will driverless cars abolish buses? (email of the month)

An emailer who wishes to remain anonymous tries to put it all together:

Your recent [Atlantic article] regarding “bus stigma”, along with the concurrent proliferation of various autonomous car posts that I’ve seen all over the web, got me thinking: I am starting to believe that a certain “technophilia” (as you put it so well) not only applies to the “rail-in-any-and-all-situations” proponents, but also to the increasing number of urbanists who have come to view the impending autonomous car future as one in which buses are replaced by Self-Driving Vehicles (SDVs), as they’ve started to regularly call them).

Case in point, this blog post [at Grush Hour].

[Grush's] opinions appear to be very similar to those of the various urban planners and urban designers I’ve met and spoken with over the past year or so.  Now keep in mind that these folks view themselves as “progressives” on the issue of the need for public transportation.  In this case, the author thinks the recent Wall Street Journal Op-Ed which espouses that rapidly developing autonomous car technology means we don’t have to build high speed rail is flawed and incorrect, and that the trip purposes of the modes and their associated distances, et cetera, are sufficiently different to mean that high speed rail will still have a place, even with SDV’s.

However, as I was reading this, the kicker came in the last part of the post (I highlighted the most pertinent portions):

The frontier benefits of the SDV will accrue during 2022-2042 as special, restricted applications such as replacing mostly-empty and oversized urban buses, expensive and poorly driven taxis and shared cars. Here is where I would like to see Winston’s call for private funding focused: urban fleets of self-driving jitneys to replace every form of motorized shared vehicle (bus, taxi, street car, shared car, vanpool) from the front door of your home or work right up to the light-rail and heavy-rail transit station and vice versa. Replace them all. Then by 2045, maybe the US Congress will be able to pass another Surface Transportation Reauthorization Bill in plenty of time to eulogize the last of the personally-operated SOVs and fix the last of the traffic signals in time to remove them all, because they will no longer be needed.

I don’t honestly know what to make of this …; it seems that people have such a “stigma” against buses that they view SDV’s as being able to replace them completely and instead we would then have an on-demand or subscription-based autonomous jitney/“Johnny Cab” system which takes you (of course) to the rail line – if it doesn’t provide for your whole trip.  To be clear, this author is not the only one I’ve read or person I’ve spoken with who believes this – rather, this rationale is what I am increasingly seeing being espoused everywhere I look.

So, as far as I can tell from reading these materials over the past couple of years, it would appear that the most “anti-urban” (and perhaps least progressive) sort of folks see SDV’s as replacing the need for any public transit whatsoever – including the rail modes right up to true High Speed Rail – while the most progressive and pro-urban folks see SDV’s as at least replacing almost any and all buses (but not necessarily rail), and that “traditional, fixed route transit will only be needed in the densest cores of our cities”. 

Now – and here there are some shades of gray – some posts and articles I’ve read say the highest volume corridors may still justify some form of “traditional” bus service, but then – I kid you not – most of these folks go right on to say that such a corridor (i.e., one where the fundamentals of an enhanced bus or bus rapid transit regime may work) should just likely be rail (or “tram”) lines in any event.  (I saw this theme especially crop up in the comments to your Atlantic Cities post.)

Again, there are many assumptions about the economics of SDV’s, their energy sources, the legal ability to have them operate without a licensed driver behind the wheel, et cetera – but the fundamentals are still there in their arguments, and I think you get the picture.  In any event, I have three inter-related questions for you stemming from this theme:    

1.       Is this another form of “bus stigma”?  Or rather are buses simply most suited to an urban transit speed/capacity niche whose days are numbered, so to speak, as being the domain of the bus?  I find it interesting that most urbanists (but not all – there are actually a few out there who think a robust and subsidized SDV system can even replace rail and other fixed guideway high speed/capacity modes) seem to salivate at the thought of getting rid of the “lowly” bus, but that SDV’s can’t, won’t or shouldn’t replace rail.  Or they at least say that bus service will be relegated to the densest corridors where (eventually) it would be replaced by a rail line in any event.

2.       The follow-on question I have for you is this: is there a future – in a world where SDV’s have been fully developed – for the “regular” transit bus service that operates along a corridor where service is only provided every 15 minutes, or every 30 minutes, or even every 60 minutes on weekends (i.e., the vast majority of North American public transit service)?  Or will SDV’s eliminate the need for such bus service? 

The argument I was told this weekend by a city planner while discussing your “bus stigma” posts and the latest “Google-car” advances goes something like this: a very significant portion of riders today – of whatever income group – use buses just because they don’t have to drive or look for parking, and the fare is reasonable.  If an SDV service can take you door-to-door, without you needing to drive, park or fuel it yourself, for a fare similar to the bus, and for a total trip time at least as fast as the bus (yet likely shorter) but just slightly longer than a purely private car, then in the vast majority of North America where densities are not all that high “the big ole’ regular bus running every 10 to 15 minutes is history, along with the horse-drawn omnibus, and transit agencies will find themselves in the same territory as buggy whip manufacturers…”  This lady even quoted you back at me: she pointed out that if “frequency is freedom”, then “think of the immense freedom and mobility the public sector-subsidized SDV can provide, while saving us the costs of big buses and their unions…”  Sheesh.  So, what’s your take on the future of the regular, non-heavy-corridor bus service that runs every 10 to 15 minutes?   

