When you talk about autonomous cars (or automated, or self-driving, or driverless cars), do you really mean vehicles? If so, say vehicles. The word cars (or taxis) explicitly excludes transit (as well as trucks and many other vehicles).
As a transit planner, I’m routinely told that we should neglect our transit systems, and certainly not improve them, because transit will soon be made obsolete by “autonomous cars.” There are very low density places where this is true, but it is geometrically impossible for dense cities, because there is simply not room to move everyone out of big transit vehicles into the tiny ones that the word “cars” implies.
However, the fantasy of cars replacing big transit vehicles can lead to serious dystopian outcomes, including higher Vehicle Miles Traveled, higher emissions, and higher exclusion of disadvantaged groups from opportunity. The fantasy is already encouraging neglect of transit systems and opposition to efforts to improve them, which is having all the negative impacts listed above, today.
None of these problems arise if the tech proponents would consistently speak of “autonomous vehicles.” In that case, the notion of autonomous vehicles replacing transit wouldn’t even arise, because it’s obvious that transit is made of vehicles too, and that the automation of transit vehicles would be part of any long-term automation project.
When I press an autonomous-car advocate on this point, they almost always say that of course autonomous vehicles will come in all sizes, including transit vehicle, etc. But they keep saying autonomous cars, which implies the opposite.
Remember, many people hate transit, love cars, and are generally OK with all the outcomes of mandatory car dependence, especially sprawl. These people love to hear that autonomous cars will destroy transit.
But if you talk about autonomous vehicles, you can promote autonomous technology without sounding like you’re advocating cars over transit, with all the problems for big cities that this implies. Why not do that? You’ll provoke less blowback from urbanists, and have more friends.
Very good point Jarrett. As you probably know, Australian cities are now in dire need of a massive revision of urban transportation policy and quite often the advocates of the status quo car centric arrangements will invoke the approaching driverless future as the great panacea for all our transport ills. Reminding them that the technology that will make cars autonomous will also be applicable to other vehicles may not necessarily change their thinking but it will blunt the simple equation of a driverless future with more cars.
The cost/benefits ratios will change among different types of transit vehicles. In general, autonomous buses will become cheaper to operate than low to medium capacity autonomous rail systems because driver expenses comprise a higher proportion of total operating expenses of current buses compared with current rail systems.
If the “first & last mile” function and most short trips are handled by car sized AVs, then express, frequent service longer distance, and commute hour buses could still have a place. Streetcars, lower capacity light rail, hybrid rail, etc. would have trouble being more then pricey affectations.
Maybe the most important factor in shared autonomous cars’ favor is that they are coming without any expectation of direct subsidies from local taxpayers other than what car owners already receive. That’s should be a pretty compelling argument in their favor, especially if operators are successful in getting customers to pool rides.
R. A. Remember that as transit evolves toward really attractive mainline services, bus and rail features converge. Buses become higher-capacity and are deployed with supporting infrastructure to ensure speed and reliability. If you focus on bus services that are intended to carry high ridership the difference is already much less. The problem is that the extreme diversity of bus services causes analysts to grab an average of bus performance that isn’t relevant to the situations where transit actually succeeds.
Unless autonomous rail gets there much faster than autonomous buses (and by that, I mean which year, not trip length).
Automating a rail vehicle, especially if it’s already in dedicated right of way, _should_ be a much simpler technical problem than automating a bus that has to operate in a very mixed urban environment. There’s just fewer degrees of freedom, 1 dimension vs. 2 dimensions.
Part of me is amazed that we’re not already there. And I wonder if it’s now only a regulatory/legal hangup.
There’s also an accessibility question. What happens on an autonomous bus if there’s no one to help the rider with a powered wheelchair board and strap it down? Will autonomous buses still have to be manned by an attendant?
Conversely, rail vehicles seem to be able to dispense with this (due to level boarding, smoother ride, smoother acceleration/deceleration).
Fully automated rail systems exist and have existed for some time. See Vancouver’s Skytrain as just one example.
Right.
I guess I meant automated rail more in the same way people talk about automated buses/cars. Sonar/Lidar to obstacle sensing, localized decision making when obstacles come up, etc.
“Part of me is amazed that we’re not already there. And I wonder if it’s now only a regulatory/legal hangup.”
they have to update the signaling first, and it’s a mess with human drivers already, but it’s happening.
This is for traditional rail of course, new underground rail systems have been autonomous for years.
