Self-Driving Cars

I am probably not in the majority in finding this wholly unsurprising:

THE GOVERNMENT revealed Thursday that a Tesla Model S crashed into a truck in Florida in May, killing the electric car’s driver.

In the Florida case, the car failed to detect a large truck that had crossed into the Tesla’s path, perhaps because it blended in with a brightly lit sky.

A lot of people who are not engineers, and a lot of people who are, have a touching faith in the ability of technology to solve every problem there is.  People tend to look at technological progress in one area and assume that it can be seamlessly transposed into something entirely different provided enough minds are working on it.  This is why people are so optimistic about self-driving cars: they look at the amazing advances in computer power in the past few decades, they see Google has a huge stack of money and a very clever search engine, and conclude that self-driving cars are just a matter of time because…well, technology, innit?  And the same people often think it is self-evident that computers will always do a better job than humans as the former are infallible.

Personally, I understand enough about how things work to know that technological progress in any given area is not inevitable, there must be a mechanism in place for the shift to happen, e.g. a step-change in technology in the vein of the PCB or transistor.  When you consider how long the humble motor car has been around and the hundreds of millions of manhours that have been spent in trying to improve it in every possible way, it is astonishing how little has changed since the Model T Ford.  The basic principles of how a car is powered, controlled, and physically laid out haven’t changed.  They even still have wing mirrors and a driver’s rear view mirror.  So much for technology.  There have been plenty of improvements and enhancements, but no step-change in motor car technology since the first one rolled off a mass production line.

Google reckon they can make his step-change by doing away with the driver, and everyone seems to be confident they, or somebody else, will be successful in doing so.  Why I’m don’t share their confidence is because of two technical reasons: the first, which I’ll write about at length in a separate post, is the cost of manufacturing, testing, and maintaining extremely reliable electronic systems.  The second is that I do not believe computers will ever be as good as humans at driving in the environment in which humans live.

The mistake people make is to assume every action in driving is one of simple measurement, and conclude that computers are far better at measuring things than humans are in terms of speed and accuracy.  However, driving is often about judgement as opposed to pure measurement (and this is why it takes a while to become a good driver, judgement improves with experience), and much of this judgement relates to the interpretation of visual information.  The recognition of objects by computers is still only in its infancy, and nowhere near robust enough to deploy in any safety-critical system.  Given the pace of development of other areas of computing abilities, such as sound recognition in apps like Shazam, object recognition is seriously lagging behind and I suspect for very good reasons: software, being made up of pre-programmed algorithms, simply isn’t very good at it.  And even then object recognition isn’t enough, a self-driving car would need to be able to not only accurately acquire visual data but also interpret it before initiating an action (or not).  Computers are unable to do this for anything other than the most basic of pre-determined objects and scenarios, while the environment in which humans operate their cars is fiendishly complex.

There are those who think that advances in computing power will solve this issue, but I think the problem of visual data acquisition and interpetation is one more akin to aesthetics than measurement, i.e. its a judgement, not a binary decision.  Are we confident a computer will one day be able to write a decent novel?  Or generate a picture which is not a pre-programmed mathematical model which the coder knew in advance produces nice shapes?  With enough computing power, do we believe a computer could write a better song than a human could?  Personally, I don’t think this will ever happen because so much of aesthetics is down to judgement and involves variables which cannot be properly defined, much less defined in advance in a piece of code.

I believe a human’s ability to determine at a glance that an object in the road is a shallow puddle and not a large rock is the same ability which can differentiate between an operatic aria and a pet shop on fire.  Computers don’t have this ability, as the failure of Tesla’s to tell the difference between a large truck and the sky shows.  What does amaze me though is that computers are being put into cars with the belief that they can do things they demonstably can’t.  A hefty lawsuit and tighter regulations can’t be too far away.

If self-driven cars have a future, I believe they will take the form of manually-controlled machines which switch to self-drive mode only once they are driven by a human onto a very tightly controlled and sterilised environment such as a motorway specifically designed to take only self-driving vehicles.  I am confident we will never see self-driving vehicles moving around cities and towns as we currently know them, ever.

Share

23 thoughts on “Self-Driving Cars

  1. Nah – there’s too much money in it, and/or Silicon Valley is too hubristic, for this not to be foisted onto the public, or at least for there not to be a very strong attempt to do same.