3.       Finally, my last question: the Gensler fantasy which you took down so effectively seems to always re-appear in some form or another; does it become more (or even less) viable with the assumption that the vehicles are autonomous? 


Jarrett here.  My answer to all of these questions is the same.  It's in my book, and it should be on the screen-saver or refrigerator of every well-intentioned urban visionary:

Technology never changes facts of geometry!

We can be quite confident that nobody (on this world or any other) is going to discover a technology that changes the value of pi or that suddenly causes large, uncompressable objects to fit into boxes smaller than they are.  We know that because we understand the special status of mathematical and geometrical facts.  Indeed, they are so much more certain than any other "fact" that we should have a different word for them.  

And this, friends, is a geometric fact:

Bus bike ped in same street

If you define a "car" as "a separate enclosed vehicle for every passenger or party", then the geometric fact about all cars, self-driving or not, miniaturized or not, is that they take vastly more space per passenger than effective public transit.  This will not be a problem in low-density suburbs, but cities, by definition, are places with relatively little space per person.  Self-driving cars will certainly improve the efficiency with which cars use space, so they will shift the calculus somewhat.  But the bottom line will still be that if you want two crash-safe metal walls between every two strangers going down the same street, you will need a lot more space than if those two people can sit next to each other on civilized public transit.

You will also need vastly more metal and equipment, which means that the self-driving-car-replaces-transit fantasy involves massive industrial production with severe consequences for energy security and greenhouse-gas emissions. 

As for the idea that somehow these cars will replace buses but not rail, this may be true around the margins.  Grush's reference to "replacing mostly-empty and oversized urban buses" is a crude approximation of the issue and misses the point about why these sights occur.  The real problem is that most "legacy" labor agreements don't allow transit agencies to pay drivers less to do the easier job of driving a small bus in a low-demand area, and given that it's cheaper (due to high maintenance costs of fleet diversity) to run a standard modular bus everywhere. (Vancouver's TransLink is a spectacular exception.)  Most transit agencies run low-ridership service that is a drag on their budget, but that meets social inclusion or equity needs.  Most agencies I have worked with would be delighted to see those predictably low-ridership "coverage" services transitioned to a more decentralized or low-cost model, or moved off their books entirely, so that they could focus their big buses in places where they'll be full.

So to sum up, the technophile urbanists who believe that self-driving cars will eliminate the need for public transit are making several mistakes:

  • They are assuming that technology will change the facts of geometry, in this case the facts of urban space.
  • They are assuming that the costs of having every passenger encased in a metal sphere (in terms of production energy and emissions) are readily absorbable by the planet.  (To be fair, the SDV discussed here is one that you don't own but just grab when you want it, so if it replaced the car there would be far fewer cars.  But that's different from replacing a bus.)
  • If they think that self-driving cars will replace buses but not rail, then they haven't informed themselves about the vast diversity of different markets that buses are used to serve.  Self-driving cars many logically replace some of these markets but not others.
  • They believe that public transit is incapable of improving in ways that make it more positively attractive to a wider range of people, despite the fact that it is doing so almost continually.

Again, the whole bus vs rail confusion here arises from the fact that technophile urbanists classify transit services according to how they look and feel, whereas transit experts care more about the functions they perform.

So yes, buses are currently doing some things that other tools could do better, especially in sparser markets.  Some agencies, like Vancouver's, already have the tools to solve that problem.  But when a huge mass of people wants to go in the same direction at the same time, you need a rail if you have tracks and an exclusive lane for them, or a bus if you don't.  I don't care whether it's rail or bus, but the need for a high-capacity vehicle running high quality service that encourages people to use space efficiently — that's a fact of geometry!

38 Responses to will driverless cars abolish buses? (email of the month)

  1. MLD August 16, 2012 at 7:44 am #

    Most arguments about some auto technology making transit obsolete make the same mistake – they assume that automobiles will improve in a meaningful way but transit vehicles will not. We see this often with the argument about environmental quality – the argument goes that cars are getting much more efficient and so they will be more energy efficient than transit, ignoring that as the technology has improved transit vehicles have also gotten more efficient.
    And so the argument continues with so-called self-driving cars. If cars can have this technology, then why wouldn’t transit vehicles be able to? Self-driving buses would reduce transit labor costs and allow for better frequency at all times (just like driverless metro does), which would likely mean more people would take transit, and therefore create more demand for improvements like dedicated lanes, etc. that would make transit service even more desireable.
    The technophile urbanists would respond do your claims about the space-using reality of cars with the argument that these magic cars will be able to speed along at 130mph with only six inches between bumpers, meaning you can fit many more cars on the road! Of course that argument might be reasonable for grade-separated interstates but I don’t think people will accept it in cities. Lane capacity in cities is basically already maxed out (during peak periods) since you have to have provisions for pedestrians to actually be able to cross streets.

