Many people comment to me how driverless technology will make transit obsolete, but I disagree, many of the advantages driverless cars have will apply to driverless transit too. Of course, we already have cities already have driverless transit by way of their metros and further, I’d argue that driverless technology will provide greater advantages to transit. This is due to the fact driver constitutes a large portion of the operational costs of running services so driverless technology will allow more frequent and therefore more useful services without increasing costs.

I also hear a lot about driverless buses, but not about driverless light rail. Having a quick google it turns out that people are in fact working on it and I found a great article & video about the technology Boshce is delivering in Frankfurt that they argue is just one step towards driverless Light Rail.

“The driver assistance systems being developed now, Bosch says, are the first step towards automated light rail. The systems are being engineered to work in all terrain, weather, and congestion scenarios.”

So when someone says driverless cars, say driverless vehicles, because transit will be able to take advantage of that technology as well to deliver cheaper, more frequent and more useful services.

Share this


    1. They aren’t trams though, not in the sense that they run in the street and have pedestrians around. They’re small metros, it’s entirely grade separated which makes it easy to automate.

      It’s a useful point though, metros have been fully driverless for decades now, railed street vehicles (trams/LRT) will be next, then buses, and finally cars/taxis. With regard to the complexity of the automation task and the economics of delivery, it’s clear we will have driverless trams in widespread use decades before we have a comprehensive driverless car fleet in action.

    2. Also Dubai metro or Skytrain Vancouver which goes up to 90km/h – but yes, they require complete separation from traffic (usually achieved by elevation or tunnelling), unlike trams which are described in the video in the article.

    3. Except that it didn’t work very well. The train’s computer simply counted the number of wheel turns to each station, stopped and opened the door. Any wheel spin meant they would stop short of the station. Stopping for obstructions was also a bit of a problem. When I was there they put train managers on who could stop and start the stupid thing. Drivers who were not called drivers because that would have been embarrassing on a driverless system.
      DLR is a great example of what happens when land use and transport are not planned together. It was intended to serve a largely residential area. Then Thatcher opened Docklands to very high intensity development that DLR could never cope with. So they had to resurrect the plans for the Jubilee Line Extension and massive cost and build the Limehouse Link. I worked with people who planned the original DLR, they told me had they known the level of development that actually occurred then they would have opted for a tube line instead.

      1. Yes DLR isn’t the best example of driverless metro, in fact its probably the worst (unless you count the Detroit people mover). Vancouver is probably the best, it has a network length and number of stations almost identical to the Auckland rail network (a hundred and a bit km of lines and 40 odd stations)… but it moves 110 million trips a year, so about six times as efficient as our trains are currently.

        1. DLR certainly had its challenges early on including several lots of upgrading to cope with the growth of Canary Wharf etc. It now shares a common SELTRC moving block signalling system with Skytrain in Vancouver and is arguably more effective, in patronage terms than Skytain.

          DLR currently has 34km of route, 45 stations and in FY2013 had an annual ridership of 110.2 million ( whereas Skytrain has a system length of 79.5km, 53 stations an annual ridership of $117. 7 mm (

      2. The DLR doesn’t operate like that mfwic, it uses SELTRAC which is a Tales product. From memory the system uses wires to form a accurate magnetic sensor of where the train is on the track.

        That’s pretty standard and even driver operated lines may have very accurate train detection (eg on lines with platform side doors that need correct alignment).

      3. The DLR doesn’t operate like that mfwic, it uses SELTRAC which is a Tales product. From memory the system uses wires to form a accurate magnetic sensor of where the train is on the track.

        That’s pretty standard and even driver operated lines may have very accurate train detection (eg on lines with platform side doors that need correct alignment).

        1. They used a GEC system up until the mid 1990’s. It required an axle to generate pulses which were counted to figure the exact location of each train. Problem is of course steel tires on steel tracks slip. The problem of trains stopping most of the way into the station was fixed by having a person on-board who wasn’t called a driver who could move the train forward. That and other reliability issues and the 20 million pound annual subsidy made the DLR a national joke for years.

        2. I think the current system uses wires that cross every 20 to 25m that detect the train but within that they still count wheel turns to get the exact locations. The reversing of polarity of the wires show when a train passes the cross over so they know which section of track but they dont know where in that section of track it is without the turn count.

