A recent New York Times article discusses the arguments for and against putting off investment into public transport, in the hope that developing driverless car technology will make these investments unnecessary. There has been a lot of hype around these cars in recent years, although it is potentially tailing off more recently as the technology progresses a bit slower than expected.

Throughout a number of American cities, driverless cars are already affecting transport policy. In particular, there are some big questions about whether driverless cars will make major investments unnecessary.

But visions of the future they’ll bring have already crept into City Council meetings, political campaigns, state legislation and decisions about what cities should build today. That unnerves some transportation planners and transit advocates, who fear unrealistic hopes for driverless cars — and how soon they’ll get here — could lead cities to mortgage the present for something better they haven’t seen.

“They have imbued autonomous vehicles with the possibility to solve every problem that was ever created in transportation since the beginning of time,” said Beth Osborne, a senior policy adviser with the advocacy group Transportation for America. “That might be a tad bit unrealistic.”

In Indianapolis, Detroit and Nashville, opponents of major transit investments have argued that buses and trains will soon seem antiquated. In Silicon Valley, politicians have suggested something better and cheaper is on the way. As New York’s subway demands repairs, futurists have proposed paving over all that rail instead for underground highways.

Autonomous cars have entered policy debates — if not car lots — with remarkable speed. And everyone agrees that making the wrong bets now would be costly. Cities that abandon transit will come to regret it, advocates warn. Driverless car boosters counter that officials wedded to “19th-century technology” will block innovation and waste billions.

Dealing with this kind of uncertainty can be difficult. The last thing cash-strapped cities want to do is make huge investments into projects that will quickly become ‘stranded assets’. Nobody wants to be the modern equivalent of someone in the early 20th century who invested in more horses!

However, what I find always surprising is how this debate seems to come down to a fight between driverless cars and public transport projects. Why not driverless cars versus new highways? It seems like a bit of a ruse for the same old critics finding something new to jump on. One part of the article highlights this perfectly:

“I expect by 2030, most transit agencies are going to be zombie agencies that exist mainly to collect taxes from people to pay down their debt,” said Randal O’Toole, a senior fellow with the libertarian Cato Institute who blogs, provocatively, as “The Antiplanner.” In the meantime, he argues that cities should put no new money into infrastructure.

He acknowledged that he believed transit was wasteful for taxpayers long before everyone got excited about driverless cars. But now he and others who say no to transit also have something positive to say. Something better is coming.

Towards the end of the article we start to get some useful answers to the uncertainty question posed earlier. And, unsurprisingly, it’s our good old friend Jarrett Walker highlighting that some things really do stay constant over time, like geometry. And, when it comes to urban transport, geometry is boss.

Highways today can carry about 2,000 cars per lane per hour. Autonomous vehicles might quadruple that. The best rail systems can carry more than 50,000 passengers per lane per hour. They move the most people, using the least space. No technology can overcome that geometry, said Jarrett Walker, a Portland-based transportation consultant.

“Let’s talk about what we can predict,” he said. “The problem of the city is a problem of sharing space. In 2100, the problem of the city will still be a problem of sharing space.”

By that logic, cities should invest even more in high-capacity rail and dedicated bus lanes in key corridors. Autonomous vehicles might handle other kinds of trips — rides from the train station home, or through suburban neighborhoods, or across the parts of Las Vegas without rail.

So driverless vehicles aren’t going to do the job of high capacity rapid transit systems, regardless of the hype.

The real benefits from this developing technology will come from finding ways of more efficiently providing public transport to low density areas, as well as squeezing much more out of motorway corridors where the risks of people/vehicle interaction are much reduced. Bus routes can be changed overnight if something more efficient comes along, but we certainly should look long and hard at whether any future motorway widening would actually be required if driverless cars came along and allowed many more people to travel along these corridors.

Share this


  1. “Driverless car boosters counter that officials wedded to “19th-century technology” will block innovation and waste billions.”

    Always weird that somehow removing the driver from a car makes another 19th century tech into 21st century tech, while functional self driving trains have been a thing since the 20th century, yet is derided as old fashioned.

    1. Trains run on tracks, it’s difficult for them to hit another train unless someone fucks up, car’s don’t run in tracks, plus there will be hundreds of millions of them (millions in NZ) which won’t be automated, mixing automated and human controlled vehicles is not going to work.

      1. I agree with your comment – ‘mixing automated and human controlled vehicles is not going to work.’

        Major cities like New York, etc are struggling with the problem, as experts have pointed out that fully autonomous vehicles can not cope with unpredictable human reactions in confined or crowd areas. The experts say, the technology to hand this, is still at least 10 years away.

        For fully autonomous vehicles to be effective, they need to have to own infrastructure that is subject to unpredictable human reactions, so the need for PT is still valid. The cost for a large cities like New York to to modified or build the necessary infrastructure for fully autonomous vehicles is very expensive and very disruptive.

        Don’t see NZTA, central governemnt or councils investing in fully autonomous vehicle infrastructure for at least 20 years. What NZ should be doing is starting planing to upgrade the existing roading infrastructure like the State Highway network and major regional roads for semi-autonomous vehicles.

  2. Great and timely post. It is the topic of a workshop that is happening in Auckland next Thursday. If you would like information about the workshop, please click this link: https://www.eventbrite.com/e/preparing-cities-for-a-transportation-revolution-auckland-tickets-47661593108. There are still some spaces available.

    The tension that many people overlook is that there will be a period of time during which there are Level 4 autonomous cars, that will be allowed to operate as robo taxis, but that most people will not want to buy as household cars. The operating costs of these robo taxis will be really low, especially on a per person basis if they are high occupancy. In order to achieve their yield needs, they will compete with public transport for their customers. They will provide a better service at a similar price – ie door to door, and faster than transit. By being high occupancy (3 – 4 people in each) they will solve the congestion challenge that you mention above.

    The workshop on Thursday will explore the sort of governance structures that cities could put in place to ensure there is a good integration of public transport and robotaxi fleets and secure the best outcome for everyone. These structures could be implemented immediately with significant benefits.

    1. Paul, why do you think that “by being high occupancy (3 – 4 people in each) they will solve the congestion challenge that you mention above”?

      Did you not read Jarrett Walker’s comparison of people per lane per hour? Even quadrupling the lane flowrate by 4 (your upper bound) the 8000 figure is no match for transit’s 50,000. And of course, this ignores the fact that in our increasingly inequitable society, many richer people will not share their AV’s, and will almost double the vkt they do by sending the car off home or to find the cheapest park when they hop out.

      1. I think the quadrupling was for single-occupancy like it usually is now. If you times it again by four, 32000. But this assumes that the average is 4, it’s most likely to be 2.5 or something.

        1. At least this efficiency allows as to strip away lanes from the cars and give it to cycling and walking. The through-put would get pretty good if this happens.

        2. Bus lanes provide better efficiency for the same purpose, though. Unfortunately whenever there is a drop in traffic volume and the time is ripe for reallocating without much backlash, it doesn’t seem to happen, and the roads fill up with traffic again.

      2. Hi Heidi
        You are talking about Level 5 autonomy, and that is much further away I think. But I was not discussing theoretical capacity, rather the reality of Auckland’s commuter flows. Sure transit can do 50,000, but where is that a reality in Auckland at the moment? My starting point is a large fleet of street legal level 4 vehicles. Ideally we would manage the situation to minimise congestion and keep people using PT – but if the pricing is attractive enough, people will ride the HOV robotaxi, door to door, rather than the train or bus. It surprised me when I did the modelling just how low the AV cost structures are likely to be.

        1. Interesting article, and an interesting comment too (by Edward Greenberg) about London’s investment in transit after they lost the bid for the 2008 Olympics, due to the lack of mass transit in key areas.

        2. Do you really think ride sharing between 3 – 4 strangers is going to be popular? I think people are much more comfortable with strangers when they are in a bigger space, such as a train or bus.

          Think of that uncomfortable silence and looking away when people get into a lift.