  2. ant6n August 16, 2012 at 9:27 am #

    Driverless (midi/mini-)buses may pop up before driverless cars. This not only makes more sense from a geometric perspective, but also economically: The incremental cost of making a vehicle driverless is smaller per passenger if the vehicle can transport more people. Or, a bus already costs 300K, a private vehicle maybe 25K. If you add in, say, 50K to make it driverless, where would the investment have less of an impact?
    I’d say driverless/on-demand buses will help making individual cars obsolete for example in the suburbs, rather than the other way round.

  3. Neil August 16, 2012 at 11:25 am #

    I was reading this and remembering vividly a GM-sponsored ride at some Disney park in the late ’80s. Which had a “cars in the Year 2000” theme, including, of course, self-driving technology. Perpetually 10 to 20 years away, I guess.
    Also brought to mind the old joke of Bill Gates vs GM…specifically the “If GM developed technology like Microsoft, for no reason at all, your car would crash twice a day” punch line.
    Jarrett, of course, notes the transit implications if such a technology ever does develop, though fails to mention that the same technology could also be applied to buses, and already exists for rail, keeping the cost difference in private vs shared vehicle operation much the same as it currently is.

  4. patfromigh August 16, 2012 at 11:45 am #

    I have seen photographs of US cities from the pre-automobile era. These pictures date to around the decades from 1890-1910. It is interesting to see downtown traffic jams of streetcars at rush hour, when everybody is trying to get home at once. The streetcars are gone now and we still have traffic jams, now some are much farther out from the urban core.
    Transit companies have had to deal with the extra resources needed for rush hour that aren’t used outside those times. We have traffic clogged roadways that are often next to empty lanes going in the opposite direction. So the SDV services will have to have a surplus of equipment for such special times. There will be rush hour problems, trust me.
    There is an assumption that SDVs will be cool. If I have to wait 45 minutes for my electric network vehicle to show up at rush hour, and then find out the previous occupant threw up in the thing before I get into it, how cool is that?

  5. NCarlson August 16, 2012 at 1:06 pm #

    I wonder if they might not have gotten it exactly backwards in terms of rail. If it becomes feasible to operate driverless buses there’s going to a be a definite loss in the attractiveness of rail from a functional standpoint. I rather suspect that something along the lines of automated battery supplemented double articulated trolley buses would ruin the economics of most new rail line short of a full heavy rail metro (and how many of those do we actually need).
    Of course a lot depends on the cost of the technology involved, but I don’t see any fundamental reason that it will actually be that expensive. I do think we’re at least a couple decades from realistically allowing completely driverless full size buses in mixed traffic though. Of course that leaves me with the thought that as long as a couple decades seems that still leaves as much of career being in the era of this stuff as not…

  6. david vartanoff August 16, 2012 at 1:15 pm #

    SDVs are the next desperate dream of the single driver who abhors social contact. whether a Prius (feelgood but still burns oil) or some midget car,they are all still personal cocoons which need somewhere to be parked/warehoused when not in use. As others have pointed out, there isn’t sufficient waste value land in a vibrant (ie worth commuting to) CBD. As long as a multitude of worker bees are expected in the office, SOV commuting is just dumb utilisation of valuable resources.

  7. Cascadian August 16, 2012 at 3:28 pm #

    I think the real effect of automated cars on space is that they obviate the need for most of the parking that currently exists. If people don’t have to drive, many won’t bother owning a car, freeing up not only garage and driveway space, but also the space for street parking. They won’t need a parking space at work, or when they go shopping. Donald Shoup (I think) estimates that there are at least three parking spots for each vehicle currently. With automated cars, even if people take exactly the same number of personal vehicle trips as now, 2/3 of that space becomes unnecessary and will eventually be reclaimed for other uses, probably rather quickly. I would not be surprised to see sidewalks widened throughout cities to take space that was formerly used for cars. Architecture would shift away from prominent garages (now wrapped into storage and workshop areas that do not face the street) and toward a revival of porches.
    Transit is still necessary for geometric reasons, but drivers will be replaced with fare checkers or transit police (or nothing). Really, the distinction between bus and train will probably blur a lot.

  8. Jeffrey Bridgman August 16, 2012 at 5:30 pm #

    SDV = PRT without a guideway?
    hehe 😉

  9. Chadnewt August 16, 2012 at 8:35 pm #

    It is amazing how often discussions of self-driving vehicles tacitly assume:
    – they are battery electrics
    – they are shared/on-demand vehicles like taxis
    These 3 characteristics are not correlated. But these common thoughts are the result of decades on sci-fi movies with electric automated taxis. And that is why “urbanists” from the suburbs imagine that self-driving vehicles are the transportation solution for the future.

  10. Eric August 16, 2012 at 9:06 pm #

    Even as private cars become driverless, buses will only be able to become driverless if the transit workers’ union allows it, and given the way such unions will work, I fully expect bus drivers to fight tooth and nail to prevent computers from taking their jobs away, regardless of the cost to the riding public.
    Even driverless taxis might have a difficult time taking taking root, as human taxi drivers will fight tooth and nail to convince regulators to ban them, to avoid being undercut. Or maybe they won’t be banned entirely, but their numbers will be restricted so much that a ride on a driverless taxi of the future will cost just as much as a ride on a human-driven taxi today, with all the money going to the driver today going to the taxi company instead.
    Private cars, of course, won’t face any of these restrictions. And the result will be a weird situation where most people could ask a friend for a ride in his driverless car for free (while the owner is at work or sleeping at night) that would cost somebody without a car-owning friend $50 or more through a commercial taxi service.
    However, the really interesting coup of driverless private cars is how they turn parking and traffic congestion up-side down. All of a sudden, everyone who commutes into the city for work or for a baseball game will realize it’s cheaper to have their car drop the owner off in front of his destination and self-drive to the nearest free parking space. In reaction to this nearest free parking space being street parking in residential neighborhoods, virtually every neighborhood will be forced to set up a resident permit parking zone, including those miles away from their city centers. So everyone’s self-driving cars will end up either circling around the block all day or driving themselves back to their owner’s driveway, whichever is operationally cheaper. While you do drastically reduce the amount of parking needed downtown, you also drastically increase congestion on all the streets in and around downtown. Freeways will become nearly as congestion in the reverse direction as in the peak direction. Congestion pricing could, of course mitigate this, although a more politically expedient solution (albeit much more expensive) may be to simply widen all the roads instead.

  11. EngineerScotty August 16, 2012 at 11:45 pm #

    Last year, Portland Transport did an article on this subject in response to an entirely predictable Randall O’Toole screed on how self-driving cars would utterly obsolete mass transit, therefore we should abandon investments in busses and trains, yadda yadda yadda. Many of the same points were raised as here (both article and comments): High-density areas will be unable to handle cars for everyone, and autonomous vehicle technology could also benefit transit. (Unions will resist, but when technology has the potential to completely replace a job, unions tend to lose. How many police officers have full-time gigs directing traffic through intersections these days?)
    And yes, I agree that self-driving busses could easily make many current applications of rail obsolete: Which is more desirable–a 2 car LRT that comes every 15 minutes, or a 40′ bus on an exclusive ROW that comes every three?
    A few other points.
    * Chadnewt assumes self-driving, electric power, and car sharing are orthogonal. Power source, yes, but self-driving autos solve many problems associated with current car-sharing arrangements. Unlike taxis, they don’t need a driver (and in most places, being a hack is a lousy job); unlike carshares and car rental, insurance isn’t a problem (simply don’t let the renter drive), and the issue of delivery/pickup is no longer an issue.
    * At the risk of inviting the proverbial plague of locusts–if there’s something that the self-driving car obsoletes, it’s something that never was in the first place–fixed-guideway PRT (personal rapid transit). Autonomous vehicles have the virtue of being all gadget and no bahn.
    * As the technology matures, I expect manually-operated autos slowly disappear from the roads. More and more streets will become limited to autonomous-vehicles only, and getting a drivers’ license will be come correspondingly harder. Eventually, manually-operated vehicles may only be permitted in places like racetracks or rural neighborhoods.
    * Another major area that might be affected is freight hauling. Long-haul trucks may not need to have anyone on board at all –long haul truck drivers generally aren’t involved in the loading/unloading of their vehicles, so there’s no reason for anyone to ride in the truck. Delivery vehicles will still need to have someone to handle packages, but a CDL will not be required for the job.
    * A while back, Jarrett did an article about Barcelona, where busses were treated like emergency vehicles and given priority over other traffic, who were expected to pull over and let a bus pass. This scheme, which required lane controls on the route (busses were NOT equipped with sirens; instead signals directed cars to vacate the lane when a bus was approaching), allows transit to be more reliable without an empty-most-of-the-time exclusive lane. On an autonomous-vehicle-only road, this could be enforced by vehicle policy; when a bus comes by, other vehicles pull over, making the setup even more reliable.

  12. Andre Lot August 17, 2012 at 12:12 am #

    I think the technology involved in self-driving vehicles on open roads is really a game changes. Contrary to previous attempts, it relies on making vehicles smart instead of purely dedicated ROW with complex technology embedded in the road/track/way.
    In order to fully assess the magnitude of how much a game changes this could be, I think we need to think a bit ahead of “having your own car self-drive”. For a starter, we could conceive a scenario in which many people opt out from owning a car that is used at most 10% of the time for subscribing to some sort of Zipcar of the future.
    As long as a car can be booked and arrive for pick up within minutes, suddenly a single facility for these shared cars can have a wide catchment area in the suburbs. You don’t need the same number of cars to perform the same trips they do now and, on top of that, the effects of parking would be wiped out (even in an urbanized area, cars would just drive themselves to some off-site to be stored). People who still wanted their own vehicles could do so, but they would probably cost much more.
    Another possible development, albeit far-fetched now, is a scenario in which all cars are self-driving (or non-self driving cars are restricted from most built-up areas). Coupled with the “subscription model”, the overall size of cars could be smaller for short trips (even short trips to a high-frequency, high-speed transit line!), or even different car fleets could be booked according to need. But as they are smarter, the amount of weight it takes to make them crash-safe would be reduced. I’m thinking on vehicles more on the golf-car size, for instance.
    Taxis would surely be an early causality of self-driving cars.
    They would also open a whole new world for things like online shopping with almost-instant delivery: you post your order, a small light vehicle leaves the regional warehouse and heads straight to your residence.
    What many people miss is that the innovation of the “Google car” and the likes are not giving the wheels and power pedal extremely precise instructions, that would have been possible at least since mid-1980s. It is their technology of recognizing obstacles, other vehicles, signs, people and positioning and reacting accordingly. This implies trams could also theoretically benefit from the same stream of awareness technologies and become self-driven.