        3. If it’s the seltrac system that is for ‘signal’ positioning on the network, but at stations there is a separate mechanism for stopping in the right place with much finer resolution. Optical sensors I believe.

        4. I think the trains are slowed in advanced of the stations based on the inductive loops so they are going slow enough for the station boards to be detected. Those on-board sensors then slow the train to a stop at the right location. All of that was added after the original DLR was opened. The original ‘driverless’ system was a bit of a joke in comparison.

  1. You are right Harriet, driverless tech applies just as much [if not more so] to PT as it does to other vehicle types.

    But in this Governments eyes; as shown, once again, by the comments made by Bridges recently about “Driverless Bus Rapid Transit” being “a promising technology” [to eliminate the need for anyone to consider LRT anywhere].
    They clearly show that they and their advisors only believe that there will be two kinds of “Driverless” vehicles on the roads – cars [and by extension Ubers] and of course trucks.

    Driverless Trains or Metro? or anything else that is actually more likely than not, to come to pass sooner than those?

    “Forgeddaboutit!” is the instant response.

    It won’t be too long before the big truck crowd and NZ Bus lobbyists to start demanding that NZTA provide dedicated roads/lanes on existing roads for their driverless trucks and buses to use (for public safety of course).

    1. AFAIK most modern LRT and trams have only drivers and dont have conductors/ guards etc. Why do you think it would be any different in NZ?

      1. I’m just thinking of our train system and thinking they’ll probably want the same thing for light rail. Not exactly what I’d wish for either.

        1. Train Managers are a legal requirement on passenger trains at the moment I believe. Light rail will start without this encumberment.

        2. That will depend on how LR is defined under the law, the easy way out makes them light trains and subject to the rules of rail.

        3. I don’t think that there are any specific legal train crewing requirements – but I stand to be corrected!

          There are already driver-only rail vehicles, on the Wellington Cable Car.

          Light rail is subject to the Railways Act, just as heavy rail is.

        4. Light rail is defined as light rail under the law. There are separate parts of the legislation to cover light rail already. No they don’t have to meet heavy rail standards, no they don’t need TMs and there is zero chance they will have them.

        5. The Railways Act 2005 does not specifically require this as far as I can see – however it is likely that the safety cases , required under the act, for organisations operating passenger trains, may specify that Train Managers are required on passenger trains. If this is the case, presumably such organisations could seek to change this requirement through a revised safety case and submit this to NZTA for approval.

  2. The LRT (light rail transit) in Singapore has been driverless since it began in 1999: It is used as a feeder service to the main MRT (mass rapid transit) lines. However, in the past few years, more MRT than LRT lines have been built. This may be partly due to the LRT breaking down fairly often, leading to a proposal to scrap it:

  3. Electric self-driving cars are likely to be mainly hired rather than owned. This means a big drop in car ownership, which means a big opportunity for PT. All those former car owners are not going to hire a car for every trip, that’s where PT comes in.

  4. If you think about the problems to be solved, a driverless rail-based vehicle must surely be an easier task than a road vehicle. It is surprising then that driverless trains and LRT seem to get overlooked in the discussion. Yet there are large potential savings. The cost of the driver is a significant proportion of the total cost and the need to schedule meal and rest breaks and return crew to their base is a significant operating constraint. I wonder whether higher safety standards inflicted on the rail sector are inhibiting development?

    That said, driverless vehicles will have the biggest impact in the taxi sector – as Uber already understands. Most of the cost of a taxi is the driver. Driverless taxis will make universal cheap public transport available at a fraction of the cost today.

    1. From what I have read the literature on driverless vehicles breaks down the task into four or five levels, from not automated at all through to fully automated anywhere all the time. The thing with taxis and private cars, they need full level five automation (have to work on any street, any time, any route, i.e. on the fly and perfectly). Driverless PT only needs level two or three, operating driverless on a fixed pre-programmed route on a fixed environment (especially fixed if it’s on rails).

  5. From a passenger’s point of view, does it matter if the PT vehicle is driverless or not? Either way, you get in, do whatever you like during the journey, then get out. If anything, having at least a human vehicle manager on board might be preferable so they can intervene when something goes wrong. This is not a problem worth solving. More frequent and reliable PT with better coverage should be job one.