        3. Jezza,
          i think that you are absolutely right that this may fail at the first hurdle; not an economic one, but just the psychology of it. I go further than your statement and say that people are uncomfortable with strangers on buses and trains and that is why you see very many people sitting singularly.

          I also look at what society is doing currently. We have shared cars, but with a driver, and we call them shuttles. They have not taken the world by storm and everyone will have their own thoughts as to why this is so. I struggle to see the technology that is going to be brought that we L4 or L5 so attractive for users.

        4. My question is why don’t we do it already then? Why don’t people share cars already and quarter their transport cost, taking advantage of transit lanes while they are at it? Why aren’t people using taxis and taxi vans for the same thing?

          …and what changes with driverless operation that would alter those factors?

        5. Agree absolutely. Ride sharing is not predicated on having autonomous vehicles – it could happen right now if people wanted but for the reasons given (that people basically don’t like strangers) it doesn’t. Time we separated the issue of vehicle autonomy from the issue of whether people will willingly abandon PT for ride-sharing services. I also think that owning a vehicle personally will remain popular for those spontaneous weekend jaunts and for going on holiday. And the rest of the time, if there’s a car in the garage I think it will be the default option over sharing a vehicle with a bunch of strangers (and possibly going on a bit of a circuitous route to meet their particular travel needs).

        6. I think it is more pragmatic that that. The very reason for using a car is point to point travel when you want to go, without delay from others.

          Car share means not really point to point travel, not quite when you want to go, and delay with others. If you are sharing with four people thats four pick ups and four drop offs, going out of your way and spending time to suit them.

          If you are happy with that you’re probably already catching the bus, which will usually be more direct as it tends not to veer off down side roads and long driveways.

    2. I’m not sure why you think they will be cheap for peak hour commutes? Either there will be enough for everyone at peak hour in which case there will be a lot of those cars doing nothing most of the rest of the day (doesn’t sound cheap), or there won’t be enough for everyone and they can charge as much as demand allows at peak (again not cheap)

    3. Paul I seriously doubt we’ll see Level 4 autonomy within the next decade or two. Maybe not for centuries. Maybe never.

      We haven’t come close the the level of AI that would be required. Even at level 4, cars would have to negotiate their way through roadworks, follow the directions of police directing traffic, predict the likely movements of pedestrians and cyclists, drive in poor weather conditions, respond to drivers flashing their lights or waving them on to negotiate a narrow gap etc. All this without having to slow to a crawl, and with 100% safety (otherwise the manufacturers will be sued out of business).

      All this is worlds away from the sort of autonomy / driver aids we see today, such as following lanes on the open road, adaptive cruise control etc. While that stuff gets better and better I don’t see any indication we’re about to make a huge leap into the sort of AI required.

      The exception might be in we have level 4 cars with extremely restricted domains, e.g. motorways only, but then that wouldn’t deliver anything like the benefits you’re talking about.

      1. Hi Nick
        I think you are talking about Level 5 automation. I am talking about Level 4, and fully expect it to be restricted domains, which is the main reason I don’t think households will buy them. What makes this field so interesting is all the uncertainties, such as how long before the huge level of investment by many companies delivers a viable Level 4 vehicle. It might take a decade, or it might be next year. If cities wait and see, they will not have the structures in place to get the best advantage when those fleets come knocking.

        There is lots of hype in this arena, but the rate of learning seems to be pretty high. The workshop is based on the book by Niles and Grush called The End of Driving, that is getting pretty good reviews as a sensible counter to the hype.

        See here for the book: https://www.elsevier.com/books/the-end-of-driving/niles/978-0-12-815451-9.

        1. Level 4! Lets be clear what Level 4 imagines: That the bot drives the car when everything is fine, but when disaster or complication strikes the human, who until now has been eating a pot noodle or watching Harry Potter, suddenly and expertly takes over to avert disaster…

          Think about that for moment.

