Driverless cars – really?

An interesting article by Christian Wolmar, former transport correspondent of ‘The Independent’ and former Labour candidate in south London.

It concerns self driving cars – and I have to say I cannot disagree with any of it.

Addison Lee, a London minicab company, has effectively spotted an opportunity and now say they are going to be first to run driverless cars!

What? This is hype.

They surely hope – and why wouldn’t they? that they get adored by investors and as they are London based, they can ride on the concept with Uber.

Wolmar is worried about the actual concept and its implementation – with which I largely agree.

Who on earth is going to insure an automatic driverless car that, going too fast to stop, avoids a group of schoolchildren on a zebra crossing only to kill a man by driving into him on the pavement on the other side of the road? Or kills the driver himself by aiming at a wall?

That is either going to be an astronomical cost of business  (if it is algorithmic and automatic why on earth couldn’t it avoid both?) or it is not, in fact, going to happen.

Unless, perhaps, automatic driving will be conducted at walking pace?

It is interesting that while the Docklands Light Railway is simple and also automatic and driverless, no national railway system on earth is driverless.

Surely running on tramlines – as it were – is so much easier than road running – yet still it has not happened.

To me, Christian Wolmar is absolutely correct.

Meanwhile KPMG suggest that fifth most prepared country in the OECD for autonomous vehicles is the UK:

Thank goodness we are not the Netherlands!

Comments

  1. Andrew (Andy) Crow -

    Peter, you just don’t get the idea of driverless cars do you ?

    It won’t be very long before our descendent look back on this age of the motor car with utter disbelief.

    “They did what ?” They will ask . “They let people derive these things unsupervised, on the public roads ? It’s a wonder they didn’t keep getting killed in collisions, and pile-ups.”

    Think about it. I don’t have current British road kill figures, but they reckoned at the time that the total death toll of the twin towers attack amounted to three months US road deaths.

    The insurance companies will cope. The dilemma you paint is likely to be false, the autonomous car won’t be driving too fast, it will have better braking reaction times than any but the smartest most highly trained expert, and the dilemma of who to kill in extremis is no different for a human driver who hasn’t time to make a decision anyway.

    Self driving vehicles have already clocked millions of miles and the fatalities to date …? One highly publicised. How many on the roads in human driven cars ? Thousands. Year in year out. Publicised? Not at all. Our world’s guilty secret. Men, women and children. Economic consequences: incalculable. Human misery incalculable, and largely quite unnecessary.

    1. Peter May -

      Your’e right I don’t. The British ‘road kill’ (good, hard hitting, terminolgy!) figures are some of the lowest per head in the world, but still too high, granted.
      I’m not impressed with millions of miles driven, After all many HGVs and vans will have a million on their clocks and sometimes more. So a few millions isn’t much and these cars always seem to have a human driver too… I think the idea that we haven’t managed to automate rail but all of a sudden we will manage to do it for cars, which is a much less controlled environment is speculative and highly unlikely any time soon. Unless we’re all travelling at walking pace or possibly ‘hooked up’ on motorways.

      1. Bruce Gray -

        I believe the lack of automation in the rail system is driven more by economics rather than technology. The ratio of revenue from passengers or cargo to operator cost is substantially higher than motor vehicles, and the economic losses from accidents substantially lower. The lack of ROI on the capital investment necessary to fully automate the rail system is the more likely impediment.

        Here in the US we now have pizza delivery by autonomous vehicle (joint project between Google and Domino’s Pizza). No paid driver and no additional 15%-20% cost to tip the driver. Not sure that’s an overall benefit to society, but one can easily see the potential economic advantage to the customer and pizza vendor.

        I don’t think anyone can reliability predict the pace of technology development given all the variables and unknowns. Frequently there is a tipping point or breakthrough that accelerates the adoption well above earlier trends. Smartphones were more of a curiosity prior to 2007 (anyone recall the IBM Simon?) but 10 years after the introduction of the first iPhone, it would seem everyone in the industrialized world owns one, including children.

  2. Bruce Gray -

    The question is not whether autonomous cars will be in accidents, but rather would they be in accidents at a greater rate than human driven cars. The data so far, with relatively immature technology, shows accident rates about 25% less than human driven cars with much reduced severity (study sponsored by Google so take it as you may). Insurance companies will ultimately base their decisions on real world data and actuarial analysis, rather than industry hype or worst case scenario fear mongering from naysayers.

    Those opposed to autonomous cars paint a picture of driverless vehicles “speeding” around, or “going too fast” to stop, as if humans never did this, which is ridiculous. Autonomous cars are programmed not to exceed speed limits and to adjust speed based on prevailing conditions. The sensors are also functioning 100% of the time, and they don’t get sleepy, distracted, drunk, or exhibit road rage. The same cannot be said of human drivers.

