A Next Step for Autonomous Buses?

Photo: David Wheatley

A fully autonomous bus is now in regular service in Scotland.  It still has employees, two in fact.  But if this technology works out, the ultimate goal is probably to run buses with no employees on board. In wealthy countries, the cost of running a bus is mostly the cost of the driver, so in theory, if and when all the bugs are worked out, a driverless bus could be far more abundant, for a given operating budget, than buses with human drivers can be.

That will be wildly controversial. I have mixed feelings about it. But driverless rail transit has existed for decades, starting in Vancouver in 1985. The lack of a driver is why trains in Vancouver come every few minutes even late at night. In emergencies people like for there to be someone in charge, but the voices coming over the intercom from headquarters often have a better picture of the situation than an on-board employee does. Driverless buses would definitely be part of a world where security is based more on electronic surveillance, and like many people I have mixed feelings about that.

But if we end up in a world with abundant and affordable autonomous taxis — still a big if — it will be very hard for cities to function without autonomous buses. When we remove the hassle of traveling by private vehicle, and reduce the cost, everyone will want to do it, and a city simply doesn’t have room for that.  The only other solution will be heavy decongestion pricing to make the affordable autonomous taxis less affordable, and/or bus lanes on every street so that buses effectively bypass autonomous-taxi congestion.  That may be the answer, but in the long run I’d rather see public transit be abundant, so that everyone can go places quickly in a space-efficient way.

20 Responses to A Next Step for Autonomous Buses?

  1. Jordi September 12, 2023 at 12:01 am #

    As an IT person myself, my sensation is that driverless bus (or even more, driverless tram) should be an easier problem to solve than driverless taxi. You only need to prepare the system for a specific route, and you can limit the corner cases a lot.
    I find it odd that driverless taxi has prototypes working at bigger scope. Is it because the companies involved see more market building and selling millions of cars instead of thousands of buses? Any security feature needed for buses that isn’t necessary for taxis? Any other reason I can’t think of?

  2. Sean Gillis September 12, 2023 at 3:56 am #

    Given that driver shortages are a major problem across North America (and staffing shortages in general), will this end up being as controversial as it might have 15 years ago?

  3. Jack September 12, 2023 at 9:14 am #

    Sean, that’s a great point. In my non-union environment, I expect the roll out would be so slow and gradual that no driver would be laid off–they’d be replaced by attrition. I think they’d all be fine with it. It’s probably different in a union environment where even a scenario that didn’t negatively affect the individuals–but did diminish the union–would be resisted.

    That said, I am 35 and I highly suspect I will retire (not early) with drivers still behind the wheel of transit buses. I just don’t see the technology being there except in highly ideal environments.

  4. Henry Miller September 12, 2023 at 2:08 pm #

    If the drivers unions do not resist we can/should retrain existing drivers. There is a lot of need for humans in a driver less system – in fact probably more need than now. Once we get rid of the drivers we can afford to run a lot more buses and that means we need more mechanics, more bus cleaners, and more people in dispatch. Without drivers we can afford to run 24×7 with the existing buses, we can probably afford more of them for more frequent service, and for running more routes. All 3 of those will increase ridership, probably to the point where a transit service needs more people than it does now. It will be easier to hire people though, because the shifts are not as bad (most of the new jobs can work 9-5, though of course some of them will be “graveyard”), most of them don’t have to fire you after a DWI (and of course with more transit there will be less DWIs as people who might drink don’t even consider driving)

    Maybe I’m too optimistic, but I think it is better for current drivers.

    • Dave September 13, 2023 at 6:46 am #

      “There is a lot of need for humans in a driver less system – in fact probably more need than now.”

      This cannot possibly be true. If it was true – if going to driverless cars and buses resulted in more humans being needed to maintain the transportation network than are currently being used – then there would be no financial incentive for all these companies to pursue driverless vehicles in the first place. Why would you create a fleet of driverless buses that uses as many humans to operate them as the current fleet of driver-required buses needs? The whole promise of driverless cars is freeing up most people from having to drive their own car… implicit in that promise is that we’re not all going to be freed from having to drive our cars so that we can spend more time acting as mechanics on our cars or manning the computers that control the highway network… so presumably the whole promise of driverless buses is freeing up most transit agencies from having to staff their own buses.