          This is the reverse of technology making driving safer (seat belts, ABS, airbags, etc) this is technology doing the easy bits, while dumbing down and distracting would be drivers then suddenly and unexpectedly expecting them to take over, asses the situation and take skilful command…

          In my view Level 4 doesn’t exist in any practical real world form. It is simply the car-cos and tech-bros trying to get the public to accept a dangerously incomplete product. Because the real deal, the all weather, all condition, all event, safe bot-car is very very hard, prohibitively expensive, if not functionally impossible.

          Level 4 are clever machines, and exist now, but will kill users, have killed users, and more concerning, civilians. People outside the vehicles. Level 4 is bullshit.

        2. Hi Patrick. No, actually what you are describing is Level 3. Level 4 expects the vehicle to be able to do everything without a human needing to intervene, but only in specific locations and circumstances. It cannot go everywhere, just in the places it has been approved for based on some as-yet undefined criteria. Here is a link to an image of the levels: https://www.google.co.nz/search?hl=en&tbm=isch&source=hp&biw=1919&bih=847&ei=3ppaW-mXAZKi-Qaz7TM&q=levels+of+automation+sae&oq=levels+of+automation+SAE&gs_l=img.….0…1ac.1.64.img..14.11.3853.0..0i30k1j0i5i30k1j0i8i30k1.0.26Nku-rQKPw#imgrc=1nkBF25MVdzt_M:

        3. I agree, based on the fact that these issues have been known for a long time and are not really possible to solve, given that the human operator will be ‘out of the loop’ when the emergency requiring their intervention occurs. Here’s a classic paper on the topic:

          Bainbridge, L. (1983) Ironies of Automation. Automatica 19(6), pp. 775-779. https://doi.org/10.1016/0005-1098(83)90046-8 (payment required).
          (Or available free here: http://www.bainbrdg.demon.co.uk/Papers/Ironies.html)

          Juicy bits from the paper:
          This paper suggests that the increased interest in human factors among engineers reflects the irony that the more advanced a control system is, so the more crucial may be the contribution of the human operator…

          The designer’s view of the human operator may be that the operator is unreliable and inefficient, so should be eliminated from the system. There are two ironies of this attitude. One is that designer errors can be a major source of operating problems…The second irony is that the designer who tries to eliminate the operator still leaves the operator to do the tasks which the designer cannot think how to automate. It is this approach which causes the problems to be discussed here, as it means that the operator can be left with an arbitrary collection of tasks, and little thought may have been given to providing support for them…

          Unfortunately, physical skills deteriorate when they are not used, particularly the refinements of gain and timing. This means that a formerly experienced operator who has been monitoring an automated process may now be an inexperienced one….
          There is some concern that the present generation of automated systems, which are monitored by former manual operators, are riding on their skills, which later generations of operators cannot be expected to have…

          The other important aspect of cognitive skills in on-line decision making is that decisions are made within the context of the operator’s knowledge of the current state of the process. This is a more complex form of running memory than the notion of a limited capacity short-term store used for items such as telephone numbers. The operator has in his head (Bainbridge, 1975) not raw data about the process state, but results of making predictions and decisions about the process which will be useful in future situations, including his future actions. This information takes time to build up…The implication of this for manual takeover from automatically controlled plant is that the operator who has to do something quickly can only do so on the basis of minimum information. He will not be able to make decisions based on wide knowledge of the plant state until he has had time to check and think about it…

          We know from many ‘vigilance’ studies…that it is impossible for even a highly motivated human being to maintain effective visual attention towards a source of information on which very little happens, for more than about half an hour. This means that it is humanly impossible to carry out the basic function of monitoring for unlikely abnormalities…

        4. If there has to be a human driver, how is it cheaper or more desirable than taking a taxi? Very few people use taxis for commutes because it’s crazy expensive. I don’t see how level 4 automation would change that.

    4. jumping in an autonomous car to go pick up 3 strangers before heading to your destination doesnt seem very efficient to me.

        1. I cannot see it being particuarly quick either, adding several km to your journey, a bit like those airport supershuttles where your shuttle picks up 3 other strangers from random addresses along the way to save you a couple of dollars but adds 30 min to the journey.