  13. Eric2 August 17, 2012 at 1:31 am #

    Eric – the amount of street parking in a typical US neighborhood dwarfs the amount of parking in a few downtown garages. It will be easy for a SDV to find parking much closer than the typical commuter’s home.

  14. xavier debidour August 17, 2012 at 2:20 am #

    Dynamic geometry matters too
    The pictures of car vs bus vs bicycle are very good but they only show the space taken when stopped or parked.
    When moving at speed V, the cars (and bicycle also) will need a lot more space for braking distance at speed V (1 for each car) while the bus will only need 1 (longer) breaking distance!

  15. Chris Bradshaw August 17, 2012 at 9:03 am #

    I agree with the point that the cost of SDV technology will make it more likely on larger vehicles that are on the road more than 1.5 hours a day — thus expect it to occur on shared vehicles of a larger size (although they could be much smaller than city buses.)
    So, even though moving vehicles could often be empty (ZOVs vs. SOVs, eh), that would occur mostly when the streets are devoid of any users. These vehicles could be used “privately” only during very low-demand periods; otherwise the users would have to be comfortable with sharing space with strangers.
    As the increased sharing literature points out, the same technology that allows for SDV also can track all passengers in real time, and monitor in-cabin behaviour, making shenanigans very unlikely.
    What I find odd is how everyone seems to think a) drivers will not fight for their rights to pilot these vehicles, and b) the pedestrians and cyclists will not fear being hit by them.
    It makes one wonder whether people will simply embrace ways to avoid the kinds of trips that autonomous or semi-autonomous motorized transport serves. Live locally.

  16. cph August 17, 2012 at 9:30 am #

    “[…] and then find out the previous occupant threw up in the thing before I get into it,”
    Maybe they’ll be self-cleaning, like those JCDecaux toilets….

  17. Gordon Ross (@gordonr) August 17, 2012 at 10:03 am #

    This discussion reminds of the remarkable book by French technology theorist Bruno Latour about the personal rapid transit system project in Paris throughout the 70’s and 80’s. A difficult read at times, but a fascinating dissection of many of the same beliefs…

  18. Marc August 17, 2012 at 12:54 pm #

    To reiterate Chadnewt’s excellent point…
    For all the babbling over SDVs, the panters consistently overlook one thing: what fuel will all these future pods run on? Bountiful, cheap gasoline increasingly looks iffy. And there’s not nearly as much shale gas as the propagandists claim. Hydrogen’s out. So is compressed air. Or algae. Or alligator fat. So far we haven’t found any way to scale up these decades-old technologies. And centuries-old battery electric technology hasn’t improved much either. So how will all those SDVs zip around?
    We forget that cars proliferated because we found a way to run them on cheap and convenient fuel: gasoline/diesel. That’s the *only* reason cars became so commonplace. Remove that cheap and convenient fuel (as is gradually and painfully happening already) and you can’t exactly “swap it out” for another wunderfuel. I’m not holding my breath for electrics, autonomous or otherwise. After more than a century of experimentation (with billions of dollars thrown in by venture capitalists, governments, and universities) we still can’t get electrics to scale up to replace gas cars. So the idea of a broad-based electric SDV is still a pipe dream.
    SDV planners and dreamers are totally jumping the gun: instead of babbling about redoing transportation networks/systems around a handful of existing science-project-scaled curiosities, maybe they’d do better to fix regular transit first? That’s way more realistic. Otherwise they might end up wasting a ton of time and money preparing for something that’ll never actually come (at the universal scale we fantasize about).

  19. Marc August 17, 2012 at 1:03 pm #

    I also wonder if Gordon Ross is onto something here… is this the newest delusional fad of the PRT kooks? Even though we implemented a few experimental PRT systems (a handful of which actually *do* work for extremely limited, reductive purposes), we never found a way to scale the technology up.
    Who knows where SDVs will go, but right now they looks like PRT 2.0: just because Google can create a couple of science projects doesn’t mean the technology can feasibly scale up. So let’s see how it evolves first before we start distracting the transpo planners with gee-whiz schemes. After all, we can’t even keep the regular roads in good shape…

  20. bo August 17, 2012 at 1:17 pm #

    Brilliant response, jw.