  2. Thing is, I have yet to meet anyone, IRL, who will admit to wanting to buy one.

  3. I have no strong view on the subject. But I can remember spending undergraduate days being just around the corner from an AI department that swore that quite soon AI would revolutionise everything. The decades have sprinted past, and yet it hasn’t happened.

    A second tale: long ago a petrochemical company wanted to hire me so that I could propose how it should revolutionise its manufacturing processes in the same way that cars had recently been revolutionised. I first offered the opinion that cars had not been revolutionised, and that what I’d learnt about them in my teens had not changed in any profound way – not the engines, not the gearboxes, and not the steering. The replacement of drum brakes by disc brakes was, I opined, an advance not a revolution. Ditto the sequence magneto, dynamo, alternator. I suspect that their notion was empty-headed: that someone had been struck by how much more comfortable cars had become, and had extrapolated from that.

  4. @Dr Toboggan,

    I quite agree, these things – like electric cars – will be shoved down our throats, until enough pile-ups and lawsuits convince our betters otherwise.

  5. @ dearieme,

    I first offered the opinion that cars had not been revolutionised, and that what I’d learnt about them in my teens had not changed in any profound way – not the engines, not the gearboxes, and not the steering. The replacement of drum brakes by disc brakes was, I opined, an advance not a revolution. Ditto the sequence magneto, dynamo, alternator.

    Exctly. I was going to expand on this in my post but felt readers would get bored.

  6. Not the best example to make your point with as that vehicle was semi autonomous. I am with you on the quantum leap bit and I love to and continue to laugh at those that have predicted for over a hundred years that battery powered cars would take over from a combustion engine.

    Yes the liability and insurance issue is a show stopper alright, driving a car is the single most risky thing we do every day and once we get a bit of momentum up and mix it with humans or other heavy objects with the opposite momentum getting in the way there is a very high potential for injury, lawyers, damage costs, loss of freedom, insurer investigations and general mayhem to be instantly created, this is the elephant in the back seat of a driver less car. Plus the learned knack of judgement hasn’t yet been created in a chip, I don’t believe that they have worked out how to read body language yet which is 90% of human communication either. That’s why computers cannot negotiate.

    If you look at the areas where fully autonomous vehicles are working it is in controlled and fully mapped areas such as mines where driver less trucks are working very well without incident. Not only is this a tightly controlled environment and workplace (which is not like our cities and towns) it is a cost saving initiative where one person somewhere else can “drive” many trucks on a far away mine site, they call this concept “one to many”. So we do have a proven business case here where one guy sits in an office and drives many trucks reducing fixed business costs which is not the case with your average driver like say me. When I drive around, something I dont class as chore by the way, its just me and the one to many business case doesn’t work in this instance. This to me is the biggest reason that we wont see a shift in this direction in the mid term. I wouldn’t go as far to say ever though as that is a long time.

    I do think that we will see the current successful driver less business models be extended wider in similar industries and in differing modes of transport such as rail, sea and air (drones). Next it could be expanded into dedicated mass transport systems. But until folk like me are prepared to sit next to some random person, and hand back the keys to our stretched Hummer then it will be a very long time before a driver-less car replaces the traditional passenger-less car.

  7. I suppose it might depend on where you begin. When we lived in Oz it was noticeable that they were lousy drivers. Maybe the robots should start there.

  8. dearieme: it’s notable that many of the current self-driving car efforts are based in the Bay Area where the average standard of driving is several notches beyond “execrable”. Making a robot better than that doesn’t seem to be terribly hard.

  9. Did some Tesla S self drive in Hong Kong recently, at 70kmh through undersea tunnels. The sensors picked up other cars, road curvature and edges easily as a tunnel is well defined & lighting uniform. Not sure how well this would work driving through Korsakov with porous road edges and additional lanes appearing willy nilly.

  10. Heh, heh. Where we live a major cause of death in road accidents is drowning. Maybe self-driving cars should be built with positive buoyancy.