        2. unless we have a really good urban form with better density. Then the passengers might be close enough for this to not be a problem.

          But then of course, road space is at a premium, far better given over to PT and cycling… and will work even better than shared AV

    5. How do you think Auckland’s existing roading infrastructure is going to cope with large number of robo taxis roaming the city?

      You know the technology and urban planners are saying mixing automated and human controlled vehicles is not going to work without the right roading infrastructure.

  3. I think share ebikes will mostly take up the role of moving people to PT stations for moderate distances. With increasing population, it’s the best chance for the infra to keep up. The share cars (autonomous or not) are best used in hailstorms and getting to stations far away (most likely for city-city, or in very very sprawled out cities).

  4. Do driverless car supporters seriously think the likes of London and Manhattan will be turning off their rail/metro? Or are they just a solution for Auckland?

  5. I don’t think it is ‘driverless car supporters’ as much as people who are thinking about how the future might unfold. The result will be different between London, Manhattan, and Auckland, I would expect. For serious damage to PT, these robotaxis would not need to take all the PT riders. This is already happening in some US cities with Uber/Lyft even at their existing price points taking away PT riders. Imagine the impact if Uber controlled an AV fleet and really wanted to fill it up with people at peak time. We would love the average occupancy until we saw where all the passengers came from.

    1. I reckon the future is pretty hard to predict and so is the rate of change. Smartphones changed everything so quickly, but conversely electric cars have been feasible for a long time now but are still a tiny percentage of the fleet.

      My personal prediction is that slow local bus routes will disappear but decent rapid PT will become more important as people will stop owning cars. But that is only one of many possibilities.
      I’m not sure how roads will have any more capacity until all cars are automated – that must be many decades away?
      Also someone still has to pay for the roads – so the amount we currently pay in petrol tax will still have to be charged to AV users in some way.

      1. my understanding is that the capacity effects are not quite “all or nothing”: If an automated and connected vehicle (AV) is being followed by another AV, then the vehicles can communicate to reduce the following distances, thereby freeing up road space. So it’s not necessary for all other vehicles are AVs to get some marginal capacity improvement.

        On the flipside, some of the safety benefits of AVs are likely to reduce road capacity in other ways, e.g AVs will always observe orange traffic signals. My hunch is that highways may see capacity increase from AVs, whereas local roads will see capacity reduce. Travel-times on local roads are likely to be undermined in other ways, for example as AVs generate additional driving from circulating to pick-up/drop-off passengers.

        My take on AVs is that they simply reinforce a view that the future is uncertain. The sources of uncertainty are many and varied, as others have noted. A strictly econo-bot approach might be to reflect this uncertainty in (1) higher discount rates and/or (2) sensitivity testing. The net effect is to raise the relative benefit/costs needed to justify investment.

        Incidentally, one of the hidden benefits road pricing is that it provides better information on driver’s willingness-to-pay for travel, which in turn reduces the uncertainty associated with transport investment more generally.

  6. Driverless cars may well cause more chaos on the roads than less. If we assume that a driverless car will avoid a crash at all costs, the logical extension of that is that a car with a driver in it doesn’t need to take any notice of it as it will always get out of the way. How long will it take for the driverless car to be paralised by the fear of a driver doing something stupid. If you know a driverless car will get out of the way why bother stopping at a red light ?

  7. Great article and an interesting discussion.

    Overall I agree with the sentiment: “Don’t be too proud of this technological terror you’ve constructed. The ability to double private vehicle occupancy and quadruple road loads is insignificant next to the power of geometric constraint.” Or something like that.

    Where lies the better part of valour?

    I’d say in investing in better transit systems and not worrying whether level 4 or 5 AVs arrive tomorrow or next century.


    1. Option value. If you are dramatically wrong, it is much quicker and cheaper to decommission a transit line you don’t need than to create one when suddenly find you do need it.

    2. Social/community form. A network for AVs is a car-centric network. A network privileging PT and active modes is more likely to create a people-centric network.