  21. Rational Plan August 17, 2012 at 2:38 pm #

    I’ve always thought a networked bus system would be a likely development.
    A lot of the assumptions about SDV’s are bunk. Unless we see some radical changes to batteries and super fast recharges, electric cars are going nowhere. SDV’s will still be gas powered.
    I like how people assume parking will no longer be needed. That cars spend all their time parked and not moving people etc. Why do people think these cars run on moonbeams and rainbows. It costs money to move a car whether it is powered by gas or electricity. How expensive would parking have to be for it to be cheaper have car endlessly search for a free parking spot or drive all the way home again after it dropped you off.
    It will take a long time for SDV’s to reshape our cities. At first it will be mainly private cars that are SDV’s and that great for accident rates but not much else. In reality people will still want their car to be parked close to where they are. If decide to leave somewhere at an unplanned time whats the point having to cool your heels while you car has to spend 15 minutes getting back to you.
    Real estate will change, but will still need lots of parking on site. Sites can be densified as there is no need for lots of surface parking. Buildings, in particular retail, can be close to the street, with auto valet zones at the front of buildings where people wait for their own cars or rentals while out back are robot car stacking cubes for where vehicles wait until needed.
    As mentioned before the fundamental physics means that all these cars can’t go where they want at peak times. Either you have gridlock, and maybe public transit lanes or some form of road pricing.
    If autocabs and then autobuses come into being that changes a few things.
    A city could just cut it’s core network to a system that is economic and cocentrate it’s investement in the city core and inner suburbs. The rest could be covered by autocabs, if they were cheap enough or the subsidy required less than the equivalent bus service.
    Bus service could peter out after 9pm, as cabs became cheaper outside the rush hour.
    A more integrated approach, requiring a much bigger centralised computer role would a networked system. This would arise if individual rental of vehicles is still too expensive for a significant proportion of people. Such a system could aggregate demand to certain destinations (hospitals, universities, malls, office parks etc) and pick up fares along the way.
    Such a system may require a few bus hubs where riders may be required to change to a different bus, as too few people on the bus they got on want to go in that direction, or they could change to higher order transit for faster travel over longer distances.
    Everyone would need a ‘travel agent app’ on their smart phone to negotiate to request a destination and find who is going there, or at least in that direction.
    Such an agent could give price and time options upon a destination request, with the highest cost being sole use of vehicle in peak time, all the way down to multiple linked trips on local bus shuttles.
    One seat rides are still preferred by everyone, but if it is often only 30 seconds to a minute for the next mini bus going in the direction you want then it won’t be so bad. As long as local auto buses are busy then they should be cheap.
    A lot of people make trades offs over cost and time. In London now with train and tube fares so much higher than bus fares, I know of plenty of people who commute to lower paying jobs by multiple buses to save money. In some cases the savings are huge and £30 to £50 a week is a lot of money on a lot if incomes.
    In the future such a network could be layered by incomes. With the top being able to afford single car service with $150 dollar a barrel running costs and congestion pricing, with the middles mixing it up with single car service on nights out and a mix of rail and local shuttle in the day and the working poor pretty much stuck with local shuttles and multiple changes.

  22. Matthew August 17, 2012 at 7:08 pm #

    Marc, one thing I suspect may be possible with SDVs is to setup a virtual “guideway” of sorts which allows vehicles to connect to a trolley wire or conduit for power. Plus, for serving a myriad of short trips, it may not even be necessary for that. Just pick up the nearest SDV and when it drops you off, if it needs charging it goes to the nearest station or guideway.
    Anyway, there’s lots of other issues for SDVs to solve, including the everpresent geometry problem, so we’ll see how it turns out, if at all. It would really be a shame if the advent of SDVs led to a new round of destructive urban renewal and the death of cities.

  23. Christopher Parker August 18, 2012 at 7:08 am #

    What the fan’s of self-driving-technology are imagining is a taxi, but automated.
    That sounds less sexy, doesn’t it. But that’s what a self-driving car that then goes along to the next person is.
    Already taxi’s have all the advantages of not needing a car for every person or parking and so on. Nothing about self-driving is different.
    In America, taxi’s are a small part of the market (Manhattan excepted). Why would self-driving cars be any different?
    Cost? Will automation be cheaper – enough to woo people to depend on self-driving taxi’s instead of their personal auto? Maybe. Maybe not. Technology isn’t cheap. Besides, taxi’s are already cheaper than the cost of owning a car for many people and American’s prefer to have their own.
    The labor cost of driving is zero. (Not really – but that’s how people think of it, and perception determines decisions). I only see this working where labor costs are a factor (which means fleets, trucks, yes para-transit, etc) and for those who really want a chauffeur.

  24. Marc August 18, 2012 at 10:46 am #

    LOL, Chris, that reminds me of a comment a Chicagoan once made to a PRT fanboy: ‘We have four-person PRT pods in Chicago too. You summon one whenever you want and tell it exactly where to go. Plus it doesn’t need a separate guideway system. We call it taxi.’
    Matthew, I’m skeptical that we’d see those infrastructural improvements implemented on a broad basis at a time when we’re having trouble maintaining our existing infrastructure. When governments are desperately selling every public surface for ad space, I doubt they’ll be able to throw money into guideways, catenaries, new traffic/intersection control systems, charging stations, etc., etc. Even with all the public and private money we’ve already thrown into conventional electric cars and charging station experiments, the private sector response has universally been “meh.” They know these ideas can’t scale up, so they only indulge in a few public relations sops to appease us and quietly scrap them when our attention shifts to the next new stunt. IMO overhead-powered SDVs are a pipe dream: there’s a reason bumper cars never went mainstream – they can’t feasibly scale up.
    The SDV might go the way of the Segway: lots of publicity over a handful of limited stunts, but it’ll quietly fade away when people realize it offers no earth-shattering advantages over cheaper existing technology. After all, for short trips there already are taxis, bikes, feet, buses… All of these allow you to go somewhere without having to pilot a two-ton vehicle and finding a place to park it. And the infrastructure for these convenient car-free alternatives is already in place… no need for the equivalent of a gigantic HSR project or new Interstate Highway Act to build another layer of fantastically expensive redundant infrastructure.

