like this are not promising: http://www.ssti.us/2013/12/new-travel-demand-projections-are-due-from-u-s-dot-will-they-be-accurate-this-time/

by Alex B. on Feb 5, 2014 9:38 am • linkreport

"The world envisioned by driverless car advocates is the antithesis of the city environment. It is "suburbs everywhere" though maybe with less parking."

Yeah, that's fairly accurate. One thing that's never addressed is how all these robocars will be powered or any of the negative environmental externalities associated with them. Call them "Auto-Determinists". They love suburbia as it is, love the status quo, hate mass transit, and are latching on to new technology as a way to keep from thinking about the real problems with single-occupant private transportation.

by RandG on Feb 5, 2014 9:55 am • linkreport

I can just see this robocar paradise. Ultra-low-density development, Broadacre City style! No sidewalks, all single family homes and 1950's style shopping malls, robocars flitting around all day wasting energy, more land committed to parking lots, highways, and above all, more auto infrastructure, because God knows we haven't turned over enough of our landscape to the almighty car. Truly the vision of the future for 1962.

by Barton S. on Feb 5, 2014 10:04 am • linkreport

I don't get the gloom and doom. I have no idea why some predict it will enourage greater suburbia.

Millenials are trending away from car ownership because the act of owning a carhas lost its importance. The increasingly consume transportation based upon the merits at any given time of any given mode. They may arrive by robocar and leave by Citibike. Arrive in 3 cars and leave in one. Travel by subway to the first ring of suburbs, then break up into 3 robocars. With information streaming back and forth to the routing service, they can be flexible about sharing rides with strangers if the result is a lowered cost--like a jitney service.

Without having the baggage of a vehicle, their thinking about "car" will change to "getting from A to B by the best means available."

by Crickey7 on Feb 5, 2014 10:12 am • linkreport

Yeah, I think RandG is certainly right: there is a segment - probably largely coincident with anyone who's advocating driverless cars (as opposed to merely noting their possibility) - who are really hoping they'll somehow help preserve the increasingly untenable and undesirable suburban status quo.

To the extent that infrastructure and planning is a political process, and not purely driven by abstract technical capabilities and constraints, the existence of that is a negative influence.

That said, I think there are some abstract qualities of driverless cars that would tend to ameliorate some of the worst aspects of suburbia - like reducing or eliminating the need for landscape destroying multi-lane freeways and inhumanly vast surface parking lots.

I mean, I wouldn't say I'm exactly sanguine about it. On it's own, driverless cars wouldn't do anything about anyone's desire to build acres of flimsy mcmansions. And energy use in a scenario where it's all driverless cars everywhere is probably not so great. But at least those housing developments and commercial centers might not be wedged so far apart by quite so many square miles of asphalt-scape. And shopping centers and civic places might be allowed to resemble something built for human beings, not automobiles.

It's certainly not an anti-suburban technology, but neither is it completely supportive of all the aspects of suburbia as we know and loathe today.

by jack lecou on Feb 5, 2014 10:39 am • linkreport

There are some other advantages. In sort of the same principle of high speed rail, enclosed highway tunnels could be built that operate at speeds well above which humans can safely drive and could theoretically reduce the need for surface highway lanes. Another topic we haven't talked about is trucks which have a high potential to cause serious injury or death in accidents. Without the need for human drivers you could have large fleets that mainly operate when traffic is lightest easing congestion and improving safety.

by BTA on Feb 5, 2014 10:45 am • linkreport

Computer-driven cars are virtually guaranteed at this point, and they seem likely to be commercially available within a decade or so. Driverless cars are a different matter--and there is a distinction between the two.

A legal distinction, possibly, at least initially. But the practical distinction seems less clear to me.

The way I see it, an autodrive system either works, or it doesn't*.

If you get one that actually works without trying to crash every 5 minutes, let's face it - people are going to tune out pretty quick. Human beings barely pay attention while driving as it is, so telling them they don't actually have to touch the wheel anymore, but asking them to be alert and diligently supervise the whole time? I give that about a week. Autodrive week 2: 90% of the cars in the morning commute are going to have newspapers or shaving mirrors or whatever obscuring half the windshield.

I mean, that is basically the whole selling point of an autodrive in the first place.

So, yeah, maybe the first systems are crude novelties - not much more than better cruise control, and still need to be actively supervised. Systems that don't work, basically. But expensive+crude=probably not so much market penetration. Such a system inherently wouldn't have any dramatic impact on either driving patterns or infrastructure planning.

Once you eventually have a system that does actually drive the car though? Even if the law still requires a human to sit at the wheel at that point, it's going to quickly become obvious that most of the 'drivers' behind the wheel are making themselves pretty superfluous. I suspect won't take that long for interested lobbies to get the law changed to reflect the de facto situation. At which point, well, here comes johnny cab.