    3. Behavioural pragmatism. If the operating costs of an AV are going to be that low, when I use one I won’t be sharing it because the cost won’t make it only a rich man’s privilege. And I sure AF won’t put my kids into one that might have some stranger hip-to-hip with them. So the passenger loading ‘benefits’ are just sales gloss as far as I can tell. (And yes, you get strangers on buses, but many more at any one time, and probability says most of them will defend your kids from perverts).

    4. Regulatory pragmatism. Once you’ve put your investment in, you use your regulatory toolkit to privilege the heck out of PT/active mode access, especially to civic and commercial centers where you don’t want lots of cars mucking about.

    5. As stated clearly in the article – time. Look how long EV production is taking to gear up and eat into sales. AVs won’t be any faster. Even if they do save us all, there is easily a decade, probably more, before they do. Investment needs to deal with today’s realities.

  8. All driverless robocars are is public transport. If people don’t want to share a bus, why would they want to share a driverless robocar? Driverless cars will simple put *more* cars on the road because now you don’t need a driver.

    Imagine the convenience of being able to stop outside the coffee shop, put the car into “orbit mode” ’cause you couldn’t find a car park, and when you’re done, use the cell phone to call the car back. I cannot see how driverless cars reduce the number of trips.

    1. Yes, like Uber etc, shared AVs plan on taking the money you hitherto spent on parking, OK… the problem here is they assume there will be free road space for them to occupy instead… public realm we instead have other uses for in city centres….so given this is all either already full or highly contested, the plan also relies on there being no increase in vehicle numbers, but also somehow huge success and growth to replace all transit, which is orders of magnitude more spatially efficient.

      It don’t add up.

  9. So it seems to me, that the take home message is that if the AV proponents’ wet dream comes true then private vehicles will become so much more space efficient that it would be pointless creating any more road space for private vehicles, given the imminent arrival of this AI-guided transport panacea. Cool. Glad we’ve cleared that one up.

    1. AVs will make freeways more efficient through closer following, but they can’t/won’t increase capacity of mixed environments, with pedestrians, lots of conflicts, etc, ie city streets. The amount of efficiency increase is therefore incremental, not substantial (and requires full fleet conversion to happen), nowhere near enough to supplant the spatial efficiency of good Transit, and can in no way touch underground rail at all, which takes up zero road space.

      They will likely be handy in the suburbs for getting to the Rapid Transit Station, if cheap enough. In fact they’ll need good quality Transit to be of much use at all. But will also bet competing with all those very very low cost little machines; shared small battery devices; bikes, scooters, boards, for the first mile/last mile segment. SSBDs

  10. I look forward to driverless tech turning up in short order in trains and buses.

    They’ll be in common use there well before most folks have access to driverless Ubers.
    As they’ll deliver sufficient benefits to overcome to costs in PT.

    And unlike most driverless car scenarios, Driverless PT also comes with a built in ability to deliver effective social outcomes.

    Unlike many driverless car scenarios, which often end up with the “owner” of it having it circling the block to minimise their inconvenience – at a hell of a high level of inconvenience for everyone else affected by such “f*ck you” behaviour. Especially in an urban environment.

    But its not just American cities where the reality distortion field is happening around Driverless tech.
    MoT and NZTA and completely and utterly enamoured with it. They let their “feelings” about how good it will be hang out for all to see in their documents.

    Unfortunately, those types of folks don’t just get to screw up a single cities transport system as they can do in the US. No these folks actually get to screw the transport system for an entire country with their inane, uncritical thinking and completely blind acceptance of such tech as the saviour for any problem you care to name.

    The same crappy thinking which assumed flying cars would solve all these pesky transport problems by now.

    I STILL don’t have a flying car. Nor any time frame for when one will be available.

    But a lot of the thinking around personal mobility that went on in the 50’s and ’60s when a lot of the Auckland Motorways and suburbs were “designed” [if such a term accurately describes what amounted little more than throwing spaghetti against a wall and taking what stuck to the wall as the plan for what to build].
    And so 60+ years on, we’re stuck with the poor, and clearly deranged thinking from those days.

