Excerpts from CNU Public Square

Ten rules for cities about automated vehicles

The adoption of AVs should not be allowed to replace time-tested places with something that would probably make our lives worse.
1) Be afraid
New technologies that increase convenience are unstoppable, whatever their impact on our long-term quality of life.
It happened with cars. Enthusiastic adoption followed by some dubious outcomes.
2) Be realistic (or, What to expect, when you are expecting AVs)
… major change is unlikely to happen for several decades … because most of the benefits of AVs really only kick in when we have full autonomy—a swarming fleet of shared vehicles that operates as a public good.
Heaven is fleets of shared electric driverless cars powered by renewable energy, and a more socialist—my word—wealth distribution system to help all of the drivers put out of work by driverless cars. Hell, on the other hand, is privately owned AVs, that, since they have nowhere to park, spend much of the day circling, doubling the traffic load.

3) Decide how much traffic you want
This is probably the key rule …
If (AVs) cut the cost of driving by 80 percent as anticipated, that’s supposed to add 60 percent of the traffic to city streets that are already at capacity. … But it’s actually worse …
If you ignore induced demand, 60 percent more trips are not a problem, as long as we have swarming. Elon Musk tells us that a driving lane full of swarming AVs can handle 3 times as many cars as it does today. So, problem solved, until you realize that, these days, traffic congestion is the principal constraint to driving.

Source: Walter Kulash

This becomes especially alarming when we realize that AVs will make driving cheaper in two ways: money and time. You will pay less per mile, and won’t mind sitting in gridlock as you work or watch cat videos.
The only answer, I believe, is to regulate it, not with laws, but with lanes. Without a commitment to limiting capacity, all our parking lanes, soon empty, will not become bike lanes and greenways as promised, but more driving lanes. And sidewalks will feel miserable.
The right solution, is to make the streets what you want them to be.
4) Plan for more sprawl pressure
Here was a Lyft ad, ostensibly for last-mile service complementing our rail network, but I’ll be darned if it wasn’t an exact map of suburban sprawl.

Only with the car did the entire landscape take on wasteful, unwalkable, disconnected forms that now, more than anything else, characterize American life.
But there is recent good news, which is that cities and towns have begun to figure out that sprawl does not pay for itself.
5) Understand transit geometry

There is no getting around the fact that low-occupancy vehicles are a tremendous waste of street space.  … Autonomous cars are a great supplement to transit, but, in congested places, they are not a solution to transit.
6) Don’t rob transit
This may be the greatest risk. In congested cities, replacing trains and buses with autonomous cars will cripple mobility. …
Unfortunately, just the prospect of future AVs is already threatening transit investment in certain American cities.
Meanwhile, Uber has set its sights on transit, and is offering UberPool monthly memberships at lower cost than a transit pass, some say to put the conventional transit out of business.
City leaders have the responsibility to teach their citizens about geometry—to teach them what even Elon Musk doesn’t seem to know—that a shift from transit to AV cars would reduce mobility.
7) Own the streets and own the data
It sounds implausible, but there is a very real worry that AV providers will ask to buy certain city streets, or certain segments of city streets, and cities will take the money.
Adam Gopnik: “Cities are their streets. Streets are not a city’s veins, but its neurology, its accumulated intelligence.” Never sell that.
Like Uber, AVs will represent a viable business model only by running on public streets. Sharing full data would seem a small price to pay for that privilege.
8) Don’t buy any urban vision that forgets urbanism
If you study modern urban history, you will see that every major transportation advance has brought with it whole new concepts of what the city is.

Ford’s vision of where people will enjoy walking with AVs.
Where people actually like to walk. Photo by Lynn Richards.

Traditional urbanism was not an invention, but evolved naturally in response to human needs. The adoption of AVs should not be allowed to replace it with something different.
9) Unify around a set of policy demands
Individual cities have little sway here, but working together cities can’t be denied. Cities can all create a protocol that to present to AV providers as a collective requirement.
10) Invest in the current technological revolution
A transportation technology available now has outperformed AVs by almost every measure.  … That technology is called a bicycle.  … Protected bike lanes are what you need to give a sense of security to civilian bicyclists.
Autonomous vehicles are the right answer to the wrong question. Why do MIT’s Media Lab, and Google keep asking how we can make cars better?
A better question is how can we provide the most useful mobility to the most people with the most positive outcomes for society? The answer includes cars, but also trains, buses, bikes, and walking—especially biking and walking.


  1. True AVs fully connected to each other are decades away. Throw in non-AVs and a hybrid mode city is the reality for at least another century. Add in humans desire to own ( and not share) and the shown socialist car sharing for all is not likely at all or certainly not before 2100.
    Interesting he did not mention road tolls at all. That is how you regulate use, ie an AV plus parking trip might be cheaper than a second & third trip empty (to park empty car and pick up user) if road tolls are high enough.
    Biking and walking makes sense for dense parts of cities, but not everywhere. A shared e-bike fleet ( also not mentioned btw ) makes a lot of sense for the last 1-2 km based on a rapid transit network.
    Downtown Vancouver could thus be entirely car free in 20 years if sufficient e-bikes or mini e-carts are available to rent by the minute. But that model would not work MetroVan wide but it could in certain dense pockets like downtown, west end, UBC, N Van Lonsdale area etc ..

  2. I understand that AVs are being designed to sense when a person or cyclist is in a vehicle lane then it will take evasive action to avoid hitting that person or vehicle.
    “Elon Musk tells us that a driving lane full of swarming AVs can handle 3 times as many cars as it does today”
    So we can be like Moses and part the traffic by just walking onto a road – all AV’s vehicles will stop! Or any vehicle or cyclist can emerge from a side street and drive straight into flowing traffic. There goes the projected road capacity in urban environments.

    1. That model will work only on a computer but not in the real world until ALL human operated vehicles are eliminated and all legal, technical and regulatory kinks have been ironed out. Maybe by 2100. Likely after that only.
      I can certainly envision highways or highway lanes that allow only AVs equipped with certain technologies, i.e. a controlled environment. Not everywhere.
      It’s a bit like speculating in the 18th century how video games or cellphone enabled parking meters operate. Pure speculation on yet unknown technology. Pure science fiction.

    2. When AVs are subject to full, life-cycle cost-benefit analysis and compared to other modes of transport, the central conclusion will no doubt be that public transit will always outcompete AVs on a per-capita basis in cities on cost, energy, emissions and system efficiency. The analysis would be incomplete if it didn’t acknowledge the cost of the public road system, its poor efficiency and return, and its effects on generating unsuitable forms of urbanism and a notorious environmental impact.

Leave a Reply

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