----
* Allowing that It might work initially only in, say, a highway mode but not an urban mode - but that doesn't affect the argument for travel within that mode.

by jack lecou on Feb 5, 2014 11:08 am • linkreport

re "liability issues": What are the liability issues exactly? I think current mechanisms in our legal system should be able to handle this just fine: driver-fault torts and negligence, and product liability laws such as negligence and strict liability for design and manufacturing flaws. These should be, for the most part, applicable when it comes to automated cars without problems.

re fear-mongering: I don't see the value of pointing fingers and calling proponents of this technology suburbanites trying to perpetuate the status quo. No evidence of this, and exactly why planners should be constructive about this and involved in the discussion instead of getting political. "Us versus them?" I don' think so. It's a technology that will be as useful or harmful as we make it.

And it's a technology that's here now. It's already passively present in many cars and being slowly phased in, not later, but now, so we do have to deal with it somehow.

by sorelo on Feb 5, 2014 12:13 pm • linkreport

In sort of the same principle of high speed rail, enclosed highway tunnels could be built that operate at speeds well above which humans can safely drive and could theoretically reduce the need for surface highway lanes.

This seems problematic to me.

If driverless cars have an advantage from a public benefit perspective, it'd be their hypothetical ability to make more efficient/safer use of our existing infrastructure. I.e., carry at least marginally more people, more conveniently, with less congestion using nothing more than the already built surface grid (maybe allowing for a little special signaling or marking infrastructure).

The minute you start talking about building expensive new tunnels, for cars, you've lost that. Why not build trains in those tunnels instead? It also all starts to take on a suspicious PRT-esque flavor.

Existing grid or it's not worth it.

by jack lecou on Feb 5, 2014 1:14 pm • linkreport

In reading these comments it seems that many commenters are "stuck" in current thinking.

Here are some questions to ponder -

We have 250M cars/light trucks in the US now. When self-driving cars become a reality how many will we have? how many would we need?

Will we still need all the highway lanes we have now? Could some be devoted to trains or bicycles?

Will we have more "no-car zones"?

What will these new "cars" look like? Will a one-seater become common?

Will most/all of these cars be electric?

If so, do we need all of current gas stations? even if we keep some do they need to be on primary real estate?

What happens to car companies? Do BMW and Ford turn into "transportation service" companies? Meaning they help you get from Point A to Point B.
Same for Car-Rental companies?

What will happen to garages/driveways?

Will self-driving cars help or hurt mass transit?

What will happen to strip malls? how about shopping malls?

Do we dig up most of the parking lots around businesses, grocery stores etc...

Will more goods be delivered?

AAA estimates that car crashes cost $167B per year. That s $500 per person in US. What will be cost of "crashes" be with robotic cars?

Will we have "bicycles taxis" - vehicles that can drive you and your bicycle ?

Will we have more or less buses? Will we have really large buses - more like trains? or will there be more smaller more nimble bus-taxis?

the average driver in the US spends over $8,000 per year on auto expense? Would a 7x24 transit package cost more or less than that? What different kinds of "packages" might there be?

by Joe Deely on Feb 5, 2014 7:10 pm • linkreport

There are lots of tasks that computers do better than people.
That's true. Like airplanes; computers could fly them "better" than pilots for a long time. That includes taking off and landing, even in crosswind conditions. So would you ride in an automated airplane?

by Bob See on Feb 5, 2014 8:12 pm • linkreport

Perhaps someone can answer this question.

In the video with the blind guy in the Google car, how in the world does the car know where to stop so that he can order at the drive through and then again at the pick up window? Absolutely amazing.

by David G. on Feb 5, 2014 9:39 pm • linkreport

1. I put the odds of 100% self-driving cars saturating the roads SOMETIME in the future at about 50%. In the next 20 years at less than 1%. I may be wrong. I hope I am. But this is

2. "Probably be the final death of public transportation in the US at any rate." Again, I'm doubtful. I'd love to be wrong. I'd love it if electric-powered robocars were so clean, fast, safe, cheap and efficient that they rendered everything but walking, biking and flying obsolete. What a fantastic world to live in. Truly fantastic. So I'm rooting for it, but I'm doubtful.

3. And let's not forget some of the things we'll be losing if this comes to be.

by David C on Feb 5, 2014 10:47 pm • linkreport

Part of that got cut off:

1. I put the odds of 100% self-driving cars saturating the roads SOMETIME in the future at about 50%. In the next 20 years at less than 1%. I may be wrong. I hope I am. But this is a really hard thing to do. So I really don't want planners thinking about this right now. They have too many actual things to worry about without having to Guess what might be an issue if a certain kind of robocar comes along. Imagine if we had asked planners in 1980 to start building the backbone of the internet - what kind of mess they would have created terrifies me.

by David C on Feb 5, 2014 10:49 pm • linkreport

Re automated planes: no. UAVs (drones) are remotely controlled. Commercial aircraft still require pilot actions. Fewer than 1% of landings are flown using autolanding systems, and even those require pilots to set up the landing. Mostly, the automation reduces crew workload.

http://www.askthepilot.com/cockpit-claptrap/

by David R. on Feb 5, 2014 10:54 pm • linkreport

From investigation into recent accident for landing at SFO.

http://www.mercurynews.com/nation-world/ci_24700728/sfo-asiana-flight-214-hearing-begins-prevent-repeat

"Schiff and other experts said the reports and testimony at the hearing raised fundamental questions about pilot training; overreliance on automation that has become a concern as airplanes are designed to virtually fly themselves."