    1. The running costs of paying a driver would be the main thing. Any money saved is money that can be used to extend the network, put on more services or simply reduce fares.

    2. Where it works for the passenger is in the second order effects. Basically if you take the driver out you halve the marginal operating costs. So that means you can either run literally twice the frequency (and capacity) all day for the existing budget, or you can run the same frequency for half the cost.

      So driverless means twice as much transit, or half the ticket price. Either of those is probably pretty big deal for passengers.

    3. In Vancouver, when there was unexpected demand, they just put extra cars into the system without needing to call in drivers first and wait for them to arrive. Also it means that it’s feasible to provide small but frequent services which are very convenient from passengers’ perspective. Think our trains are quite large and they can move a lot of people per hour but the convenience factor isn’t there (you have to wait up to 15 minutes for a train in peak) where in Vancouver they come every 2 minutes.

  6. In Vancouver, when there was unexpected demand, they just put extra cars into the system without needing to call in drivers first and wait for them to arrive. Also it means that it’s feasible to provide small but frequent services which are very convenient from passengers’ perspective. Think our trains are quite large and they can move a lot of people per hour but the convenience factor isn’t there (you have to wait up to 15 minutes for a train in peak) where in Vancouver they come every 2 minutes.

    1. In Auckland drivers and their availability are not the issue, the availability of the rolling stock and trying to run things outside their allocated slot on the other hand.

      Auckland trains run at 10 minute intervals during peak on all lines except Onehunga and Pukekohe.

      1. Really? So how come on New Years eve the last train left britomart with hundreds of people still waiting on the platform. Why didn’t they just take six more sets out of stabling to run everyone home?

        1. It is up to someone in AT to run extra services and they then need to be on a bulletin as they will be running outside assigned slots something that doesn’t happen just because there are people that miss the last train.

          On NYE there was only limited services running due to track maintenance anyway.

        2. Nick as I said the issues are availability of the rolling stock and trying to run things outside their allocated slots.

        3. Bigted – how is rolling stock availability an issue at midnight on New Years Eve when they were running half hour frequencies on three lines and none on the Western line? There would have been heaps of EMUs parked up that could have been used.

        4. Yep that’s complete bullshit Teddy, plenty of stock and slots at midnight on NYE, in fact plenty of both anytime outside of about four peak hours on weekdays.

        5. There is no assigned slots, anything not planned and running outside of normal slots needs to be on a written bulletin and that is not something that can happen in a short time just because people miss the last train.

          The point that appears to missed here is it is not a lack of drivers so a driver less system would be of no advantage when people can’t make the last service.

        6. Bigted – so rolling stock is not an issue as you claimed it was in a comment above?

          TBH – this was not a surprising outcome with numerous trains coming into town throughout the evening but just one on each of the Southern and Eastern lines leaving after midnight, it could have easily been planned for. I think they were half hour frequency IIRC, with the last train leaving at 12:28 the timetable could simply have been extended to 12:58, and one extra train could have been brought in for a run leaving at 12:43. This could have all been scheduled in advance and would have minimal impact on operations.

        7. Ted, you just listed a bunch of things they never need to do on Vancouvers driverless metro system. They just put another train on with literally the click of a mouse.

        8. Jezza yes it could have been but it wasn’t, quite simple really.

          Nick that is one of the many differences between Vancouver and Auckland but as I had been saying that drivers are not the issue in Auckland everything else is.

        9. Planning to run the last service at 1AM is the problem. What kind of half assed operation is this?

        10. Bryce – it wasn’t even 1am, it was 12:28, probably only 20 mins after the fireworks had finished, I think the Southern line was 12:15 and the Onehunga line was around 11:45! AT really do beggar belief at times.

        11. Every new year for the past several years, Wellington has managed to run an hourly service on all lines except Melling right through the night!

      2. From my observations I conclude that Bigted is a train driver or belongs to train driver union of some kind. That’s the kind of vibe I get from his comments. So no surprises here really hearing that there is no problems currently… yeah right.

        1. Also wrong but I do work in real world transport and in the real world (not the fantasy world that many here live in) things happen the way they do for a reason and changing is not as easy as thinking up something and doing it.