  25. Joel N. Weber II August 18, 2012 at 8:51 pm #

    If you have a bus route today with 20 minute headways in the mid day, operated with a 40′ bus which is not excessively full, would a person who argues that frequency is freedom, favors driverless operation for metro rail systems to increase affordable frequencies in off-peak hours, and tries to avoid false rail vs bus distinctions argue that replacing the 40′ bus with much more frequent, smaller driverless vehicles would be a clear improvement over the bus?
    Once you have small vehicles with only three or four people anyway, does it make sense to allow them to deviate from a fixed route in order to save passengers walking time, and possibly also to add some flexibility so that passengers who want to pay more can get a one seat ride where they might otherwise have had to transfer in the 40′ bus era?

  26. Gavin Seipelt August 19, 2012 at 12:25 am #

    I think a good current example that illustrates of the capacity limitations of PRT/SDVs are many typical amusement park rides.
    This ride at Disney California Adventure could be described as being similar to PRT in many aspects…A small 4 person vehicle, one of which is dispatched every few seconds, and a loading area where the vehicles move through slowly bumper to bumper (and riders quickly step across into it)
    But the point is, amusement park rides are famous for getting lines, sometimes an hour or more.
    In a busy city you could just imagine the exact same types of lines forming at PRT/SRV stations as hundreds of people try to filter into little 4 person vehicles

  27. Marc August 19, 2012 at 5:59 pm #

    Who knows how SDVs will evolve, really. The point is that all we have right now are a handful of Google experiments, so any discussion on “bus eras,” or the impending obsolescence of buses (or any other kind of transit), or on the urgent need to reorganize transit orgs to focus on accommodations/infrastructure for SDVs is laughably, absurdly premature.
    O’Toole can crow about the supposed obsolescence of transit in lieu of SDVs when (and if) it actually happens, but right now SDV talk is just a silly distraction from the important task of fixing regular transit (which is probably all that O’Toole is aiming for). *If* the technology evolves to the point where there are tens of thousands of SDVs running out there on the existing networks, then maybe we can start distracting the planners and engineers with grandiose infrastructure schemes to better accommodate those SDVs.
    We have plenty of time: even the personal-conveyance-obsessed US didn’t initiate the Interstate Highway scheme in 1896 right after Ford wheeled his Quadricycle out of the garage. We had to wait a while and observe how the technology progressed to understand exactly where and how we would have to accommodate it. Not so with the hype coming from some in the SDV crowd: ‘We have to start planning/building/accommodating for them NOW!’ is the general meme I get from the more enthusiastic people, even though not a single SDV yet exists in ordinary private use. Spend prematurely, and you might just end up wasting a colossal amount of money.
    Hell, I think those west coast states pouring money into experimental charging stations and other doodads for conventional EVs are *already* pissing money away, because there’s no evidence yet that that ancient technology will ever scale up to meet the expensively-provided infrastructure. (In contrast, we refined, built, and expanded the gas station network *after* gas cars had become a predictable mainstay!)

  28. Eric2 August 20, 2012 at 1:58 am #

    Re oil:
    1) Natural gas can be used instead of oil.
    2) At some time in the next couple decades, electric car technology may pass a tipping point. Increased quality will lead to increased demand which will allow for development at a faster pace, in a feedback loop.
    3) The fuel costs per user are much smaller when multiple passengers use the same vehicle.
    Re PRT:
    The unavoidable problem with PRT is the need for a comprehensive expensive guideway. SDV avoids this by using existing roads.
    Christopher Parker:
    Labor is the highest cost of light rail, which has huge vehicles carrying large numbers of passengers. For a taxi, which has similar labor costs but much smaller equipment and fuel costs, labor is by far the dominant expense. SDV will make taxis MUCH cheaper, and they will change from a luxury good to something comparable in price to current public transportation.
    Gavin Seipelt:
    You don’t need SDV stations. They can stop at the curb anywhere. Probably more efficiently than human drivers.

  29. francis August 20, 2012 at 10:29 am #

    Something tells me that before IT technology makes self driving cars mainstream, it will have already made most personal transportation unnecessary through telecommuting, online shopping, etc.

  30. Wad August 22, 2012 at 10:43 pm #

    @francis, the technology to make electronic economic transactions a substitute for the meatspace is here now and is fully mature.
    The revolution has happened, and whatever could be substituted has been substituted.