      • Dave September 13, 2023 at 7:05 am #

        I’ll give a real-life example: when the phone companies switched to automatic routing, they got rid of lots of operators who used to manually route your call across the country. We used to have 420,000 switchboard operators in the US in the 1970s, today with automated routing, there are only 4,000 operators left in the US. And there’s no way that 416,000 of those laid off operators went into other jobs like maintaining the phone lines or (nowadays) cell phone towers. The entire telecommunications industry – which includes telegraph, telephone, cable, and internet providers – used to employ over 1 million workers prior to the 1980s, but today only employs 637,000 people nationwide. It looks like the laid off telephone operators just evaporated when automation eliminated the need for their specific tasks. (Anecdotally, my aunt was an operator for Bell of Pennsylvania, then Bell Atlantic, and finally Verizon. She held on as long as she could, but her job was eliminated in the 1990s and she simply just retired early in her 50s, never working again. There’s no way she could’ve been retrained to climb telephone poles or cell towers, even if she was willing to!)



      • RossB September 13, 2023 at 1:19 pm #

        “Why would you create a fleet of driverless buses that uses as many humans to operate them as the current fleet of driver-required buses needs? ”

        Because you are operating a lot more buses. This is the nature of mechanization. As soon as Ford figured out the assembly line, it meant a lot fewer people were needed to make a car. But they soon built a lot more cars. There are other limits (as Henry mentioned) but overall, running the buses a lot more often *for the same amount of money* is a good thing.

        • Jack September 14, 2023 at 8:17 am #

          In my small system we have 46 drivers, 3 mechanics, and 2 dispatchers. That balance has worked well for years.

          Even if we quadrupled the amount of service on the road (a big expensive in capital and fuel even without drivers), and we matched that in mechanics and dispatchers, we would only need 15 staff in new mechanic / dispatch roles. In reality, our dispatchers have the capacity to handle a quadrupling of service with maybe one more position.

          So no, there will be far fewer paid positions in such a system.

  5. Dave September 13, 2023 at 6:41 am #

    Presumably, by the time these autonomous buses are ready to drive around with no employees at all on them, driver shortages happening this year will be over. (It doesn’t sound like fully-autonomous cars – much less buses – are going to become a real thing in the next 5-10 years, given that they currently can be easily disabled with traffic cones or loss of cell phone service, as seen in SF this year.)

  6. RossB September 13, 2023 at 1:28 pm #

    I believe Jarrett alludes to exactly what I’ve mentioned before in terms of automobile automation. The best analogy is self-checkout at the grocery store. Cashiers haven’t been eliminated, but there are far fewer of them. It is essentially a robot doing the work. We don’t think of it that way, because the change has been gradual, but imagine what it would be like for a modern Rip Van Winkle. Back in the day, the cashier looked at a price tag and pressed buttons on the cash register. They took the cash, and made change. Now machines do all that. The only thing a customer does is bag the groceries (which many of us were doing a long time ago). But it isn’t completely automated. If something goes wrong, then someone else will notice, and do what is necessary to fix it. Sometimes this requires stepping over to the register — sometimes it can be done remotely.

    I expect the same thing to happen with buses and trucks. They will be automated, but monitored by a team somewhere else. There will be people in various places, ready to intervene as necessary. The buses (and trucks) can also be remotely operated. Thus you have to have a lot of cameras (along with the other sensors) to see what is going on at any moment. Most of the time, the machine can do the work. Even when there are problems, they are usually fixed by remote control. Rarely do you need to send someone out there, although that option is available.

    Some people are nervous about the cameras. Personally, I’m not. Like it or not, people take pictures of us all the time. Sometimes, it is civil libertarians who insist on it (e. g. police body cameras). Google takes pictures of people all the time on the street. It is what folks do with the photos that matter. As long as reasonable safeguards are taken, I have no issue with cameras on buses.

    From a labor standpoint, I think it is actually good. Increased mechanization can lead to fewer jobs, but as long as you have a strong union, it can lead to better pay. A good example is longshoreman. As mechanization has increased (and the number of jobs have gone down) the salaries have gone up. It pays really well, as the work has become a lot more technical (it isn’t like people are just hauling stuff off the boats). I could easily see the same thing happen with buses (and trucks). There could be a substantial increase in buses on the street, but far fewer “drivers”, as the nature of their work changes. I see the result as being overall quite positive.

    I also wonder if it will eventually make for smaller vehicles in some areas. Right now, things are based on the cost of the driver. If that gets reduced by say, 90% (as each operator can handle an average of ten vehicles) then it might make sense to run more vans, and fewer buses, especially in areas that will never have that many riders, even if the bus is coming by every few minutes. My guess is you end up with a mix, with smaller vehicles serving less densely populated areas.

    • Henry Miller September 14, 2023 at 6:47 am #

      It is an interesting question how can cheap automated vehicles change transit.

      What I come up with is a couple electric mini-vans that run a short route down a cul-de-sac to a central station. The entire route is 2-3 miles, and the only purpose is to do the last mile. A mini-van is almost always plenty of capacity in a suburb. The central stations at 5 mile intervals have large cachement area with those small feeder, so you can then run express services all over (to other nearby central stations, the city center, and whatever else). Since the mini-vans are cheap you can run a lot of them and so give each cul-de-sac frequent service 24×7, and the central station has great transfers to useful service. This is the service level suburbs need to places where you don’t have to own a car.

      Of course the above is about less dense area. I do not claim it scales to the dense areas of your city (with larger vehicles it might, but I’ll let someone else think about that). I’m not sure how rural you can make it work, though it should scale down to small hobby farms just outside the city (remember that the rest of the city has such good service they don’t think about a car and thus the network effects bring in demand to reach exurbs).

      Remember the above is all a thought experiment. it seems like it could work, but anyone implementing it will need to work out a lot of details, some of which they won’t find until they start running it.

      • RossB September 15, 2023 at 4:30 pm #

        I think we are on the same page here. One of the big fundamental problems with transit in the United States (and Canada) is that there are so many people living in moderately low density areas. If the density was higher, it would be fairly easy to serve them. If the density was lower, there would be fewer people in that situation. These are the folks that could easily be served by vans, especially is they are relatively close to higher density areas. That is often the case, and simply the result of the strange zoning that permeates typical U. S. cities.

        Thus you end up with a mix. Vans for a large area, along with buses and trains connecting the more densely populated places. They complement each other, in the same manner as the buses complement the high capacity rail lines in Vancouver, BC. The trains are really not that extensive, while the buses cover most of the city. But the key is that they work together really well, forming an almost perfect grid (https://humantransit.org/2010/02/vancouver-the-almost-perfect-grid.html). The result is arguably the best transit system on the West Coast, and the best system for a city that size in North America. It isn’t always easy to build such a system, but it could be a lot easier if the cost of running a bus (or van) drops.

  7. Gaelan Steele September 14, 2023 at 5:51 am #

    I was on this bus on launch day. Through discussions with staff afterward, it was only running in autonomous mode for the motorway portion of its route, and even then the driver had to intervene to change lanes when a car was stopped in the road. I get the impression it was no more autonomous than most cars sold today.

    I suppose there’s the possibility the technology’s improved since then, but it seems doubtful, and certainly Stagecoach isn’t being upfront about what the technology can do.