    To think that autonomous cars should be 100% safe is naïve. Technology can fail and one can always devise worst case scenarios for driverless cars, but ultimately the question comes down to what would the statistical outcome be for autonomous vs. human driven cars under the same set of conditions over time? Of course, the ethical questions about whether to save the “passenger” or “pedestrian” still need to be resolved, but that shouldn’t be barrier to discovery and development. The technology is still in its infancy and performance will likely improve significantly as technology and the regulatory framework advances. Ultimately the success of autonomous vehicles will be based on the safety performance and cost efficiency in the real world rather than sensationalist stories in the media.

    1. Peter May -

      You may have a point about the economic return for fully automated complex rail, though I wonder still why it is economical for simple lines but not for complex ones, where one might presume the savings and possible reliability gains would be greater.
      I wonder is there a link for Google’s autonomous pizza delivery?

  3. Graham -

    I feel that autonomous cars are a distraction. Far more important is to get people out of their cars and onto public transport, bikes or even their own 2 feet. We need to stop people making the millions of unnecessary journeys that could be better made in more environmentally friendly ways.

    We bought an electric car earlier in the year, but even though it is zero emissions (while driving), and I can charge it at home utilising my pv panels there are still problems with pollution, if they are not charged by sustainable means, mining and environmental degradation associated with the battery technology. Hydrogen technology may be the answer, using renewably generated electricity, but without more investment and research the breakthrough to make it mainstream is some way off.

    The UK doesn’t have an integrated transport policy worthy of the name, rather it kowtows to private enterprise and the motoring lobby. We really need to start thinking about transport from the bottom up asking questions about, among other things, new road building, vehicles in towns and cities, cheap, reliable, accessible, frequent public transport for all, including so-called “remote” communities; how far people should have to travel to get to work; whether globalisation of transport where people and “things” (such as viruses, alien species) as well as goods and food can travel unrestrictedly across thousands of miles and of course the malign effects of car exhausts both on the environment and human health – to name but a few.

    1. Bruce Gray -

      I agree that greater priority should be placed on public transportation, but it doesn’t always work well for many people. Here in the U.S. the public transportation system is abysmal. I was recently on a two week trip to Scotland and managed to get to all of my destinations via public transportation; that would not be possible in U.S. unless you were confined to a certain urban areas.

      Most people are missing the point of autonomous drive. Long term trends in car ownership is downward due to the increasing cost to own and operate them, along with environmental impact considerations. Over 2/3 of new cars sales in U.S. are to people over the age of 50, and among millennials, car ownership is at an all time low for the respective age groups. As a result, ride sharing will continue to increase, especially for those with poor access to public transportation. Given the current trends, it is inevitable that automotive companies will look to offset lower sales by entering the ride sharing business and offering “cars as a service”. The only way this becomes a viable business model is through autonomous drive.

      1. Graham -

        Your second paragraph suggests that the move to autonomous vehicles is a (cynical?) attempt to restore car manufacturers’ profits. And of course the software companies hope to cash-in as well.

        I understand your point about public transport in the US, but it is pretty poor in many places in the UK as well. So do we just give up on it or argue for improvement?

        Another idea that was floated a few years ago was platooning on motorways. And another suggestion was high speed coaches operating on motorways and trunk roads from hubs at interchanges rather than in town centres. https://www.alanstorkey.com/a-national-integrated-coach-system-for-10-20-billion-in-five-years/

        Of course none of this has the “gee whiz” of having your pizza delivered autonomously, presumably by a drone for the final few feet through an open window.

      2. Peter May -

        Thanks, Bruce, for the links re the self driving delivery cars but these quotes “Ford will train self-driving car safety drivers locally” and “set aside his predecessor’s plan to have fully autonomous vehicles by 2021, indicating that this target was unrealistic.” leads me to think nobody is really there yet, but, as often, Ford are probably the people we should be looking to, rather than Google or Uber. Ford look to have a much better understanding to me – the delivery by driverless car would necessitate, not a drone, but going out to the car to collect via a PIN!

    2. Andrew (Andy) Crow -

      “I feel that autonomous cars are a distraction.”

      I’m not sure they are a distraction exactly, but I would agree that they are not the be all and end all.