It’s really a cultural war. When people are intoxicated by the oppressive californian ideology, even the most intelligent have a tough time getting the point.
This summer, I was blessed to have a looooong discussion(during a wedding) with a few “professeurs de classes prépas”(i.e. amongst the best teachers of mathematics & physics in France). One of them suddenly wondered if public transit was about to die. Because of autonomous cars. I made you point about geometry. Even they had a tough time processing the information. You don’t find better in maths than those gals & guys in France, a maths-loving country. Still, they had all been blinded by ideology, and could not make sense of “cars take too much place, you cannot replace the subway by autonomous cars, and hope things will work that well”.
I once had made a quick computation, probably rather unaccurate, and found that you’d need a 2*10 lanes urban highway to replace Paris’s subway line 1 by cars. Something unthinkable, of course. Even with that argument, they had tough time dispelling the dream of autonomous cars. They’re our best, and they’ve been brainwashed. Very, very, very worrying.
Considering that most of Paris is a stunning visual of Jarrett’s point about geometry, ‘blinded’ is right.
There are massive gains from automation of transit vehicles such as trains, trams and buses.
What is unusual is how few transit systems have converted over their train systems to full automation.
Automation is an established technology that has been around for more than 30 years now.
In most places, you just cannot fire drivers and hope to be reelected.
As an aside, perhaps the word “autonomous” needs a rethink. Usually, this term refers to “self-rule”, that is, something is autonomous when it decides for itself what to do. So, your autonomous car may decide to take you to Portland even if you say “Seattle”! Think of what a fleet of autonomous busses might do! The term “self-driving” seems more accurate.
Jarrett, how do you feel about automated trucks? There is a tv news story about Coors cooperating with Otto (a self-driving truck firm) on running self-driving beer trucks on routes between Colorado cities. Also this story on Medium.com: https://medium.com/basic-income/self-driving-trucks-are-going-to-hit-us-like-a-human-driven-truck-b8507d9c5961#.ib043xfbk
Correction: The beer company is Anheuser-Busch, not Coors.
Furthermore, I have discovered that Otto is a unit of Uber. So Uber is also getting into trucking.
With regard to self-driving car fantasies, a University of Hawaii civil engineering professor who has always championed roads over transit has made a bold prediction: Honolulu not only doesn’t need the heavy rail system it is building, but will not need a bus system in 20 years. Car-sharing will take care of all needs. Oahu Transit Systems (founded in 1971 as Mass Transit Lines after a strike shut down the privately owned Honolulu Rapid Transit Co.) has been awarded the APTA Best American Transit System Award twice (a record) and has been a Yelp favorite over the years will be uncessary in 20 years? I don’t think so.
With regard to self-driving car fantasies, a University of Hawaii civil engineering professor who has always championed roads over transit has made a bold prediction: Honolulu not only doesn’t need the heavy rail system it is building, but will not need a bus system in 20 years. Car-sharing will take care of all needs. Oahu Transit Systems (founded in 1971 as Mass Transit Lines after a strike shut down the privately owned Honolulu Rapid Transit Co.) has been awarded the APTA Best American Transit System Award twice (a record) and has been a Yelp favorite over the years will be uncessary in 20 years? I don’t think so.
It should be noted that for most (all?) of the automated rail systems currently operating, the goal of the automation is to accurately run trains at very high frequencies, not to reduce staffing costs. That’s why we see only the busiest most heavily trafficked networks automated, not little branch lines.
The jump from an automatically driven (but still human staffed and supervised) metro to all kinds of rail vehicles being fully autonomous is huge. To give an idea of the problems that would have to be overcome, there was a relatively recent case in Sydney where a train hit someone who had fallen from the platform at night. The driver saw an unknown object on the track approaching the platform, but assumed it was just a piece of rubbish until the train had become too close to stop in time. Preventing that kind of accident with automation is incredibly challenging.
Actually the systems I am familiar with run at high frequencies because of their low staffing costs
Actually this problem has already been solved 30 years ago. In Vancouver the only place you could possibly get to the track are the stations. Along the edge of the platform run 3 infrared laser beams at different heights. If you were to cross one at any time when there isn’t a train at the platform all incoming trains would immediately be stopped. It is impossible to enter the track area without crossing those beams. And the system works – at one time I was on a train when somebody entered the track area without authorization and the train braked immediately. Vancouver’s SkyTrain also hasn’t had a single major accident in its 30 years of automated operation. How many transportation system of any kind can make that claim?