Skip to content

Destroying the city to save the robocar

January 18, 2018

By Paul Homewood


h/t Dave Ward


A thoughtful article in the Register about driverless cars:



Special report Behind the mostly fake "battle" about driverless cars (conventional versus autonomous is the one that captures all the headlines), there are several much more important scraps. One is over the future of the city: will a city be built around machines or people? How much will pedestrians have to sacrifice for the driverless car to succeed?

The battle over the design and control of urban infrastructure pits two distinct ideas against each other. One narrative of "networked urbanism" envisages the city driven by data analytics and networks controlled in part by machines. In this "smart city", technological solutionism is rampant, with everything connected and automated. This is Googleville: a posthuman urban laboratory.

You might expect the car makers to be happy with this as a future, but even here, car ownership and use may fall. Networked urbanism can be dressed up as faster, smarter and greener but it is still pushing the corporate panopticon into our streets and lives. The life in such a world sounds like that of the "Insiders" in Michael Frayne’s novel A Very Private Life – a life tended by the kindness of corporate automata.

The other vision of the city celebrates "walkable urbanism". This is gaining in popularity round the world. Detroit is copying the bike-friendly, walkable examples of Copenhagen and parts of the Netherlands. Earlier this year, urbanist Jeff Speck gave a terrific talk at the US Conference of Mayors called "Autonomous vehicles: the right answer to the wrong problem". The wrong problem is: How do we make cars better? The right problem is: How do we make cities better? And when it comes to cities, there are simple limitations of geometry. The real disruptor is the bicycle, not the robocar.

But the driverless car has to deal with pedestrians, as Christian Wolmar discussed at The Register last week: "The open spaces that cities like to encourage would end as the barricades go up. And foot movement would need to be enforced with Singapore-style authoritarianism."

Dealing with pedestrians safely is difficult, expensive, and culturally alien to the nerds building the cars; Melissa Cefkin at Nissan is a rare anthropologist in the business.

"The randomness of the environment such as children or wildlife cannot be dealt with by today’s technology," admits Volvo’s director of autonomous driving, Markus Rothoff. The driverless car can’t hear you scream. Tests are not being conducted in real pedestrian-congested conditions.

The cheat is: just get rid of the people around cars, so you don’t need to solve these problems.

‘Just a little tweak, here and there…’

The slippery slope starts with "modest changes" of course. Two leading artificial intelligence gurus, Google’s Andrew Ng, and Yyanqing Lin admit in a piece ominously titled "Self-Driving Cars Won’t Work Until We Change Our Roads—And Attitudes".

"Safe autonomous cars will require modest infrastructure changes, designs that make them easily recognized and predictable, and that pedestrians and human drivers understand how computer driven cars behave," they wrote in 2016.

There are reports of dedicated infrastructure already. A plan to transform Interstate 5 between Seattle and Vancouver, BC envisages a three-step process.

"Road signs and lanes disappear, with roadway intelligence built into vehicles. Highway lanes expand and contract automatically for high-traffic times," dreams John Jones, Fjord’s VP of design strategy.

Completely unfounded expectations of performance and safety being used to influence infrastructure, driverless cars will be able to travel safely bumper-to-bumper, advocates argue. And when that fails, as Wolmar points out, there’s always moral blackmail.

We can already see the pavement become political. "Sidewalks are often a hotly disputed space," this article on delivery robots explains. "People live in urban centers not because they want to sit at home in their house and have their toothbrush delivered to their door, but because they have a pharmacy around the corner that they can walk to," says Nicole Ferrara, executive director of pedestrian advocacy group Walk San Francisco. Moves to rid the streets of people are already under way.

Research has shown a lack of demand for autonomous vehicles – nearly six in ten Americans do not want to ride in one. MIT found that drivers, even millennials, want clever technology to help the driver, not replace them (PDF). Ultimately, driverless cars are part of the tech utopia that nobody wants. But rather than technology, it is money and momentum behind them; they keep the share price up in the face of Google and Tesla.

And it’s a utopia that may never happen. Jeff Speck reminds us that the predictions of full autonomy are decades away. "I would challenge anyone in the automated driving field to give a rational basis for when level 5 will be available," says Dr Gill Pratt, head of the Toyota Research.

If we want walkable urbanism (and we should), we will have to make a stand. ®

Brian Sherwood-Jones has 40 years’ experience in all of the Human Factors aspects of complex systems, helping to design and improve the safety of ships, helicopters and nuclear power stations. He blogs about usability and design, where a version of this piece first appeared.