    We don’t want to repeat that experience anytime soon. Do we?

    1. Yes, “MoT and NZTA [are] completely and utterly enamoured with it.” This is what has me most worried. Like the Uber Flying F**kwit Fantasy, and the call for Hi-Tech Solutions to Road Safety when Low-Tech Solutions have been ignored systematically, the people who should be raising critical eyebrows at this nonsense – and protecting us from the wasted resources even considering it – are welcoming it in like their sex lives depend upon it. Which doesn’t surprise me in the slightest given some of the people pushing some of these things.

  11. I see AV’s doing three things mainly:
    1. Reducing private vehicle ownership
    2. Improving links to mass transit
    3. Reducing the need for parking in cities

    It’s a nonsense to think that filling our roads up with AV’s will solve congestion – they are congestion! There is also the question of what they are doing in between trips – are they driving around the streets (creating congestion) or are they fighting for car parks? Most people use their cars something like 1% of the time, and for the other 99% they are off the road. If the AV’s are not individually owned then it should decrease private ownership, but not traffic as the same number of people will be wanting to move the same way as they do now. There will just be fewer vehicles doing it due to that 1% thing.

    Which begs the question, where is the capacity going to come from when everyone wants to go to work or go home at the end of the day? Even if you have the pipe dream of 4 users per vehicle, you still will have a complete shortage of capacity during peak hours, or a complete overcapacity off-peak. The AV people never ever put that much thought into it, it’s a bit like the Skycab concept – people who don’t know enough get excited about it but when push comes to shove it’s just not practical as an alternative.

    1. The over/undercapacity at peak/off-peak hours is a problem. But as you point out, it will be an improvement on the status quo, because the utilisation will be higher. Each vehicle might be used for 2-3 runs per peak instead of 1. Maintenance and charging schedules can presumably be arranged to have all the cars on the road during peaks.

      I agree with others here who have said that AVs are not 100% positive e.g. they will make congestion problems worse. But they will reduce the amount of city space wasted on car parking. I think less vehicles that are professionally maintained will reduce resource use (both steel and energy) so that’s a good thing too.

      1. Yes that’s the whole point that people are trying to make – AV’s are potentially useful and could lead to improved PT efficiency and less demand for parking, but they will never be a replacement for mass transit as they are designed to replace private car use more than anything.

        1. and therefore will undermine PT, which is also designed to replace private car use. We have an unhealthy mode share, and unless AV is regulated to limit its use to very specific situations, it will exacerbate our problems.

  12. The big advantage I see in autonomous vehicles is that they won’t need sleep or holidays or extra pay for unsociable hours. So all of the fleet not in maintenance can run continuously and the money saved on employees can be spent on more vehicles – allowing more routes/more frequency/extended hours. Smaller autonomous vehicles can service narrower/less used routes, taking people to where bigger vehicles run along broader paths.
    I’m also hoping for re-scheduling of at least some deliveries – what if “overnight” from Fishpond actually meant our AV will drop it in your secure box during night time hours? – when the people are off the streets. Day time deliveries could left for the urgent stuff, freeing streets for people/dogs/bikes/trees.

    1. Careful though. Nighttime deliveries by AV to reduce daytime courier trips equates to adding capacity. Same as any other form of adding capacity (road capacity, intersection priority to cars, parking capacity). It induces more traffic. So the traffic noise won’t stop at night. There won’t be a quiet time to have a walk in the dark. And the capacity relieved during the day will just induce more daytime traffic.

      Of course, if you reallocate road space at the same time, any freeing up of the streets can actually be given to people/dogs/bikes/trees. But that option already exists: we have the option of more space-efficient travel already (bus lanes) but to free up this space for people/dogs/bikes/trees, it’s a fight every time. What’s stopping us isn’t lack of AV, but lack of understanding or will to change.

      1. I seem to recall a story about ancient Rome trying to reduce congestion by switching deliveries to nighttime. They went back to day time after the noise meant residents couldn’t get a good night sleep.

  13. When you read “libertarian Cato Institute” in an article it is good indication that you can safely skip the next paragraph or so.

  14. True, induced demand is a problem for every transport system. But these AV’s will surely be electric (so silent at night) and, since their every move is tracked to the mm and the second, road pricing will manage demand. You’ll choose “overnight” because it’s cheaper than day time.
    Of course, if if road pricing actually worked (technically and politically), the battle of understanding to reallocate street space might also be won. Friday afternoon dreaming before I face the congestion again…

  15. Just as conventional transport modes are predicted to be swept aside by a swing to autonomous vehicles, so too will AV’s soon become outmoded once teleportation techniques are perfected. We should be very wary of wasting valuable resources on AV technology which will rapidly become a stranded asset.

    Beam me up Scotty.

  16. All of this AV talk often conveniently ignores the huge costs that will be required to retrofit a city to make it AV friendly – all the work to date suggests they can’t operate in an urban environment (motorway/ rural is a different story) it wont be cheap going in and signalising most intersections or pedestrian crossings across a city. Full adoption is also forecast to be decades away – we need to be focussing on addressing existing issues now rather than kicking the can down the road for the next generation(s).

    1. I don’t think it is feasible to retrofit all the cities in the world, especially during the changeover period where most cars are still human-operated. However I think there is an incredible amount of work going into making AVs capable of operating in urban environments. It is more challenging, for sure, but I am confident that the technology will be ready in a few years. It may not be perfect, but it will be better than the average human driver.

      1. the main work going into making AVs capable of operating in urban environments is lobbying to impose further restrictions on pedestrians and reducing liability of AV car companies in the event of (inevitable) deaths.

  17. In the 1980s I was involved in carpooling, in a publishing company I worked for whereby an employee with a big car – me – would circle around half of Auckland and pick up five or so staff members and drop them off at night. It worked well for a little while, but then the little issues started – someone would have to work late to finish something, another person wanted to socialise at night, another had their started time altered. In the end it became just too much of a hassle so they decided to drop the scheme and drive their own cars to work and I decided to catch the bus.

  18. So here is how we prepare for an AV future: 1) remove all parking in the city, because it won’t be needed 2) Put in light rail, because that can be made autonomous (and maybe we’ll allow AV’s to hitch a ride on a team track) 3) separate cyclists in their own lanes to avoid confusing the AV, 4) stop widening motorways because AV’s won’t need as much space. There, the future is proofed.

  19. I could see driverless cars integrating nicely with trains and buses and replacing some lesser used and off peak buses on routes where adding bus lanes are difficult. A driverless car could pick one up and drop one off at the train or main bus station and leave. This could reduce also take some pressure of the Park and Ride car parks.

    It seems mad to think that we’ll only need driverless cars in future.

  20. I wonder if a perverse outcome of driver less cars, and particularly driver less taxis could be even more congestion. The cost is reduced from a normal taxi fare, the passenger selects “non share” on his/her app and lo and behold more cars.

    It seems to me that there are two very different objectives between PT operators and driver less car operators.PT is looking to maximise an outcome for the city; and driver less car operators are looking to maximise outcomes for themselves. One of those outcomes maybe that the more cars potentially the better financial outcome i.e. is it better to carry three passengers to the one destination for one split fare, or have three journeys with three fares?

    I approach this matter with a healthy skepticism.

    1. I think the consensus is that AVs will lead to many more cars on the road. In some places, average occupancy will be even below 1, as some of the vehicles drive around empty. Extra trips will be possible e.g. kids can order a driverless taxi to get to school/sports/etc. Elderly people will have better mobility options.

      The question is whether reduced following distances can make up for more vehicles on the road? The analysis in the article suggests that it can – up to a factor of 4, but that is still a lot less efficient than mass public transport.

  21. Regarding the Sharing or the AV system, has there been any feedback on the uptake of the AT initiative “Match-a-ride”?

Leave a Reply

Your email address will not be published. Required fields are marked *