  31. Wad August 23, 2012 at 1:40 am #

    @Andre Lot, you know why self-driving vehicles will not be a game-changer? Because you are expecting it to be a game-changer.
    That’s the paradox. You’re concentrating your expectations on an event that will be nothing less than awesome when it occurs. At the same time, you don’t notice smaller causes that build up to a larger event. You’d keep your eyes out for high tide at the sea, but not notice how the raindrops make the nearby river rise.
    IFF there comes the day when self-driven vehicles become viable for mass consumption, the event would be banal. You have to have hundreds, if not thousands, of incremental technological processes to be viable first in order to make self-driving vehicles even possible.
    And that’s just technology. How do you know if society is ready for yielding their autonomy to a machine to do their driving? As technology becomes more complex, acceptance has not become any more easier. Look at how addle-brained society is becoming over privacy and “Big Brother” concerns — the risk-threat calculus is out of whack.
    So many little things have to fall into place for the “if the day comes” to become a “when the day comes.”

  32. Bern Grush September 12, 2012 at 9:06 pm #

    re Grush Hour. Here is the full-article version of the offending blog at Grush Hour: Kiss Your Bus Goodbye

  33. Charlie - Vehicle Recovery March 5, 2013 at 8:41 am #

    I never realised that if driverless cars become available, it will mean that people would be more likely to purchase a car than take public transport. However, this could also mean that there would be too many cars on the road…

  34. wood items June 8, 2013 at 2:50 am #

    We have special wood item experts having years of experience in creating these items. They always take extra care while creating them and give them a beautiful and attractive look

  35. James C. October 22, 2013 at 5:28 pm #

    > They are assuming that technology will change the facts of geometry, in this case the facts of urban space.
    No, that is not the assumption. The assumption is that it will not change the fact that it will happen anyway. People will get in their $3 smart taxis and arrive at their location faster than with a bus.
    > They are assuming that the costs of having every passenger encased in a metal sphere (in terms of production energy and emissions) are readily absorbable by the planet. (To be fair, the SDV discussed here is one that you don’t own but just grab when you want it, so if it replaced the car there would be far fewer cars. But that’s different from replacing a bus.)
    Again, that is not an assumption. It just won’t change the fact that people will use smart taxis.
    > They believe that public transit is incapable of improving in ways that make it more positively attractive to a wider range of people, despite the fact that it is doing so almost continually.
    Unless buses become as fast, private and direct as taking a car, no.

  36. JB March 11, 2015 at 1:10 pm #

    The elephant in the room as far as I’m concerned is capacity. I’ve heard some SDV enthusiasts wax poetic about how this is going to solve rush hour congestion in a big way, but the fact is this – in a self-driving vehicle where all of the accident liability is likely on the manufacturer of the vehicle, ALL of them are going to strictly enforce a safe braking distance that will include some fudge factor to account for road conditions.
    As it stands now, with driver-controlled vehicles, we tacitly allow people to tailgate (follow at much less than safe braking distance) because the police will rarely ticket someone doing 70 mph 50 feet from the next car’s bumper. Manufacturers will *not* accept the liability of explicitly allowing a car to follow at less than safe braking distance + fudge factor. At higher speeds, the braking distance increases exponentially. So, you wind up with a 70 mph operation on the highways with about 5 or 6 cars per mile density (and therefore crappy throughput), or you ratchet ALL the speeds down so you get to a “sweet spot” that allows the highest density with still barely acceptable travel times. How do you think folks will like being down to 30 mph on the Interstate where they used to do 75 mph 200 feet off the leading car’s bumper?

  37. Eloisa April 2, 2018 at 7:49 am #

    There are definitely challenges to be overcome with regard to driverless cars in the realm of technology, the regulations, and the acceptance by the public. From one perspective, people are already comfortable with getting into a car over which they have no direct control – you do it every time you get into a taxi, a bus, a Lyft, etc. There you’re faced with the variable of a driver for which you have no knowledge and simply are simply placing your trust in them. Driverless cars have the potential to remove that variability and uncertainty. In fact, ride sharing may be the initial “killer app” to generating acceptance among the public but there needs to be an understanding of and trust in the technology (including the sensor fusion that brings together a host of sensors such as radar, lidar, odometers, GPS, inertial navigation systems, and more) and an expectation that there will be a high level of performance provided by these vehicles.

  38. Rapid Transient May 2, 2019 at 11:54 am #

    Autonomous cars won’t be able to replace transit.
    Could autonomous vehicles replace public transit?
    Not all autonomous vehicles will be personal cars; some will be buses.
    Without the labor cost of one driver per vehicle, buses would be much cheaper to operate. Buses could also be smaller, as there’s no need to spread the cost of a driver over a number of riders. (They’d be small but dense, i.e. having a high number of passengers per square meter.) These factors would make it easier for private transit operators to enter the market (if legal restrictions were loosened or lifted). Smaller buses would also mean that companies could compete along the same route, each conceivably earning a profit and providing frequent service.
    To ensure that traffic doesn’t get snarled by too many companies, cities could implement dynamic tolls (decongestion pricing) that responds to demand. The cost of the road would (among other factors) tell companies whether they should squeeze seats into vehicles or give people more space, the former in order to lower the toll-per-passenger ratio and the latter increasing passenger comfort.
    Could Autonomous vehicles replace regional (commuter) rail and subways?
    Rail lines could be paved over and run as dynamically tolled roads inhabited by minibuses. Passengers per square meter
    in a minibus vs a train would take only a minor hit, and minibuses would be able to run much closer together. There could be an increase in capacity, frequency, and flexibility and a decrease in travel times.