  11. Nobody is going to convince me they have seen worse driving than what I witnessed daily in Lagos.

  12. The sensors picked up other cars, road curvature and edges easily as a tunnel is well defined & lighting uniform.

    Interaction with other vehicles is relatively straight forward. It’s the interaction with pedestrians and other unexpected obstacles that is tricky. Presumably there were no toddlers wandering along the pavements in these tunnels?

    Not sure how well this would work driving through Korsakov with porous road edges and additional lanes appearing willy nilly.

    Heh! They’d not managed to put up decent road signs last time I was in Korsakov, and there were no road markings. It might be a while before self-driving car infrastructure is installed.

  13. Where we live a major cause of death in road accidents is drowning. Maybe self-driving cars should be built with positive buoyancy.

    Scares the hell out of me, that. It seems all the electrics stop working when the car submerges, meaning you can’t even open the door. I’m not sure what the recommended course of action is here.

    When I was working in Sakhalin, one of the Russian workers on the LNG construction site managed to roll his vehicle into a large pond of water. The car sank and he found himself unable to open the doors or electric windows. Fortunately, Russians aren’t prone to panic and this chap was a diver and so coolly hopped into the back where the windows were manual and escaped. I’m sure most people aren’t as cool-headed or lucky.

  14. “Or a firearm”: only if I want to deafen myself by discharging it in a confined space.

  15. only if I want to deafen myself by discharging it in a confined space.

    Dunk your head under the surface before pulling the trigger. 🙂

  16. While i am pessimistic for sure on autonomous vehicles until legal/insurance issues are resolved (and the issue of the car stopping when it a child is in front of it, versus stopping when it sees car exhaust at -10 C which is an opaque cloud), technology will deliver us some form very soon.
    Welcome to the “revolution” via AI in self driving cars: It’s here and it’s real:
    https://blogs.nvidia.com/blog/2016/05/06/self-driving-cars-3/
    Volvo is going to soon start accepting applications for 100 drivers to test 100 self driving cars on specific Gothenburg Sweden freeways in 2017..
    http://www.volvocars.com/intl/about/our-innovation-brands/intellisafe/intellisafe-autopilot/drive-me

  17. Welcome to the “revolution” via AI in self driving cars: It’s here and it’s real:

    Let’s see:

    The vehicle drove along paved and unpaved roads with and without lane markings, and handled a wide range of weather conditions. As more training data was gathered, performance continually improved.

    Nobody is saying cars cannot self-drive by taking measurements against fixed and moving objects which it is taught to see. The question is whether computers will ever be able to replicate the ability of a human to obtain and interpret visual data and make a sensible judgement. Roads filled with cars, trucks, buses, trams, pedestrians, animals, and a million other objects are full of surprises: computers are generally not very good at dealing with surprises.

    If the technology behind self-driving cars was so advanced, we’d be seeing driverless trains being rolled out as standard by now: a railway track is a far more controlled environment than a road. Yet the only places we see them are on a few metros (Line 1 in Paris, for example) and airports.

  18. Plus the learned knack of judgement hasn’t yet been created in a chip, I don’t believe that they have worked out how to read body language yet which is 90% of human communication either. That’s why computers cannot negotiate.

    Good point that, Bardon. A surprising amount of interaction between vehicles is based on body language.

  19. Hello cos999, I don’t doubt that the technology is being developed but it wont fly until there is a business case. Like I was saying there has to be a cost saving over teh status quo or it will not penetrate an existing market.

    The driverless trucks are successful because they have the “one to many” cost savings benefit ie one driver is now driving a significant number of trucks at the same time which is a cost saving or increased profit, whilst driverless cars are fantastic technological achievements, until there is a bob in them they wont fly. I am sure there is a “one to many” solution out there and the person that finds it will be extremely wealthy but until then it just wont fly.

    I found this very short video below where business people explain the business case for driverless trucks which in a nutshell is reduced costs and the analysts talks a bit about the “one to many” benefit. This benefit is not yet apparent for driverless cars, assuming that we do eventually get them.

    http://www.abc.net.au/news/2015-10-19/remote-control-trucks-the-start-of-automated/6865026

  20. “Volvo is going to soon start accepting applications for 100 drivers to test 100 self driving cars on specific Gothenburg Sweden freeways in 2017.” How long before the Guardian claims that they’ll be programmed to run down Syrian Refugees?

Comments are closed.