      Public transport will be a major beneficiary of autonomous vehicles. The expensive and unreliable bit is the driver. Take him/her out of the equation and you could easily double the frequency of buses on a given route without doubling the running costs. Besides which more buses could be smaller buses, so reduced capital costs per vehicle. Lots of variables here for beancounters to have a field day with. Autonomous vehicles will completely alter the whole business of transport. Which, incidentally should be free. It makes no sense to add transport to the daily overheads of the population. (but that’s a very fundamental economic argument for another thread)

      Generating enough electricity to power all these electric vehicles is not a problem. We have the technology available albeit in its infancy and we won’t necessarily need nearly as many vehicles as we have now. Why own a car when you can dial one on your mobile and not have to pay the overhead charges when it’s not in use ? No need to park or pay charges for parking.

      Better frequency leads to better passenger service leads to more passengers using the service. Certainly the opposite applies and is the standard means of closing-down services.

      I can see luggage being a problem. I rather fancy that what we will need to make transporting our personal effect rather easier. Vehicles with a clip-on boot perhaps ? (Mini, even micro, container transport systems)

      Freight can equally well be shifted in autonomously driven vehicles.

      We are looking at the prospect of a lot of (hundreds of thousands) jobs disappearing and the biggest challenge we will face is how these ‘spare’ workers are going to constructively spend their time.

      The solutions will come from fundamental restructuring. Market economies don’t do social infrastructure so until we change our views on political economy we are not going to be able to make these changes.

  4. Robin Stafford -

    With the benefit of a modest amount of inside knowledge, it’s Ford and a number of the big, traditional internal combustion based manufacturers who are in deepest trouble. That’s because EVs and AVs are going to turn their business models up side down and require very different skills. A few points:
    EVs are far easier to make and far more reliable. Knowing how to make an internal combustion vehicle is of little use. It also removes a big source of revenue for their dealerships.
    The best guess is that far fewer people will buy and own their own vehicle but will call up a vehicle as and when needed. Uber are clearly in the lead when it comes to the logistics of managing fleets of vehicles and getting them to the customer.
    Whilst public transport works well in urban areas (and I’m a fan of it and a cyclist) it’s a different problem in rural areas where it tends to be dire. Just being able to call up an AV, managed Uber style, to take you the 5-10 miles to your nearest town to shop or see your doctor, at any time would be hugely convenient, especially for those who can not afford their own car. What if that Uber service was publicly owned…?
    Then the mapping and tracking that is critical – who are the deep experts on that? Google of course and no-one else really comes close.
    As for the AI, rhere are certainly some challenges that have not been fully thought through or solved such as insurance and accountability for accidents. As far as safety goes, the question is maybe whether AVs are safer than human drivers, whose track record as others have said is not good.
    I’d be worried about seeing them introduced gradually in a way that allows the problems and unanticipated consequences to be addressed. I’d also worry about the emerging monopolies such as Uber and whether those should be public services in future instead of or alongside existing public transport.

  5. Peter May -

    Thank you for some of those informed comments.
    However I’m not convinced that public transport works well outside of London – where it has easily the biggest subsidy per passenger and is sprecified by the Mayor! I believe over half of all UK bus miles are in fact in London.
    also I’m not convinced that ‘Uber are in the lead…’ I think they are in the lead only on a recast of Victorian labour exploitation. I prefer Ford, who’ve been through that and now know they have to innovate to survive (as you suggest).
    If only Britain had the ability to remain part of the EU satellite/mapping programmes we wouldn’t have to rely on the US defence industry which Google seems somehow to have monoplised …
    However I entirely agree that such services should be for public use not private exploitation.
    And also I do think there are problems with calling up any Uber – private, electric or artificially intelligent! – down some Devon lane 15 miles or more from the nearest populous destination.
    I can see how that might be useful but not how it might be profitable.
    Yet down that Devon lane they are probably producing our food!

    1. Andrew (Andy) Crow -

      “I can see how that might be useful but not how it might be profitable.”

      They neoliberal types ‘ave gotten ye by the short and curlies then, Peter.

      The be-all and end-all is can we make a profit ?

      Oh dear, Oh Dear.

      The profit in transport is to get goods or people from where they are to where they want or need to be.

      If you insist on it being financially profitable all you do is burden the system with something it has no need to be bearing.

      It has to be centrally planned and managed to be in with a cat in hell’s chance though. And costs (if there be passenger costs) that are not punitive in Devon (or any of the rest of a country which is predominantly sparsely populated and rural.) Rather as the Royal mail used to have just the one stamp price for any address in the country. It would, naturally be a little more complicated than the postal service but all that worked for years before we had clever machine computers to do the hard work of sorting and arranging data.

  6. Peter May -

    Point taken, although I was actually thinking of Uber’s profits (I’m all heart) rather than the absolute need to be profitable in general. I can see Uber just creaming off the most profitable bits although I suppose we could make their licence conditional on living up to their name!

Comments are closed.