        2. Real 20th century then. While we’re talking about 21st century transport (that is already running in other places), bigted is always behind…

        3. Clipboard operator based on another comment in a previous post, which pretty much fits the image I expected, the technical knowledge sounds too second hand to be a driver.

        4. So Ted’s maybe with AT or Transdev? Many comments seem to demonstrate the resistance we see to any suggestions to improvements with the service.

          What’s your motive for blogging here, Ted? Constructive dialogue or setting us all straight on how unrealistic we are with our thinking?

  7. Completely driverless roads are viable now. A spreadsheet shows enormous profits and no extra capital investment. The risks attached are far less than the risks with the present system. Less than 10% of the vehicles needed as they are all hired by the trip. No driver costs. Existing roads used as no risk of being hit by an imperfect driver…and all drivers are imperfect. Less damage to roads and cost of bringing roads to high standard easily covered.

    1. Are you advocating the banning of non driverless cars as I can see there being a good deal of resistance to people no longer being able to ‘drive’ the tin box themselves.

      People don’t spend $80k on a BMW to have it dawdle about autonomously stopping for cats and other things on the road.

      1. I have gone into deeply over the last couple of years. We are talking huge savings here of several trillion dollars a year for the usa which is thousands per person. Trump could clear the National Debt fairly quickly instead of increasing it. I accept it needs popular support but it should be possible to persuade people by giving them a fair part of the savings and convincing them that they and thieir families will be far safer. The excitement of driving a tin box is because it is dangerous for you and others. Not funny when your number comes up or that of a loved one.

        1. Does that utopian world with driverless cars also have no bikes or pedestrians possibly wanting to share the road in some kid of form?

    2. I’d love to see what Uber morning peak ‘surge pricing’ looks like in a world where the alternatives (walking/cycling/driving-yourself) are banned.

      Why pay for enough cars to meet surge demand when you can just make people queue. Kinda like how AT deals with March Madness.

  8. If fully autonomous cars are feasible, then driverless tram systems should be an order of magnitude easier to implement. It could be a real game-changer for public transit. Presumably existing tram systems could be retrofitted to run driverless, so even if Auckland invested in a driver-based light rail system now, there would be the potential to move to a driverless model in the future. The system picture looks like more of a “driver-assist” system – I wonder if a fully-autonomous system would need more sensors besides those pointed straight ahead?

  9. Rail’s own conservatism in regard to staffing practises makes rail and light rail its own worst enemy in regard to safeguarding a competitive edge. We only need to consider the delayed roll out of DOO in Auckland, and the whole door opening and closing system, and associated operating practises on our commuter rail network.

    Another key factor is a lack of understanding that rail and light rail are part hi-tech engineering led and part customer service led organisations. In the past a over emphasis on engineering led to a lack of awareness of customer service and general commercial acumen. In more recent years, customer service and a more business like approach has come forward in leaps and bounds. But a number of recent decisions in Auckland and on KiwiRail’s national network, indicates a lack of knowledge or awareness of the importance in correctly specifying technical engineering equipment. Add in a reluctance to address those long term staffing “habits”, and one can now witness the lost opportunities to boost long term productivity.

    Get the internal dynamics of the organisation right and driver less rail and light rail will come.

    1. I’m also beginning to think that the issues are cultural issues as much as technical, in considering the many insights of Bigted who, it appears, works in this industry.

      If so, these cultural issues need greater discussion.

      1. Yes there are cultural issues there. In many ways the front line staff have the best ideas and have had many years of running a safe network on the smell of an oily rag. But, they are often also the people who didn’t move on during various re-structurings, and who have the most “dug-in” entrenched attitudes, and least able to accept change. The greatest threat to rail and light rail’s long term survival comes from within as well as outside threats such as MOT/NZTA et al.

      2. That’s been my observation from these discussions over the last few months as well. I fully believe the trains are capable of doing exactly what we want them to do, it’s just that there is often inertia and risk aversion inside organisations, even more so if multiple organisations are involved. I see this at my work as well.

        The good news is that this inertia and risk aversion often breaks down when heat is applied from high above, I have seen this to my benefit recently. I don’t think we should underestimate the importance of Phil Goff specifically mentioning dwell times, ahead of thousands of other issues that bug Aucklanders on a daily basis, I doubt he has just said it publicly and left it at that.

Leave a Reply

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