"Jay Joseph of Joseph Aviation Consulting, in Kyle, Texas, said the testimony revealed "some of the problems we have now with a new generation of pilots who are accustomed to everything being provided to them electronically, even to flying with the autopilot."

Just saying.

by Joe Deely on Feb 6, 2014 1:33 pm • linkreport

You saying a lot, Joe.

by Thayer-D on Feb 6, 2014 4:11 pm • linkreport

FWIW, I'm looking forward to the self-driving cars because I think they will be transformative.

- Drivers will, as always, react to improved safety by trading it for speed, so they will cut in front of any car they think is self-driving, assuming it will stop. We may conclude that driving and self-driving can't mix or we may have to create laws designed to change driver behavior (we don't have laws like that now). Either way a transformation of the sucky status quo.

- A self-driving carshare system can be point-to-point and can rebalance itself. Maybe we can get some self-riding bikeshare bikes too.

- Self-driving cars will not add enough capacity to the system to overcome the advantages of transit (high capacity, daily exercise, human contact). The latter two seem only to be appreciated by current (rather than future) transit riders. It is the capacity that forces the issue.

IMO, planners are sensible to not plan for self-driving cars. The whole point of the shift towards transit, supported by walking and biking (TWB) is to make cars irrelevant and to therefore allow re-allocation of space towards TWB. An irrelevant self-driving car is still irrelevant.

by Jonathan Krall on Feb 7, 2014 11:17 am • linkreport

Great topic. This may very well be the next paradigm shift in transportation.

The author highlighted many of the potential harms to the centralized, auto-alternative environment. Any increase in attractiveness and efficiency of automobiles will likely absorb potential mass-transit use (ie. often the reason I'll take the train over driving is so I can catch up on some reading, something I could then do in a driverless car).

If the driverless car does in fact come into popularity, planners should focus on garnering some of its potential benefits:

-Justify fewer lanes in pedestrian areas with the increase in efficiency from reduced following distance and shorter intersection delay.
-Work with the designers of automatic navigation systems to enforce traffic regulation. Speed limits can be better enforced, as well as bypassing through traffic flow around congested areas.
-If incorporated with car share programs, it may make them more attractive, especially to people who drive too little to justify having a license. They can also be made more convenient by picking passengers up rather than them having to go looking for a car share.
-Could offer a means to drop passengers off in heavily trafficked areas, then parking in an offsite location with lower land value. This also has the negative impact of a greater overall travel distance of these cars.

Overall though, I am not optimistic in the speed at which this technology will incorporate into daily life. It is more of a logistical challenge than technological. Department of transportation and public works will have to scrutinize their safety and liability issues before allowing them. Infrastructure will take awhile to adapt, especially as automated cars will trickle in mixing with non-automated cars in the beginning. I recommend planners take advantage of this time. Cities should also try to get ahead of the private automobile by incorporating the automated technology into mass-transit, which should be easier to implement and control.

by Chris Allen, PE on Feb 10, 2014 9:44 pm • linkreport

Well written post..You make a good point about increased demand, but you overstate when you speak in absolutes. Driver-less cars won't rubberneck, constantly change lanes and change speeds things that cause congestion. These cars won't have ego tied up in jackrabbit starts and horse power could improve the cars ability to live up to it's full potential on efficiency and and transit planners should simply acknowledge the issue. True, If I am saying that the Americans may not love their cars as much as they used to but they still prefer them to that available alternatives and I don't expect that to change much in my lifetime.

by SierraB on Feb 25, 2014 11:11 pm • linkreport

Is it really hard to imagine, that quite a few people actually enjoy driving and being in control of an automobile? To many, it is a feeling of freedom to get behind the wheel and operate a wonderfully engineered piece of machinery. Those that dread getting into the car and driving anywhere, I can see the lure of self driving, but quite a few people look to this time of day, driving home or to work or school, as a small break from other tasks and a bit of time out from the grind. The college road trip, the Sunday drive in the country, the family vacation, these experiences with the human driving, would not be so enjoyable to many. Driving is enjoyable

by James on Feb 27, 2014 2:44 pm • linkreport

I feel that in the future, people wont actually own any cars. Instead, car ownership would be limited to competing companies that operate self-driving car "taxi services." Whenever you need a car, you just hail it from a phone, it drives you where you need to go, and the car goes to its next job. Hopefully all the cars would also be electric, and whenever the car's battery s running low, it stops taking passengers, and goes to charge itself, before driving again. When this happens, there will no longer be a need for garages or places to park, and they can be repourpised.

by spj on Jul 8, 2014 11:36 am • linkreport

Add a Comment

Name: (will be displayed on the comments page)

Email: (must be your real address, but will be kept private)

URL: (optional, will be displayed)

Your comment:

By submitting a comment, you agree to abide by our comment policy.
Notify me of followup comments via email. (You can also subscribe without commenting.)
Save my name and email address on this computer so I don't have to enter it next time, and so I don't have to answer the anti-spam map challenge question in the future.

or

Support Us