  8. Stephanie Stout September 17, 2023 at 12:18 am #

    The ideal routes for first generation automated buses would be physically separated Bus Rapid Transit lines (if they exist) where BRT vehicles do not travel on ordinary streets and mix with erratic, random traffic. Houston’s first BRT line has its own lanes in the transit median of Post Oak Boulevard but crosses ordinary streets and terminates in bus transit centers at both ends, so the sensors and operating program would be more complex than a completely separated route. Perhaps a bus operator of a 99% automated route would simply be an emergency “fail-safe” operator and not be fully rated as a driver. Other routes might be fully automated for a substantial portion of the route but have pilots (like harbor pilots) that step aboard to drive the complicated portion of the route and get off at the entrance of the automated segment. Bus routes that are straight and simple as Jarrett Walker advocates would make better automation candidates than the meandering bus routes that complicate too many of our transit systems. OTOH, bus operators also have to assist, tie down, and unhook passengers with wheelchairs and reconfigure the seats in those accessible spaces.

  9. James Garner September 18, 2023 at 6:15 am #

    I think the whole transition will be bumpy, acrimonious, and will take place well after if should have. The key is part of the federal code. It used to be known as Rule 13 (c). Now it is know as 49 U.S.C. 5333 which basically says: “When federal funds are used to acquire, improve, or operate a mass transit system (public transportation), federal law requires arrangements to protect the interests of mass transit employees……The Department of Labor (DOL) must certify that protective arrangements are in place and meet the above requirements for all grants of assistance under of the Federal Transit Law before the Department of Transportation’s Federal Transit Administration (FTA) can release funds.”

    Currently, labor must be protected. So, transit systems are stuck with the costs unless they can negotiate out of it, or have a very labor-UNfriendly Congress and President get rid of 13(c). I’ve seen and/or worked with very few unions who will voluntarily let their authority weaken.

    Also, with the big complaints about security in the Covid and Post-Covid world, how will transit systems convince riders that Driverless Vehicles are safe? Yes, the buses could have attendants, but this is really no different than having operators, in terms of labor.

  10. TransitPlannef September 18, 2023 at 2:01 pm #

    In European bus systems regulations require that handicapped including wheelchair users have to be transported by a public transit bus. Usually the driver has to support boarding and unboarding of those passengers.

    It is strange to read about protection of labor in the US from European perspective. The shortage of drivers worsens from year to year and many German systems can only operate with drivers who previously drove buses in Belgrade or who grew up in Aleppo, Kabul or Accra and still there is a shortage, cause there are enough better paid jobs without working constantly changing shifts or simply living on welfare (called now „citizens money“).

    Self driving buses will also mean there are self driving cars available and that may even be the end of public transit as we know it today – unless private ownership of cars will be restricted.

  11. Henry September 19, 2023 at 9:37 am #

    Eh, I feel like a driver or some attendant will still be necessary for ADA compliance.

    The ramp requires a human to operate, and some wheelchair mechanisms need the driver to secure the wheelchair in the space.

    • Jonathan Monroe September 29, 2023 at 3:00 am #

      On London busses, the ramp is operated by pushing a button from the drivers’ cab, and weelchair users are expected to maneuver themselves from the ramp to the designated weelchair space (which is directly in front of it). The wheelchair is secured using its own brakes. The population of people who can’t do this but still have business going out unaccompanied is non-existent.

      In a world of automated busses, you would presumably need a swipe card reader by the accessible entrance or some other way for the weelchair user to activate the ramp without the driver needing to press the button. But I don’t see why wheelchair accessibility needs to be an obstacle to driverless busses – it certainly hasn’t been for driverless trains.

      Unfortunately, I can see the disability lobby becoming useful idiots for either the bus drivers’ union or the anti-transit micromobility grifters and making it an issue. The US definitely isn’t equipped to say no in this situation, and I suspect the UK isn’t either.

  12. Ramona September 29, 2023 at 3:32 pm #

    Another instance of technology used to disrupt working class livelihoods, especially for those who don’t have a college degree. Just like the elimination of Wal Mart greeters and gas station attendants in Oregon, this will be a disaster for those who need to put food on the table without a skilled degree.

    • Adrian October 1, 2023 at 7:13 pm #

      @Ramona … don’t worry, AI tech is and will displace lots of high SES white collar jobs/livelihoods too: just Google “AI endangered jobs” and see for yourself!