  1. quaesoveritas permalink
    January 18, 2018 1:24 pm

    “Moves to rid the streets of people are already under way.”
    I have already noticed a tendency in the UK to downgrade the pedestrian.
    In news reports, people are not “hit by cars” any more, they “are in collision” with cars, or “collide with” cars, as if they were another vehicle.

  2. Derek Buxton permalink
    January 18, 2018 2:25 pm

    Where are all the asylums when they are required? Pedestrians are already at risk from bicycles, now they want us to fight off cars as well. We do not want a system that ruins what is left of our towns and cities but these ideas are so ridiculous as to be unworkable. Computers are not infallible!!

  3. Ian permalink
    January 18, 2018 5:19 pm

    In the UK, Sheffield City Region and Transport for the North are currently consulting on regional transport strategy and autonomous vehicles feature prominently. I’ll have to send them a link to this in comments!

  4. keith permalink
    January 18, 2018 7:09 pm

    Umm, equate this to the paperless office which 50 years ago was forecasted to be complete in 5 years, and yet we are still not there. So automated cars, if it’s as quick as the paperless office, we still probably won’t be there in 100 years.

  5. perkscan permalink
    January 18, 2018 7:35 pm

    So “the randomness of the environment such as children or wildlife cannot be dealt with by today’s technology” – at last the elephant in the room is being remarked on. Robot cars are software-driven and software can only react in a predictable way to situations it is programmed to deal with. Anything else and its response is indeterminate, which means it might do anything. So a 100% predictable environment is needed, either that or over the first years of its life expect many accidents due to encountering the infinite range of possibilities in the real world, in particular the ones which the programmers didn’t think of. That will result in many software updates of its software, each one due to finding a “bug”. Remind you of anything? Remember we have Windows version 10 now, not counting all the updates for every version in-between. Is that where we’re headed?

    • jim permalink
      January 19, 2018 3:20 am

      The program for driverless cars reminds me of Asimov’s 3 laws of robotics.
      Who will ever get in a driverless car that may be programmed to ‘sacrifice’ the life of its passengers if faced with a higher risk of danger to life of another road user?
      This is a complete nonsense, add your comments about typical software update problems, and we have a recipe for disaster.

  6. January 19, 2018 11:33 am

    Reblogged this on ajmarciniak.

  7. Gerry, England permalink
    January 20, 2018 4:52 pm

    Tesla already have the death of one of their owners on their hands when in auto mode the Tesla failed to ‘see’ a white trailer unit across the road in front of it. And they are also responsible for life changing injuries to a motorcyclist in Norway after she was run down presumably as the Tesla didn’t ‘see’ her either. Killing the occupant of an autonomous car is fair enough if you choose knowingly to ride in one, but killing others won’t be tolerated and should any vehicle that does so removed from the highway until such time as it can be shown to be safe in the same way that aircraft are stopped from flying following a crash. Legislation is going through Parliament that is supposed to deal with the issue of responsibility and insurance.

    As far as streetscape goes, the wonderful ‘shared space’ idea has already run into trouble as it doesn’t work for blind people as there is no reference for them as to where the ‘carriageway’ and ‘footway’ meet since there is no kerb-line. When a driver lost control of his vehicle in Exhibition Road, London it was reported that it mounted the kerb, except that it didn’t as Exhibition Road no longer has a kerb that might have slowed the vehicle down. As the article points out, with no kerb for reference how will the automated vehicles cope? Or will we see money spent on shared spaces and then more money spent ripping them out for the next looney idea.

    The article hits the main urban problem correctly – bicycles. Too slow to be mixed in with motorised traffic, with no ability to accelerate out of harm and difficult to see; too fast to be mixed with pedestrians; and often ridden with complete disregard to any traffic rules. Dedicated cycle lanes are an appalling waste of space as they are only used for a couple of hours morning and evening. At a recent public inquiry it was hard not laugh at the cycling lobbyists stressing to the inspector not to do a site visit during the middle of the day due to lack of cyclists.

  8. perkscan permalink
    January 20, 2018 5:14 pm

    Before software was permitted to be used in aircraft airworthiness people got together with the wider industry to formulate the basis on which software could be permitted as “safe” to use. That was then promulgated as a mandatory set of requirements which drove the aerospace software design, development and testing process. Where faulty software could actually kill people, the requirements were particularly rigorous. Does an equivalent exist in the motoring world? Or are governments just opting out and leaving it to the insurance industry to decide which products they consider safe enough that they might not lose money on? And given that the insurance industry probably doesn’t have the expertise to make those judgements up front, how many people will need to be killed before an insurer removes its cover?

Comments are closed.

%d bloggers like this: