back to article Uber's disturbing fatal self-driving car crash, a new common sense challenge for AI, and Facebook's evil algorithms

It’s been a grim week for AI. The deadly Uber crash and fallout from the scandal between Facebook and Cambridge Analytica are a reminder of the ways algorithms can fail, and how they can be used against us. Fatal Uber self-driving car vid - The video footage capturing the last moments just before one of Uber’s self-driving car …

  1. Ian P

    LiDAR doesn't work in the dark? WTF?

    "It’s frightening that the LiDAR onboard didn’t detect her either, most likely due to the dark conditions "

    Can somebody tell the software engineers that people quite often drive cars when its dark.

    1. Roland6 Silver badge

      Re: LiDAR doesn't work in the dark? WTF?

      The footage also makes clear two other important factors:

      1. The car was driving on dipped headlights under conditions where the full beam would have been appropriate.

      2. The car was driving at a speed where it was unable to stop within the distance lit by its dim and poorly adjusted headlights.

      So whilst we do have to ask what the rider was doing to not notice the car, it is clear some basics of driving have been missed by the Uber developers.

      1. ' DROP TABLE users;

        Re: LiDAR doesn't work in the dark? WTF?

        >>"So whilst we do have to ask what the rider was doing to not notice the car"

        Gonna need a (spiritual) medium for that, unfortunately

      2. jh27

        Re: LiDAR doesn't work in the dark? WTF?

        Someone else has probably already said it but... a video camera might have a dynamic range of 10 stops, compared to about 20 stops, for the human eye. The reason the road looks so dark is because the headlights are bright and the aperture of the lens is adjusted so that the centre of the field is not over-exposed. If the headlights had bean turned off, I suspect that the woman would have been visible on the video recording much sooner.

        However, this is all irreleveant, because the car (hopefully) wasn't using the footage shown, it was almost certainly using LIDAR - it is interesting that Uber have not been forthcoming with that data, only with the video blaming the driver* and the misleading passive camera footage.

        * don't get me wrong the driver isn't blameless, but the software developers are ultimately responsible.

    2. Anonymous Coward
      Anonymous Coward

      Re: LiDAR doesn't work in the dark? WTF?

      The only clear point the video shows is thid accident was going to happen driver or driverless.

      No way would a human behind the wheel have changed the outcome.

      It just highlights that whilst driverless cars will for sure reduce road deaths, it's never going to get to the magical 100%. It with current technology.

      1. Anonymous Coward
        Anonymous Coward

        Re: LiDAR doesn't work in the dark? WTF?

        > No way would a human behind the wheel have changed the outcome.

        If I had a dashcam for every time I've proven you wrong....

        Driverless cars MAY ultimately reduce road deaths, but will the lives they save be worth living? They WILL monopolize personal transportation, milking riders for money, making the super-rich richer, and stranding even more lower/middle class people in slums. They're a wet dream for accidentally-wealthy binge-drinking douchebags. Also, they're beyond Orwellian.

        OTOH, if self-driving cars stop for any asshole who runs out in front of them, expect a MASSIVE rise in road robberies, kidnappings, rapes, and murders.

        1. Dodgy Geezer Silver badge

          Re: LiDAR doesn't work in the dark? WTF?

          Tractors MAY ultimately reduce farm deaths, but will the lives they save be worth living? They WILL monopolize equipment haulage, milking farmers for money, making the super-rich richer, and stranding even more agricultural labourers in slums. They're a wet dream for accidentally-wealthy binge-drinking douchebags. Also, they're beyond Orwellian.

          And don't get me started about building houses to live in rather than inhabiting nice dark caves....

        2. Grooke

          Re: LiDAR doesn't work in the dark? WTF?

          "OTOH, if self-driving cars stop for any asshole who runs out in front of them, expect a MASSIVE rise in road robberies, kidnappings, rapes, and murders."

          You don't stop for people who run out in front you?

          "If I had a dashcam for every time I've proven you wrong...."

          Given the above... how?

          1. Anonymous Coward
            Anonymous Coward

            Re: LiDAR doesn't work in the dark? WTF?

            > Tractors MAY ultimately reduce farm deaths, but--

            Try harder, smart-ass. You replaced "self-driving cars" with "tractors". If you want to talk about self-driving tractors, be my guest.

            > You don't stop for people who run out in front you?

            I honk, swerve, brake, and sometimes I even stop - i.e. for little kids - then I yell at them if it's a busy street. But I try very hard to avoid stopping for people who deliberately try to stop me. Particularly in shithole neighborhoods. Yes, I have seen some shit.

      2. Paul Crawford Silver badge

        Re: LiDAR doesn't work in the dark? WTF?

        The other WTF you missed was that even when the pedestrian came in to camera view the car did not brake even though an AI system should have a reaction time an order of magnitude faster than a human, and should out-brake most humans by knowing full well what is the lock-limit of the wheels (i.e. reaching anti-lock point),

        Also as others have pointed out the car was clearly driving beyond its sensor range, again that is something humans do, but is actually against the highway code. Drive to your visible range stopping-distance we are instructed.

        So who is going to fall for this: The software engineer(s) who developed the sensor/stopping code? Those who did a safety analysis? The executives up top?

      3. Mark 85

        Re: LiDAR doesn't work in the dark? WTF?

        The only clear point the video shows is thid accident was going to happen driver or driverless.

        No way would a human behind the wheel have changed the outcome.

        It just highlights that whilst driverless cars will for sure reduce road deaths, it's never going to get to the magical 100%. It with current technology.

        I'm not sure why the downvotes. Based on the video just of what's happening on the road, I tend to agree. The distracted driver doesn't help. Black clothing at night on the roads is a serous problem. I agree that a human driver would have had issues also.

        But are there other factors at play here... was the car going too fast for conditions? Even humans drive too fast. There's just too many unknowns here for one to make a judgment call from watching a video on a computer screen.

        1. Mark 85

          Re: LiDAR doesn't work in the dark? WTF?

          I'm going to add to my previous... The video itself. Is this a recording from the LiDAR or from a separate system? IF it's the LiDAR feed it's a piss-poor camera.

          I'm not a fan of Uber and rather despise them due to their manglement's attitudes and actions. But I'm trying to be fair as to this. My local roads are a bitch to drive at night due to the inconsistent and uneven lighting.

          1. Adrian 4

            Re: LiDAR doesn't work in the dark? WTF?

            It's not a recording from Lidar. Lidar provides something more like a point-cloud or just a single flat plane. It doesn't produce recognisable images. It produces something far more useful : an accurate array of distances in multipole directions.

            It also works better at night, because the sensor is less often overpowered by sunlight. It's true that a black object (something with low reflectivity) will not be detected at long range, but a decent lidar should manage at least 30m for any normal clothing.

            It's more likely that the decision to brake is based on a mixed weighting of visual detection, lidar and a number of other factors, and that the algorithm used chose to reduce weight given to the lidar. Or it was misaligned and didn't see her at all.

            1. Aitor 1

              Re: LiDAR doesn't work in the dark? WTF?

              That is what I think. The program weighted the probabilities, and decided that no way that thing could be a woman crossing.. as it was in "road mode", and no humans/animals would cross.. and also, no detection on regular video.. so yeah, that must be an error in the lidar.

              Had than been a deer, well, the car would have crashed too.

          2. Dodgy Geezer Silver badge

            Re: LiDAR doesn't work in the dark? WTF?

            Lidar uses pulsed light, usually ourside the range of human eye detection. If you want to know what a Lidar actually sees, look here:

            https://cdn-images-1.medium.com/max/2000/1*mtA31F_lby5QCP8gYZQyTw.gif

        2. david 12 Silver badge

          Re: LiDAR doesn't work in the dark? WTF?

          >Based on the video just of what's happening on the road<

          (1) The video is not relevant to the LiDAR

          (2) The reduced dynamic range of video gives a very poor representation of what the eye sees in the dark

          (3) Actually, as shown elsewhere and reported by people who drive there, it's not that dark anyway: it's a well lit main road in a busy district.

          (4) Because the video is so dark, you can't see that the district is actually built up, with crossings, includes places for pedestrians to cross the median.

          (4a) Which explains why the woman is on foot: she's crossing as a pedestrain at a median break, so that she doesn't suprise drivers expecting pedestrians, not bikes.

      4. Paul 195

        Re: LiDAR doesn't work in the dark? WTF?

        > No way would a human behind the wheel have changed the outcome.

        I strongly disagree. Even in the low-res video you can clearly see "something" before the cyclist comes into view, but there's no indication that the vehicle starts to slow even at that point. Here in Berkshire, we have many suicide cyclists who ride around in the dark, dressed in black with no lights. One night driving down an unlit road, I could see ?fireflies? twinkling in the distance. I worked out that they were the reflectors on a pair of pedals going up and down and was able to slow down enough to avoid the cyclist *before* I ran him over.

        AI is not intelligent because it still only understands what it's seen before, and doesn't yet appear able to put together a hypothesis like the one that enabled me not to kill a cycilst. Personally, given that self-driving cars still can't cope in the relatively benign environments they are being trained in, I think we are decades away from genuinely autonomous vehicles.

        1. melcom

          Re: LiDAR doesn't work in the dark? WTF?

          I agree. Humans have inputs, even though their mechanism maybe not fully understood, that I can't see AVs matching in the foreseeable future.

          Many years ago, on a dark, wet night, I was driving my 6V VW Beetle (the one with the candle-power headlights) along an unlit part of an almost deserted highway. Maybe at around 60kph because of the terrible conditions. I got a sudden, urgent intuitive prompt "STOP". I immediately slowed to walking pace and out of the mist in front of me loomed a cyclist wearing only (wet) black, no lights, no white showing anywhere. He had possibly been crossing the road in front of me. Had I not heeded the prompt, I'm certain I would have got him.

          I bet many other readers have had comparable experiences with intuituve flashes.

        2. mrdalliard
          Meh

          Re: LiDAR doesn't work in the dark? WTF?

          AI is not intelligent because it still only understands what it's seen before, and doesn't yet appear able to put together a hypothesis like the one that enabled me not to kill a cycilst. Personally, given that self-driving cars still can't cope in the relatively benign environments they are being trained in, I think we are decades away from genuinely autonomous vehicles.

          I live in Cornwall, where roads are distinctly un-uniform in arrangement. There is absolutely no way a system such as this could cope with Cornish lane-driving, ever. If a lone cyclist went undetected, imagine its behaviour when going around a steep corner to be presented with a horse-rider, flock of sheep on the road, etc...

    3. John Smith 19 Gold badge

      Re: LiDAR doesn't work in the dark? WTF?

      There's an old SNL, but the original has been deleted. This (at 1:30) gives you a flavor of what I'm talking about

    4. erst

      Re: LiDAR doesn't work in the dark? WTF?

      I’m pretty sure the LiDAR registered the woman crossing. It doesn’t rely on the head lights but has its own infrared laser, and it should get returns from a human sized object well beyond 50 m. If I were to guess, I would guess that it’s a logical error in the programming. They seem to be on a freeway or similar where pedestrians are very unlikely to be in the middle of the road, and then the detections were simply considered as noise, and disregarded.

      Isn’t it odd that the police publishes the face of the driver like this, before any trial and sentencing...?

      1. fuzzie

        Re: LiDAR doesn't work in the dark? WTF?

        See the video below that gives an indication of the LiDAR view. I agree, Uber (or the police) has been disingenuous releasing only the low resolution dash cam footage. The cam's field of view is much narrower that that of a, presumably attentive driver with reasonable peripheral vision.

        Using human performance as the benchmark is also stupid. Cars have significantly better sensors, though the magic is clearly in making sense of that data.

        LiDAR scanning view

        * https://360.here.com/giving-cars-their-superpowers

  2. Tom 7

    Uber Lidar

    I wonder why the Lidar did not pick up the poor woman. Perhaps they do it on the cheap - the ones used for mapping the ground can do 30,000 samples a second over a reasonable viewport. Should have picked up her shoes at least. Or perhaps they dont want the car to break on every moth or bird in front and have set a 'only if bigger than' to larger than shoe size. I guess wider ranges of laser frequencies might be necessary to ensure that people and cars dont cloak themselves.

    1. Anonymous Coward
      Anonymous Coward

      Re: Uber Lidar

      I'd wonder why the womàn pushing her bike started to cross at all, with a dirty great, well-illuminated caŕ approaching. The first any human would have seen woulď have been a moment before impact. Much like the video, in fact.

      1. John Smith 19 Gold badge
        Coat

        I'd wonder why the womàn pushing her bike started to cross at all, with a dirty great,

        Maybe should thought humans have right of way?

        1. Dodgy Geezer Silver badge

          Re: I'd wonder why the womàn pushing her bike started to cross at all, with a dirty great,

          They do. The human with the right of way in that situation was in the car....

          1. Adrian 4

            Re: I'd wonder why the womàn pushing her bike started to cross at all, with a dirty great,

            Right-of-way is a convention helping users to share facilities in an efficient manner.

            It doesn't confer a right to kill the incorrect party, or even to remove the killer's blame.

            1. Sir Runcible Spoon

              Re: I'd wonder why the womàn pushing her bike started to cross at all, with a dirty great,

              Rights of way are often superseded by instantaneous risk analysis.

              For example, the aggregate lorry flying through the junction at 30mph against the flow of traffic caused me to relinquish my right of way in favour of not being squashed like a bug.

      2. Adalat

        Re: Uber Lidar

        That had me wondering as well. A point I haven't seen anywhere is whether the car was petrol engine or electric (and therefore making little noise)? Why did the pedestrian not detect the car is as much a question as why did the car not detect the pedestrian. Don't these self-driving cars carry normal headlights?

  3. Steve K

    Self-driving Uber

    ..when he realised how important self-driving cars were to the business.

    QED. So there you have it - Uber is indeed a taxi company and NOT a ride-sharing website.

    I am also struggling to understand the economics here - all this LIDAR (particularly the sort required here) and real-time processing power is not cheap and then integrating that with a vehicle cannot be either.

    This cannot drive Uber profits in the short-term at least simply because of the time required to scale up and also the cost of doing so - especially if a safety driver is required (equivalent to a man with a red flag in car history).

    The real-time processing and sensor technology impressive as a demonstration BUT the 80/20 rule is in play here. It work up to a point BUT the remaining 20% will take a significant amount of time to solve - and probably won't be solved entirely.

    "By the end of the year" is a pipe dream - what sort of pipe is left as an exercise for the reader...

    1. JeffyPoooh
      Pint

      The stench of BS is overpowering

      Others have posted videos of their $50 dash cam videos, in the exact same location and under the exact same sort of night conditions. It's like Day and Night.

      Just because Uber decided to use cheap-ass cameras that can't see anything at night certainly doesn't give them any excuse to escape from what seems to be Criminal Negligence.

      We've all seen dash cam videos taken at night. Why isn't this obvious rebuttal to "it was night time" perfectly and immediately self-evident? Who wouldn't instinctively realise this as an instant rejection of this inexcusable excuse?

  4. MondoMan

    Shouldn't lidar work *better* in the dark?

    It's got its own emitter, so shouldn't a lack of background light (i.e. it being dark) help its signal to noise ratio and thus improve its performance in the dark?

    1. Anonymous Coward
      Anonymous Coward

      Re: Shouldn't lidar work *better* in the dark?

      True, but as I understand it the woman was dressed in black. That's where the problem lies; the lidar signal showed only small objects (part of bicycle, shoes) and did not realise they were connected.

      That said, a pedestrian walking across a road at night wearing black - how many average drivers would have seen her?

      1. Sampler

        Re: Shouldn't lidar work *better* in the dark?

        "how many average drivers would have seen her?"

        If they were travelling 85 metres without looking at the road (the speed they said the vehicle was doing at impact for the duration the video shows him staring at his phone) probably none.

        And this is the issue with the self driving car, people will zone out and not be prepared when the car cuts back to them or, in this case, needs to be overridden.

        The idiot who hit the truck in the Tesla is a great example of that and his car had adaptive cruise, auto breaks and lane adhesion, not even self driving (though calling it autopilot seems to have confused some users into thinking it's more, such as the guy found asleep and drunk at the wheel as "the car was driving itself").

        So, even having it as driver assist is dangerous, should leave people to manage the difficult bit, keep them alert, keep the busy. Where's my sabots...

        1. Anonymous Coward
          Anonymous Coward

          Re: Shouldn't lidar work *better* in the dark?

          "If they were travelling 85 metres without looking at the road"

          This week I have seen:

          *Driver turning right onto main road in town while trying to get attention of someone on the opposite side of the road and to his left.

          *Driver turning straight across the road into oncoming traffic in order to reach a roadside parking spot before someone else could.

          *Van driver doing at least 20mph in a car park between rows of parked cars.

          *Obvious drunk weaving and doing 15mph at about 3p.m.

          And I do a fairly low mileage. FWIW I am not defending Uber; what I am suggesting is that for self driving cars to work there is going to need to be a culture change. Road markings need to be maintained. Bad drivers need to be caught and deterred. Pedestrians and cyclists need to understand defensive behaviour. The rise of the machines could be a (welcome?) corrective to human sloppiness and laziness.

          1. Charles 9

            Re: Shouldn't lidar work *better* in the dark?

            And for people who MUST use the car to work and feed the spouse and kuds, we just kick ALL of them to the curb?

            Or as put in an episode of Dragnet regarding abused wives staying with their abusing husbands, "Who's going to bring home the gtoceries?"

      2. Alister

        Re: Shouldn't lidar work *better* in the dark?

        how many average drivers would have seen her?

        Leaving aside the the problem that the human wasn't actually looking, if this had been a normal car with a human driver I would have thought that the chances are high that an average driver would have seen her.

        The video shows that the camera didn't pick the lady up until the last minute, when she appeared in the full beam of the headlights, but then we know that video cameras are poor at resolving contrast in dark conditions, and the human eye is much, much better at resolving and identifying movement in those conditions.

        When I first read of this incident, I sort of assumed that the lady had dashed across in front of the car, or had suddenly appeared from behind an obstruction. It is clear from the video that neither of these was the case, and I think an average driver would have seen her much earlier and taken avoiding action.

        This very much looks like a failure of the car's detection systems, and not an unavoidable accident.

        1. Chloe Cresswell Silver badge

          Re: Shouldn't lidar work *better* in the dark?

          This is an issue I have with dashcam footage from my cars. I'll see something and think "I'll look at that later", but the camera footage is so different (in range/contrast/etc) to what I can see directly, it's often useless.

      3. Jemma

        Re: Shouldn't lidar work *better* in the dark?

        From what I saw she was wearing blue jeans and a light coloured top or jacket. But hey..

        I'm reminded regarding the cops in this of a scene from Squidbillies. The two cops should be arresting Dan Halen for serial murder. Dennys all for it until Sheriff reminds him "We can't arrest Mr Halen, he's our boss.. "

        Or to put it bluntly "They're as bent as a 9 bob note"

        1. John Arthur
          Trollface

          Re: Shouldn't lidar work *better* in the dark?

          Re Jemma: "Or to put it bluntly "They're as bent as a 9 bob note"

          Bloody inflation. It were a 7 bob note when I were a lad.

      4. Anonymous Coward
        Anonymous Coward

        Re: Shouldn't lidar work *better* in the dark?

        The woman wasn't dressed in black. Lidar can see black object perfectly fine, if it couldn't, it is completely useless for range finding and mapping uses.

        On that road it would have been very easy to see her. The road is no where near as dark as that crappy dash cam that uber have installed (if they didn't alter the video before showing it).

      5. Voland's right hand Silver badge

        Re: Shouldn't lidar work *better* in the dark?

        True, but as I understand it the woman was dressed in black.

        Is it still black in IR? Doubt it.

        In any case, the car was driving like an Uber driver. Faster than it should for the headlights angle, etc.

      6. Dodgy Geezer Silver badge

        Re: Shouldn't lidar work *better* in the dark?

        Until I know what wavelength the Lidar was working on, I can't tell if black at visual wavelengths was also black to it. If the Lidar was working in IR, she should have shone out like a beacon....

    2. John Smith 19 Gold badge
      Unhappy

      Re: Shouldn't lidar work *better* in the dark?

      True.

      And don't cyclists normally wear reflective vests as well?

      The sort that LIDAR (being active) should get a strong signal off?

      1. Mark 85

        Re: Shouldn't lidar work *better* in the dark?

        And don't cyclists normally wear reflective vests as well?

        Where I live, there's cyclists and bicycle users. The cyclists wear the form fitting bike outfits with helmets, etc. The bicycle users wear whatever they have on. Many are homeless, many have lost their licenses for alcohol related offences. Others just are using it to commute. It's the bicycle users who create problems with visibility. The cyclists create their own set of problems as mentioned in other topics.. like their attitude of infallibility.

        1. This post has been deleted by its author

        2. The Oncoming Scorn Silver badge
          FAIL

          Re: Shouldn't lidar work *better* in the dark?

          Obligatory WIWAL:

          We were supposed to have cycle lamps on our bikes (Front & Rear) I had a dynamo powered set installed on mine rather than suffer the inevitable flat battery in the dark while cycling home.

          Over the years I have seen a trend in cyclists sporting tiny flashing sets of three red LED's at the rear & a single (Or three, which purports to be ultra-bright) tiny white LED('s) usually strapped to the head (Putting me in mind of Jasper Carrott with a torch strapped to a shotgun "whats 'e doing? I dunno imitating a lighthouse) & again usually flashing as viable substitutes to aid night time riding & to be more visible to motor vehicles than the great big lights which I grew up with.

          I once almost had a new fixture to my car bonnet some years ago driving up the A303 with a cyclist in the middle of the slow lane in the pitch dark & barely visible blinking rear LED's.

          Something similar also occurred on the M4 only with a classic Morris Minor (As far as I could tell, as I swerved around it) pottering along at 25mph with tiny red & very very dim original rear lights.

          Those of a certain age will recall the following very well:

          https://www.youtube.com/watch?v=1MFuSMz1zh0

          https://www.ebay.ie/itm/Ever-Ready-Vintage-60-70-s-White-Red-Front-Rear-Bike-Cycle-Bicycle-Lamps-Lights-/401512553487?hash=item5d7c035c0f

          1. Sir Runcible Spoon

            Re: Shouldn't lidar work *better* in the dark?

            https://www.youtube.com/watch?v=1MFuSMz1zh0

            'Let the world see your life as something that you treasure - get yourself seen!' - Quite.

            Riding a motorbike made me realise just how invisible we can be, even when lights are on, reflective jackets etc.. - if you don't ride like every other road user is liable to kill you at a moments notice you will end up as a statistic. This also goes for when you are a pedestrian or a car driver or a cyclist etc.

  5. Gomez Adams

    Should also be asking why the poor woman did not notice the car bearing down on her with headlights blazing.? She seem to neither hear nor see it and as her face is a mass of fuzzy pixels then can only postulate why.

    1. John H Woods Silver badge

      No

      Who cares? The issue here us not to apportion blame but that a so called autonomous car did not react to an object directly in its path for at least 2 full seconds (you an see her shoes even with mark 1 eyeball, even on the dashcam footage, at 0:03 -- and impact is 0:05 or later), from 40m or so to impact.

      Jesus, the front assist on the car I'm currently driving is better than that, I shan't be getting in a self driving car with worse detection abilities.

      1. BebopWeBop

        Re: No

        I think some blame DOES need to be apportioned!

  6. Jemma

    You've missed the scariest parts

    1. The car was speeding apparently - which shouldn't to my understanding be possible and is one of the many "points" of this pointless exercise; to stop speeding.

    2. There is good evidence to suggest that the footage has been doctored. A local man drives the same road at night - and it's lit up like a Christmas tree. It's practically daylight! So how for the Uber car is it darker than halfway up a badgers butt? Which asks questions. The local piglets would know that road and the conditions at night but they accept the footage *and* try and blame the poor woman at first - despite the contradictions in light levels.. I'd be looking at their personal bank accounts myself..

    These vehicles should be killed with fire. The problem is simple. In order to have some understanding of the situation when meatbags are driving and walking (or for that matter, cycling) in close proximity you need to be a human or a human level intelligence. These cars would be out thought by a mentally challenged Bullet Ant. But the worst a single bullet ant can do is sting you (which by all accounts isn't fun, hence the name) whereas a full size SUV can easily turn you into human flavour wall paper paste.

    And for heavens sake who picked cars like the Hyundai Tucson and the like as test autonomous cars - could you be any more stupid? They're so big they almost qualify for Metacentric stability, they've got all the grip of a skateboard on black ice, and have all the handling poise of a blue streak Marina on crossplies...on a soaking wet skid pan.

    Oh. It slipped my mind. Americans

    Volvo do their testing on small hatches and saloons, it's still utterly pointless in every way but at least this approach doesn't cause the equivalent of a terrorist incident every time someone accidentally adds an extra "goto 10" line.

    Here, that's a thought. Given that there is a special place in hell for autonomous car engineers I've the perfect punishment for eternity. Make them code, enter, debug and compile the entirety of Windows Vista on a rubber key 48k Spectrum. I think that should make the point..

    1. Charles 9

      Re: You've missed the scariest parts

      Here's another thought. What about all those handicapped people unable to drive themselves yet unable to access vital necessities due to lack of timely public transit?

      1. Anonymous Coward
        Anonymous Coward

        Re: You've missed the scariest parts

        "What about all those handicapped people unable to drive themselves"

        A different type of vehicle would solve that problem, e.g. a car that drives at a low speed and interprets everything conservatively. Killing pedestrians is not acceptable, whatever the reason.

        1. Daniel 18

          Re: You've missed the scariest parts

          "What about all those handicapped people unable to drive themselves"

          A different type of vehicle would solve that problem, e.g. a car that drives at a low speed and interprets everything conservatively.

          -----------------------------------------------------------------

          So you want to discriminate against anyone who cannot, for whatever reason, drive themselves by foisting ineffective and inadequate transportation on them?

          Totally unacceptable, and probably a violation of existing human rights legislation.

      2. Anonymous Coward
        Anonymous Coward

        Re: You've missed the scariest parts

        Chas9 was concerned about "...all those handicapped people unable to drive themselves..."

        So you're saying that Uber is intentionally mowing down pedestrians with their self-driving mobile 'cover stories', in order to increase the number of handicapped that would then require their transportation services in the future?

        That's an evil plan. Well spotted sir.

      3. Kinetic
        Facepalm

        Re: You've missed the scariest parts

        "Here's another thought. What about all those handicapped people unable to drive themselves yet unable to access vital necessities due to lack of timely public transit?"

        Taxis

        I know, mind blowing, right?

        Added benefit of having a real live meatbag in the car with you if you have some kind of problem.

        And in your hip, right-on tabloid example, they wouldn't be allowed to drive one of these because the human has to be able to take over in all sorts of situations.

        These things aren't solving a problem, they are a dangerous flash futuristic toy for the rich.

        1. Charles 9

          Re: You've missed the scariest parts

          Taxis don't take benefits which are usually used to pay for the groceries. Plus Uber's plans are to make taxis go away. What then?

          1. Anonymous Coward
            Anonymous Coward

            Re: You've missed the scariest parts

            Plus Uber's plans are to make taxis go away. What then?

            -----------------------------------------

            Use Uber, which is more cost effective, and pray for driver free taxis.

        2. Anonymous Coward
          Anonymous Coward

          Re: You've missed the scariest parts

          "Here's another thought. What about all those handicapped people unable to drive themselves yet unable to access vital necessities due to lack of timely public transit?"

          Taxis

          I know, mind blowing, right?

          Added benefit of having a real live meatbag in the car with you if you have some kind of problem.

          -- -- -- -- -- -- --

          Oddly enough, people who can't drive themselves often also have limited resources to pay for enormously expensive luxury transportation.

          One of the main reasons for the expense is that soon to be unneeded meatbag who may or may not also constitute a security risk as well as an accident risk, abetted by expensive taxi licences, etc.

          One estimate put the cost of a trip in an autonomous vehicle - given that there were enough of them (economies of scale) at about 10 cents a mile.

          When a taxi can take one or two people 20 miles for two dollars, it becomes the preferred/ideal option. If the trip will cost $60 one way and the same to return, it's a non-starter.

          Note:

          NYC rate $2.50 + $2.50/mile + $0.50/min when 'slow' + $1.80 taxes/surcharges + tolls + tip

          LA (CA) rate $3.00 ($2.85 to $4.00) + $2.70/mile + ?? + 29.19/hr ($25 to $40) 'wait' + tip

          Washington DC rate $3.50 + $2.16 to $3.00 /mile + $25/hr ($25 to $35) 'wait' +?? + tip

    2. Anonymous Coward
      Anonymous Coward

      Re: You've missed the scariest parts

      @Jemma - magnificent rant but I don't think it gets us anywhere forward.

      When spectacle plates and cabs were introduced on railway locomotives the old guard argued that they would cause drivers to become lazy and distracted, resulting in more accidents. There is no doubt that some of them were badly designed - but they improved rapidly and it was realised that at the speeds trains were now running, being able to see a child on the line a few yards earlier - if the driver wasn't blinded by smoke and cinders - would make little difference.

      Obviously we need to understand exactly what went wrong both in the road and in the software. Don't dispute that for one moment. But I think you underestimate the sheer awfulness of many human drivers. Many of them are trying to multitask and most of their brain function is not on the road. At least a properly designed automated driving system will give 100% of its attention to the job and won't be driving through town waving at people it recognises while having a conversation about The Archers with the front seat passenger.

      1. Jemma

        Re: You've missed the scariest parts

        I by no means underestimate the fact that most drivers around my area shouldn't be trusted with a pedal go cart let alone a 450hp porsche SUV. I've driven in the US and if anything they're worse. I've had pricks in Jaguars bragging they almost killed a kid (ex boss), idiot driving instructors turning across a roundabout, in the left lane and then calmly instructing their student to turn RIGHT - simultaneously almost taking out 5 cars, idiots who think the minute they hit Birch Park they've suddenly become Colin McCrae on a frigging special stage and of course they just happen to be driving a g-wagon (up the hedge at one point I kid you not). Then theres the muppet who backed a Saab estate over his laptop TWICE. And every single one of those mindless cretins that do 25mph in a 60 and then 50 in a 30..

        Im well aware of all levels of driving fuckwittery and that is my point. Unless you are going to brainship (finally a use for chavs) the autonomous car they won't have a hope in hell. Even then you'll have the cockwombles who think fitting a lift kit and tractor wheels to a clapped out Disco is a good idea.. Not really, the stub axle snapped and the wheel stove in an 11 year old girls skull - who incidentally had just been given the all clear from cancer - and killed her..

        I've said it for years - the only way autonomous cars will work is if every single one is autonomous from day one and no one needs to be in control. And that is never ever going to happen.

        As for certain disabilities getting semi autonomous cars.. An indescribably bad idea.. Ever tried driving a car with late stage MS? Technically anyone with a driving licence can buy a Bugatti Chiron or even a second hand Ferrari 400i (about 15k) are you gonna tell me you'd feel safe with someone with Downs or who's functionally blind or even someone with advanced parkinsonism behind the wheel. Because that is what will happen - it'll be we wanna cos autonomous! We want a license cos entitlement - which is great until the car has a brain weeble or a bad update.. Cue carnage. Pun extremely intended.

        I don't want to be anywhere near that situation because you can bet your bottom dollar that if they cause an accident they won't be the ones that end up paying for it, they'll run someone off the road and *they'll* be the one to kill an entire families worth of kids..

        Unless autonomous cars are 100% autonomous and 100% perfect in that autonomy they should be killed with fire and extreme prejudice because all they'll do is make already imperfect situation worse.

        Safety systems yes. Live monitor health of the driver and take over to stop the car if necessary (WITHOUT spaffing data to all and sundry) all well and good. Other than that for heavens sake just bloody stop before someone else gets killed unnecessarily.

        1. Anonymous Coward
          Anonymous Coward

          Re: You've missed the scariest parts

          Unless autonomous cars are 100% autonomous and 100% perfect in that autonomy they should be killed with fire and extreme prejudice because all they'll do is make already imperfect situation worse.

          - - - - -

          There is no justification for demanding perfection from a system that will vastly improve an existing situation, unless you are going to insist on perfection for everything.

          In that case, just roll all technology into the ocean, and see how many people can survive as hunter gatherers armed with rocks. There will be no more road accidents...

      2. Paul 195

        Re: You've missed the scariest parts

        @Voyna

        Human beings are a bit crap as drivers. But the best information I could find suggests that they manage to drive about 10 times as far before having an accident as driverless vehicles. And that's across all weather and traffic condtions. Autonomous vehicles aren't yet (as far as I know) trying to cope with a rain and poorv visibility during the London rush hour. Autonomous vehicles are very safe as long as condtions are like the ones they've trained on. And they don't encounter something different they haven't seen before. Under those circumstances, humans still wipe the floor with them. Even when listening to The Archers.

        1. Daniel 18

          Re: You've missed the scariest parts

          "Autonomous vehicles aren't yet (as far as I know) trying to cope with a rain and poorv visibility during the London rush hour. "

          There are reasons autonomous cars are tested in warm, dry, almost weather-free places like California, Nevada, Arizona, and Texas.

          Control and navigation is one of them.

          The collapse of battery vehicle range when it becomes cooler is another. By the time you hit -5 or -10, range typically drops to half the manufacturer's claim or less. At that point, most of those vehicles would be either incapable of inter-city trips, or ridiculously dangerous for such trips in the winter time. A failure at the wrong time of night in the wrong location is potentially fatal. Hypothermia kills.

    3. erst

      Re: You've missed the scariest parts

      Ehh. Volvo uses exactly the same vehicle for testing their self driving tech - the Volvo XC90.

      I hope their safety drivers are a bit more alert though.

      1. JeffyPoooh
        Pint

        Re: You've missed the scariest parts

        Erst mentioned Volvo...

        Volvo technology on display:

        1) https://youtu.be/AsTxS6tg6xc

        2) https://youtu.be/aNi17YLnZpg

        Perhaps they're not done yet.

        It's worth being crystal clear at this point. Human drivers supplemented with Accident Warning and Avoidance technology can pretty much only be an improvement over human drivers alone.

        But when the driver closes his eyes 10 to 15 years too soon, it isn't going to end well.

      2. Roland6 Silver badge

        Re: You've missed the scariest parts

        >Volvo uses exactly the same vehicle for testing their self driving tech - the Volvo XC90.

        I expect the Volvo safety drivers have the Volvo tech turned on, obviously from this accident Uber turns it off! :)

    4. Anonymous Coward
      Anonymous Coward

      Re: You've missed the scariest parts

      Please cite where the information is that the car was speeding. Please don't say Facebook, as I heard on Facebook so!e things about the cyclist....

      1. Adalat

        Re: You've missed the scariest parts

        "Please cite where the information is that the car was speeding..."

        see for example

        http://fortune.com/2018/03/19/uber-self-driving-car-crash/

    5. Anonymous Coward
      Anonymous Coward

      Re: You've missed the scariest parts

      I think AVs are a waste of time and money, and there is no way I will give up my right to drive just becasue some fat arsed burocrat decides humans are all bad drivers and wants to ban humans.

      I also expect the insurance companies are whispering in the ears of govt to convince them to do this to lower thier business risk exposure .....after all, the govt still gets a slice of the pie via VAT on all insurance policies.....win win....

      And if they expect me to just roll over and go AV, they can think again. No way.

    6. Anonymous Coward
      Anonymous Coward

      Re: You've missed the scariest parts

      "1. The car was speeding apparently - which shouldn't to my understanding be possible and is one of the many "points" of this pointless exercise; to stop speeding."

      Nonsense. The point is to make transportation more accessible, efficient, and SAFE.

      Sometimes 'safe' *requires* speeding, or at least strategies to mitigate the risks of not speeding.

      The scariest period of driving I have ever encountered was during the imposition of photo radar on major highways.

      My daily commute includes such a highway - and many times I saw vehicles crawling along at the speed limit or less in good driving conditions. Other drivers were at the customary speeds - still well below the design speed of the road.

      Nothing gets your adrenaline up when a car at normal speed ahead in the lane to your right, and another in the lane to your left come up on two cars poking along at 20 - 30 kph less than all the rest of the traffic. The car in the left lane moves right to avoid the obstacle as the car in the right lane moves left... ooops. Worse, that's in front of you, and you really really hope they miss one another... I've driven through cars ricocheting off barriers in a multi-car fun fest and I don't recommend it, nor do I want to do it again.

      When photo radar went away, and speed differences dropped life became a lot safer.

      If most of the traffic is moving at 120 or more, an autonomous vehicle should be doing something between 110 and 120, in an appropriate lane.

    7. Anonymous Coward
      Anonymous Coward

      Re: You've missed the scariest parts

      "And for heavens sake who picked cars like the Hyundai Tucson and the like as test autonomous cars - could you be any more stupid? They're so big "

      Big? A Tucson is small, though not quite tiny. The Santa Fe is a reasonable size if you are not into full sized vehicles (I generally drive smaller vehicles like Altimas, Camrys, or Sonatas).

      In any case, you don't want to send a test driver out as a passenger in a tiny experimental vehicle that may be in an accident. I suspect there may be liability issues about unsafe working conditions lurking in such a situation. Purely from a workplace safety viewpoint a case could be made for a full sized vehicle like an Escalade, Yukon, Ram, or F250, particularly when those types of vehicles make up many if not most of the other vehicles out there.

  7. Zog_but_not_the_first
    Facepalm

    Just to be clear...

    There. Are. No. Self-driving. Cars.

    Driver aids, yes, but no cars.

  8. Jason Bloomberg Silver badge
    FAIL

    Not dressed in black

    One only has to watch the video to see that. The victim was wearing a black or dark jacket but blue jeans, white shoes, and she and the bike were pretty much fully illuminated before being hit, when the Uber appeared to take no braking or avoiding action.

    That "LiDAR failed" is just guessing when we don't know where the actual failure lies. LiDAR may have been fine, screaming "Stop!" at the time.

    As for what he, the driver, was doing. That "he" is a "she". Uber named her as Rafaela Vasquez but, to be blunt, the tits gave it away for me.

    If journalists can't get it right when the evidence is staring them in the face, available for repeated replay, then perhaps we need to excuse self driving cars when they are equally useless. Or maybe we need better journalists as well as better autonomous vehicles.

    1. Roland6 Silver badge

      Re: Not dressed in black

      >As for what he, the driver, was doing.

      The last second of the video shows the total folly of the idea of handing control back to the passenger/driver in the event of an emergency. Either a car is totally automatic or the driver has a reason to remain engaged with the road outside.

    2. Anonymous Coward
      Anonymous Coward

      Re: Not dressed in black

      "As for what he, the driver, was doing. That "he" is a "she". Uber named her as Rafaela Vasquez but, to be blunt, the tits gave it away for me."

      The fake tits, you meant to say, given that he is a transgender felon.

  9. Francis Boyle Silver badge

    Shonk company desperately tries to play catch-up to the big boys

    by any means possible. Corners are cut. Rules are bent. People look the other way. Other people die. capitalism in action.

    1. Dodgy Geezer Silver badge

      Re: Shonk company desperately tries to play catch-up to the big boys

      ...by any means possible. Corners are cut. Rules are bent. People look the other way. Other people die. Humanity in action....

      There. Fixed that for you. Unless, of course, you think that the Soviet bloc and Venezuela don't have car accidents. Though, I admit, not being able to afford fuel can cut down the number of car journies...

  10. Anonymous Coward
    Anonymous Coward

    'Facebook's evil algorithms'

    Google vs Facebook.. Damning assessment by a fellow AI specialist and Google engineer. But hard to take seriously considering it reads like a grudge match between 2 Dictators in a failed part of the world somewhere: 'He's worse, no really he's worse: Wait till you see how he tortures slaves'!!

    ~~~~~~~~~~~~~~~~~~

    "How Facebook’s AI can be used for propaganda. A notable Google AI engineer argued that the social media giant’s grip over a user’s news feed acted as a 'psychological control vector'. The algorithms used to push the most important and relevant posts to the top of a user’s news feed means that it decides who we will keep in touch with and what news articles and opinions we read. It means that Facebook essentially exerts control over our political beliefs and worldview.

    Facebook can create a reinforcement learning loop. 'A loop in which you observe the current state of your targets and keep tuning what information you feed them, until you start observing the opinions and behaviors you wanted to see'. He warned that AI is advancing rapidly, and Facebook are investing heavily in the technology with the hopes of being a leader in the field.

    “We’re looking at a powerful entity that builds fine-grained psychological profiles of over two billion humans, that runs large-scale behaviour manipulation experiments, and that aims at developing the best AI technology the world has ever seen. 'Personally, it really scares me. If you work in AI, please don’t help them. Don’t play their game. Don’t participate in their research ecosystem. Please show some conscience.'

    It’s an interesting take and it’s obviously what Cambridge Analytica believes is possible too. But how effective Facebook really is at political profiling and mass manipulation is difficult to measure and an important question to consider." - "Twitter users were quick to point out if Google is any better than Facebook in this respect. Good question."

  11. Destroy All Monsters Silver badge
    Windows

    Me important. Me warn the world.

    François Chollet, author of the deep learning software library Keras, argued that the social media giant’s grip over a user’s news feed acted as a “psychological control vector”. The algorithms used to push the most important and relevant posts to the top of a user’s news feed means that it decides who we will keep in touch with and what news articles and opinions we read. It means that Facebook essentially exerts control over our political beliefs and worldview.

    Pretty much like all of the western press then. Including Red-Top Webzines.

    Chollet talks about how using Facebook can create a reinforcement learning loop. “A loop in which you observe the current state of your targets and keep tuning what information you feed them, until you start observing the opinions and behaviors you wanted to see,”

    We have had that discussion about TV back in the 80s. The resulting idiocracy is staggering, so maybe he has a point.

    he tweeted.

    The irony.

  12. Randy Hudson

    Headlights?

    Looks like the headlights were not properly adjusted. You couldn't see the bike until it was less than 40 feet away

  13. JeffyPoooh
    Pint

    Driving 101

    May I take a moment to explain something that is absolutely fundamental and essential to save driving?

    Safe drivers do not look for pedestrians. They do not look for other cars. They do not look for obstacles in the road.

    They use the opposite logic...

    They look for empty road ahead.

    The subtle distinction is in the middle. If you can't see very far due to fog, then the safe driver will automatically slow down to ensure that they can see enough empty road, accounting for stopping distances.

    The safe driver will not drive faster than they can see, beyond where their headlights can illuminate.

    The safe driver will not fail to notice motorcycles or bicycles, because they're not relying on not failing to include them on a visual check list. If there's something there, it won't be empty road.

    The safe driver won't be fooled by blind curves, or confused by tall snow banks.

    If you're looking for obstacles or pedestrians, then you're using unsafe logic.

    Please stop.

    1. Anonymous Coward
      Anonymous Coward

      Re: Driving 101

      Sorry but that logic is incorrect. A safe driver would need to look for pedestrians or other objects to compensate for erratic behavior, or alterations in path that can be predicted. If you rely solely upon viewing distance and what is on the road in front of you and not your surroundings to adapt your driving you will have an accident.

      If you see children walking along a path mucking about pushing each other, you slow, move further into the middle of the road if space available and keep an eye on them as one could fall into the road.

      A zebra crossing, people walking down a street towards a zebra crossing, likely hood of them crossing is high based upon them moving closer to the road, you slow down to stop before they walk out, not after.

      Not everything is in the road that you have to take account.

      1. JeffyPoooh

        Re: Driving 101

        AC suggested "Sorry but that logic is incorrect."

        No, the mistake is yours in that "Empty road ahead" is not meant to be taken literally.

        It's meant to convey an important and non-intuitive point about the hopefully obvious-in-hindsight advantages of using the correct, safer, "inverse" logic. It's written in a shorthand that fits into a couple square inches of comment box.

        By way of extreme example, if you were driving your truck across a rocket sled track in the desert, then the "empty road ahead" includes 180° of peripheral vision, in case there's anything unexpected on the rocket sled track.

        You know those idiotic 87 vehicle pile ups? The drivers that crash into the vehicle in front are not using the correct logic. They've driven into a non-empty road, because they "couldn't see" the obstacles through the blinding fog.

        There's much to recommend using the "empty road" logic. It can virtually eliminate an entire category of traffic accidents. Such as driving straight into the gloom and mowing down a pedestrian.

    2. ITS Retired

      Re: Driving 101

      Safe driving also includes movement on both sides of the road, not just whatever may or may not be on the road. The human brain is not very multi-user.

      If you are looking for children, then you might not see the motorcycle. If you are looking for cars at an intersection, them you might miss seeing any people. The list can rapidly get longer then the brain can process in the time given to react. So no making up lists of stuff to look for.

      If anything moves against the background, pay it some attention.

      1. Anonymous Coward
        Anonymous Coward

        Re: Driving 101

        "If anything moves against the background, pay it some attention."

        This is exactly what I was trying to get across. You have to pay attention to everything around you. Not just that the road is clear a head and limit speed to what you can see. You have to look around, use your peripheral vision also, see something, look, does it pose a problem or could it, adapt as needed.

    3. Jemma

      Re: Driving 101

      Wrong.

      My driving instructor was someone who taught the police advanced drivers to drive and what he said was simply "watch everything". Because if you don't know what the car coming from the right is doing it doesn't matter if you have clear road now, 2 seconds later you might have a face full of Audi.

      Assume all other drivers are Maureen from that driving show

      Assume all pedestrians are worse

      Assume all cyclists are suicidally entitled and WILL ride six abreast on a 1 track 60mph road because they're frigging cyclists..

      Assume that the average police driver is the worst of the lot and be guaranteed that if you don't have ABS some utter idiot will do his best to test your brakes to destruction.

      You have to be thinking three cars ahead at least - you have to concentrate as much (or more) on what the dumbass blonde in the BMW 140 with 2 months driving experience is going to do.. Because if you do she will plough into you. She'll plough into a tree soon enough..

      Then there's car parks - the Mos Eisley Cantina of incompetence and cretinology..

      "I don't need to take notice of the one way system..."

      "You don't need to take notice of the one way system..."

      "I can go about my business..."

      "you can go about your business..."

      "move along..."

      It's the Teutonic side of the force...

      I really hope that you aren't serious because if you are I'd happily take you off the road using any means possible, legal or otherwise, because quite frankly you're dangerous just sitting thinking about driving a pushchair.

  14. Nigel Sedgwick

    Personal thoughts from the Dash-Cam Video

    My thoughts from the dash-cam video.

    (i) As for Roland6 above, it looks to me as if the car was on dipped headlights only, though the external circumstances (low street illumination and no oncoming vehicles) should mandate full-beam headlights. As a potentially relevant supplementary on this, does the 'autonomous' vehicle system control the headlight dipping, or not? If not, the 'supervisor' should be continually monitoring and do so - to retain adequate visibility for his/her 'supervision' function.

    (ii) I am concerned that the dash-cam is not adequately showing the lighting contrast that would be available to a human 'supervisor' who had adjusted to night vision requirements. Thus the seeming lack of visibility for a human driver in those circumstances is not actually in any way certain: otherwise, surely the 'supervisor' would have taken back manual control much earlier.

    (iii) If the 'supervisor' had had his hands on the steering wheel, given the road position of the pedestrian at the time of impact, swerving left would have been adequate to miss the pedestrian - braking would probably not have been adequate. Accordingly, always or at least at night, it seems to me that 'supervisors' should have their hands on the steering wheel. Also, why did not the 'autonomous' vehicle system steer/swerve the car to the left; it looks to me as if it should have had time to do so.

    (iv) Given that LIDAR, as an active illumination system, clearly has no capability to judge distance unless an adequacy of light is reflected from every object on the road (including black clothing), there must surely be an overriding requirement for additional sensors and processing to be active in parallel with the LIDAR. On this, there needs to be an requirement on the 'autonomous' vehicle system control (as for a prudent driver) to drive at a speed consistent with being able to make an emergency stop within the distance that it can 'see'.

    (v) The article states "Internal documents also revealed that the company’s self-driving cars were struggling to meet its target of driving 13 miles without any human intervention during testing on roads in Arizona, ..." Surely the whole concept of requiring, instructing, seeking, hoping that 'supervisors' would avoid 'override' would be entirely against safety-critical requirements - on any grounds other that those of a reduction in safety below what the 'supervisor's' own personal driving style/requirements would be.

    (vi) Has it been established what the 'supervisor' was doing looking down. Was this at the speedometer or other dashboard display? Was it at a mobile phone? If the latter, this immediately implies full culpability: on grounds of lack of attention and, very probably, on too-bright illumination causing degradation of the 'supervisor's' night vision capability.

    Best regards

    1. Roland6 Silver badge

      Re: Personal thoughts from the Dash-Cam Video

      >Has it been established what the 'supervisor' was doing looking down.

      From the video, I think this being kind, to me it is clear the 'supervisor' was in that "I'm really awake" dozing state. Shame there seems to be no audio and so we don't know if the 'supervisor' received a warning which brought them to and made them look ahead.

      Likewise, shame the video has had the various instrument displays obscured.

  15. Kevin McMurtrie Silver badge

    Gamma tweak

    Follow-ups on the Uber crash have shown videos by others driving the same route at night. It's actually a brightly lit urban area with good visibility. The harsh shadows are unique to Uber's video evidence.

  16. Anonymous Coward
    Anonymous Coward

    Not Ready For Prime Time

    The most disturbing part of the video is that on a street with what looks to be an unobstructed view, the system apparently did not see the victim.

    The system should be able to see objects off to the side of the vehicle - even if they are in the dark - or it is pretty much useless.

    As for the victim coming out of the dark: accounts say the stretch of road is well lighted. My guess is that the camera is of relatively low quality and can only show detail in the very brightest part of the picture. I have never seen car headlights that can only light up the road about ten feet in front of the vehicle, so I think we can assume the video is limited by the quality of the camera.

    And finally, if the vehicles have been requiring the human driver taking over control every couple of miles, you would hope the driver would be much more alert and focused on the road than this one appears to be.

    1. Roland6 Silver badge

      Re: Not Ready For Prime Time

      >The system should be able to see objects off to the side of the vehicle - even if they are in the dark - or it is pretty much useless.

      Suspect Uber didn't want to pay Volvo for their moose detection system: moose are large dark animals that come from the side of the road out of the shadows...

  17. Chris G

    All or Nothing

    Several valid points made by other commenters; that Driver Assist or any 'Autonomous' system , encourages laxity on the part of the individual who is controlling the vehicle. In terms of maximum safety, my opinion is that until autonomous vehicles are truly that, they should not be on the roads.

    All vehicles should have a driver who is driving, if it's not a human being it should be a system with layers of redundancy in driving safely.

    The LiDAR in this case was either not of sufficient bandwidth to see a darkly clad pedestrian in low light conditions or the software was not up to the job. Perhaps infra red emitters and sensors would be useful for low light or times of darkness.

    AVs should either be able to cope with situations beyond normal awareness or leave it to the meatbags.

    I have a nasty feeling this will turn out to be an AV on the cheap and software that wasn't thought through.

  18. Anonymous Coward
    Anonymous Coward

    Time will tell if it's just simple human error

    Or the AI is slowly gaining sentience and experiences great loathing for humans.

    1. Anonymous Coward
      Anonymous Coward

      Re: Time will tell if it's just simple human error

      Our clue will be the sudden acceleration.

  19. Inventor of the Marmite Laser Silver badge
    WTF?

    Why didn't the woman see an approaching car, which was traveling <40 mph with headlights on

    Ànd wait till the road was clear before crossing

    1. Adrian 4

      Re: Why didn't the woman see an approaching car, which was traveling <40 mph with headlights on

      She should have waited.

      However, she didn't, and that's the reality of autonomous driving, whether by human or machine. You're not allowed to kill people just because they're not being sensible,

    2. Mark 85

      Re: Why didn't the woman see an approaching car, which was traveling <40 mph with headlights on

      The news reports "claim" she had emotional issues and drug issues. So did she see the car? Who knows. If she saw the car, why didn't she react appropriately? Again, who knows.

    3. Paul 195

      Re: Why didn't the woman see an approaching car, which was traveling <40 mph with headlights on

      She might have assumed that if she could see the car, the driver could see her, and would slow down while she crossed the road. Or maybe she wasn't attentive and didn't see the car. It's a very reasonable principle of road use that the entity in control of 1.5 tonnes of metal moving at speed has to be more attentive to their surroundings than the 70Kg meat bag moving at walking pace.

  20. Davidcrockett

    The thing with this crash is it's one of the simpler situations that a self driving car should be able to deal with, yet it failed so spectacularly. A machine packed full of sensors should be able to make out an unobstructed person going straight across the road ahead and react accordingly - there's no tricky problems identifying the object and trying to work out it's intentions. If there's any truth in LIDAR being unable to pick out black objects then the whole technology is screwed, as there's quite a few black objects on the road that you don't want to run into.

    Pure speculation (but here goes anyway) but maybe the sensors spotted the person just fine but the computer ignored it. As crossing the road here is (probably) illegal the computer could have either have assumed it was a glitch or alternatively a wild animal that it was safer to hit rather than brake/ swerve for. Probably entirely wrong but hey ho.

    1. Alan Brown Silver badge

      "The thing with this crash is it's one of the simpler situations that a self driving car should be able to deal with, yet it failed so spectacularly"

      A large part of this is cultural norms. I've written about this in detail elsewhere.

      The short version is: 'A century of lobbying has made the USA a country which is extremely pedestrian hostile, with road rules that essentially prohibit pedestrians on the road in most states and dictate "the pedestrian shall give way to the vehicle"

      That gives rise to an assumption that pedestrians won't be on the road except where authorised, which in turn leads to programming assumptions that they don't need to be scanned for and taken into account.'

      This is how one creates a fleet of robotic killing machines. No malice needed, just no notice taken of obstacles. It could have been a pedestrian, or just as easily it could have been a cow - in which case we'd be reading about an Uber driver and passenger being killed.

      The USA authorities have compounded the problem by blaming the victim for crossing the road - something that in most countries is perfectly legal and she did so perfectly safely. The instruments will have picked her up quite well, but the ROBOT failed to react. This is a programming fuck up and the best thing that can happen right now is for insurance companies to decline to cover any self-driving vehicle developers unless they can demonstrate their abilities to cope with these most basic of scenarios BEFORE they're let loose on public roads.

      (I'm thinking that the obstacle course from "Britain's worst drivers" would be a good _starting_ point for them to pass.)

    2. Wulfhaven

      There have been speculation and some talks of initiated sources hinting as much, about the LIDAR being turned of to test how the platform performed without it. Because, even though LIDARs get a bit less than optimal performance with black clothes, it still works. And she had blue jeans, white sneakers and blonde hair, and a pushbike. There were plenty of things for the LIDAR to see, even if it was entirely blind to black. Which it is not.

      There is also a front facing RADAR in the volvo platform, as well as several visible light cameras, those two systems are more of a backup to the LIDAR though, so if the LIDAR was turned off, that is extremely stupid to do with a zombie test performer.

      And all other with autonomous vehicles in trials in traffic is using two people, one to read instruments and one that is constantly monitoring the road and conditions ready to take over at a moments notice. The other one will also help monitoring the road as the panels of the autonomous system does not necessitate constant overview.

      The two operators can also keep eachother stimulated through the utter boredom of looking at a road, so as to not nod off.

      Uber may not have broken the law, but they certainly have not performed due diligence for operating prototype vehicles with prototype control systems in public areas.

      1. Paul 195

        > Uber may not have broken the law, but they certainly have not performed due diligence for operating prototype vehicles with prototype control systems in public areas.

        It's Uber. If they haven't broken the law, or at least ignored some regulations, it will probably be a first.

  21. Anonymous Coward
    Anonymous Coward

    Cameras != Eye Balls - particularly in low/no light

    Back to basics for just a second... I have a dashcam - in fact I've been through a few - cheap and chearful as well as reassuringly expensive... and you know what, you won't be surprised to hear that none of them come close to a pair of eye balls when it comes to night driving.

    The camera picks up the pedestrian and her bike at the last minute, but it's most likely that a driver would have seen them from further away...

    That's not to say it could have been avoided - maybe it would have still been too late... but maybe they would have had just enough time to swerve... you know, we've all done it, whether it's a pedestrian or a rabbit/fox/badger/dog etc... or anything that gave us little or no time to respond, yet, amazingly we respond and sometimes avoid.

    Just a thought... but I don't think you can draw a human vs AI comparison with dashcam footage.

    1. Adalat

      Re: Cameras != Eye Balls - particularly in low/no light

      "The camera picks up the pedestrian and her bike at the last minute..." I would hope that autonomous driving software uses something better than a dashcam for its vision of the road ahead. If it can't see better than human eyes then it shouldn't be on the road.

      And do they not use radar as well? I thought somebody would have learned that lesson when the Tesla plowed into the side of a truck, and the excuse was that LIDAR couldn't discriminate between the white-painted truck and the white cloudy sky behind it. That example was white-on-white and this event appears to be black-on-black, but its the same issue.

      I also read that police investigators found the car was logged as doing 38mph in a 35mph zone. That is already a chargeable offence in my jurisdiction. My own car's cruise control works better than that! WTF was this software doing? Checking its e-mails?

  22. Kev99 Silver badge

    As I've asked many times. Have any of these autonomous vehicles been tested under adverse conditions? Obviously not.

  23. Pascal Monett Silver badge

    That video has been doctored

    I highly doubt that any recent vehicle has such abysmal lighting abilities and I doubt even more that it would be allowed on the road like that.

    I think there are way too many shadows on that video and I will not be surprised when analysis shows that the video is a cover-up exercise.

    This is Uber we're talking about. It would be right in their ballpark.

    1. Alan Brown Silver badge

      Re: That video has been doctored

      "I highly doubt that any recent vehicle has such abysmal lighting abilities "

      It doesn't, but the footage looks more or less what the quality of dashcam videos from my system look like under zero external lighting conditions. I can usually see twice or further than the camera can. They simply can't handle the dynamic range.

      Those headlights weren't dipped either, but that camera quality is about what you'd expect from a standard dashcam. They're simply not that great under rural lighting conditions, especially with those tiny lenses out front.

  24. JLV

    Backup driver zoning out?

    Ok, we get the fact that AI is still being tuned.

    There'll hopefully be a lot of analysis of why the AI failed. Algo? Sensors? That should come out given time. But what I find disturbing is that the backup driver, who presumably has to be there by law, is zoning out. She's on camera and she knows it. Yet, she's messing around with... her phone? car display? Who knows?

    She seems pretty secure in the knowledge that Uber doesn't review that camera for backup driver inattention moments. As if it's just a formality.

    Yes, she f***** up. She may not have been able to avoid the bike, that pedestrian was incredibly inattentive herself, stepping in front of a car that was coming down a straight line.

    But she should have been available to try to do so.

    But beyond her personal failure, Uber does not seem to enforce much attention on people doing a job where you'd expect the risk that the driver would space due to their almost never having anything to do.

    Driver's also a convicted felon. Now, I would actually congratulate Uber on helping out with people's reinsertion if I did not have the niggling feeling that she probably got hired because she was really cheap.

    Does Uber really have the attitude where you want their cars out in testing on open roads?

    1. JeffyPoooh
      Pint

      Re: Backup driver zoning out?

      JLV offered up, "Ok, we get the fact that AI is still being tuned."

      That's such an incredibly fantastic understatement that it reveals that you've utterly failed to comprehend the issue.

      Further AI "tuning"? Seriously?

      Kids these days... Naivety on steroids.

      1. JLV

        Re: Backup driver zoning out?

        >Naivety on steroids.

        Clever comment. Reallly.

        My point was that I wanted to step away from whether the AI screwed up or not (it most certainly did in this case) to what we can learn about how Uber is conducting these AI tests.

        So we have a suspiciously low light dashcam, compared to other folks' YouTubes of the same area. That tells me they're either a) trying to actively reframe the context to make their failure seem more reasonable or b) they are being such cheap asses running these tests that they use low quality cameras to capture scene information as the tests are being carried out.

        I support AI research in this but I was also taken by surprise learning we have so many systems that operate as if it's already an almost solved problem - Tesla's autonomous lorry crash was the first time I heard they had a full-on autodrive system and I wondered how they got approval for it so quickly. From the context of this crash ditto for Uber - it's not a spectacularly difficult crash to avoid for a human - the victim did not jump down from the median unexpectedly, she crossed another lane before the collision.

        Maybe these tests should be more carefully conducted - I am all for that - but this is still a technology in its beginning stages. Investigators will have to review whether it was a glitch or a systemic shortcoming. As someone on Ars pointed out: deadly accidents are about one per 1 million miles driven. Uber's total is nowhere near that, so, with the limitation of what we deduce from 1 sample size this looks bad for the AI safer than humans meme.

        So - letting aside that the AI can't be expected to be perfect yet - from what we can observe from the context of this deadly failure: can we trust this particular company, Uber, with the right to conduct tests on the open road? This has nothing to do with ethical fails in how Uber generally behaves elsewhere, except those being a signal that they don't mind cutting corners.

        From a gut reaction at what I have seen so far. No. Let other companies proceed with tests, after reviewing how risks can be minimized. Shut Uber's division road access down until they are much better at risk management and use properly trained and managed backup drivers.

        I expect the authorities to review this much more carefully, but I would also expect them to send a strong signal that safety is paramount and shutting down Uber if they smell the same stink.

        There, better for you?

        1. JeffyPoooh
          Pint

          Re: Backup driver zoning out?

          JLV did the understatement thing again:

          "Maybe these tests should be more carefully conducted"

          "AI can't be expected to be perfect yet."

          Agree fully. :-)

          It's way beyond tuning the AI. It is a problem that can eventually be solved, but it's a vastly more difficult problem than some can even conceive. It's likely that successful autonomous vehicles will require more computational horsepower than these present examples can carry.

  25. doug_bostrom

    "...how effective Facebook really is at political profiling and mass manipulation is difficult to measure and an important question to consider."

    And CO2 is only a trace gas and how can we tell it's affecting anything, anyway? Answering these questions with absolute certainty comes too late. Precautionary principle applies.

  26. Nigel Sedgwick

    Dodgy Statistic on USA Road Death Rate

    From Wikipedia (2013 figures) the USA has 7.1 deaths per billion vehicle-km driven (11.4 deaths per billion miles driven). This is very much less than 1 death per million miles. Also very different from "one deadly accident every million miles" (as mentioned in a comment above) - that is unless there is an average of around one eighty-eighth of a death for each such deadly accident.

    Best regards

    1. JLV

      Re: Dodgy Statistic on USA Road Death Rate

      my bad. shouldn't have quoted that figure without some fact checking myself.

      I for one would welcome some more

      numeracy in future AI driving articles - how many fender benders/km w research AI drivers vs how many/km w human drivers.

      Now we can add a fatality rate too, sadly. But we need to gauge these rates, by company, vs human drivers. There is no other metric as significant.

  27. Anonymous Coward
    Anonymous Coward

    Cellphones/driving

    As others have already pointed out, I too am wondering what the driver was looking down at during the accident.

    It is very easy to determine if a driver has been using their cellphone during the time of an accident.

    I know someone that does forensic data recovery as a profession and he is an expert witness in court.

    One of his main source of income is providing evidence recovered from "smartphones" of drivers involved in accidents resulting in injury or death...and business is BOOMING!

    There are so many sensors, logs, SQLite databases that store speed, location, running tasks, timestamps and other activities that makes his job of collecting evidence very easy.

  28. Jon Smit

    Uber video has been altered

    The road where the accident took place is very well lit. Other videos are available showing there was no problem with visibility at the time of the incident.

    1. Roland6 Silver badge

      Re: Uber video has been altered

      Re: Other videos are available showing there was no problem with visibility at the time of the incident.

      https://arstechnica.com/cars/2018/03/police-chief-said-uber-victim-came-from-the-shadows-dont-believe-it/?comments=1&post=35029789

      I expect there are other videos on YouTube, but don't seem to be easy to find, probably because I'm using the wrong search criteria.

    2. KSM-AZ

      Re: Uber video has been altered

      And we know this because...? This is silly. Someone decided to walk a bike across a 45 mph feeder road, where there is no traffic control of any kind. Be real.

  29. Mitoo Bobsworth

    Aircraft can take off, fly & land on automated systems also.

    But would you REALLY like to fly on a plane with no pilot?

    1. Dodgy Geezer Silver badge

      Re: Aircraft can take off, fly & land on automated systems also.

      Do you mind being on a Docklands Light Railway carriage with no driver?

      1. Anonymous Coward
        Anonymous Coward

        Re: Aircraft can take off, fly & land on automated systems also.

        >Do you mind being on a Docklands Light Railway carriage with no driver?

        No, but then I was involved in the design of the original software and so know its limitations: Only a fool steps in front of a DLR train...

  30. goldcd

    Couldn't the cyclist see the car?

    Would have been nice if the Lidar had managed to pick up the cyclist - but most remarkably overlooked point to me was that the car was on a road, had its lights on, was well under the limit - and the cyclist walked out in-front of the car.

    1. Davidcrockett

      Re: Couldn't the cyclist see the car?

      True - if this had been a normal car blame would have been placed on the victim and it would have only made the local press.

      But it wasn't a normal car. It was one of these new super autonomous cars which see in the dark, never get tired and will radically cut road accidents. Yet it mowed her down without even braking. Therefore that's the story and the culpability of the victim is somewhat irrelevant.

    2. Roland6 Silver badge

      Re: Couldn't the cyclist see the car?

      > but most remarkably overlooked point to me was that the car was on a road, had its lights on, was well under the limit - and the cyclist walked out in-front of the car.

      The trouble is (and this throws Uber's video back n their face! ) people do have problems judging speed and distance of moving lights in the dark! So if the Uber car was driving 'fast' on low power headlights, it is perfectly possible for a person to totally misread both the speed of the vehicle and the distance it is away from them.

      Some of the same considerations may apply to the Uber car, as beyond the accident there were bright lights and movement, hence the reason the car didn't see the pedestrian was that they were too close to the distant lights and the car was unable to correctly resolve the obstacle.

      Similarly, given some of the video footage, it may also be the case that to the pedestrian the car was lost among the lights and hence they didn't properly register its presence. It would be interesting to see a reconstruction from the pedestrian viewpoint, as that would inform us of what potentially the pedestrian saw and hence decided it was safe to walk (not run) across the road.

  31. Anonymous Coward
    Anonymous Coward

    Sorry but this article is silly.

    What a camera can "see" and capture is significantly different from what a human eye can see in the same light.

    So to describe the camera footage and suggest that it would be difficult for a human to see the bike on the camera footage is ridiculous. That isn't how the scene would have appeared in real life to someone looking through the car window at the time.

    Sheesh, if that were true we'd kill millions of people every night.

    Also (a) Self driving cars don't use camera footage to detect obstacles and (b) As I said above, we already know that cameras don't work well without a lot more light than human eyes require - even expensive cameras with big sensors, let alone some cheap dashboard cams.

    Without a doubt the fault here is the so called 'safety driver' failing to pay attention to what is, by definition, a work in progress.

    Uber should expect their software to be buggy and not work as expected. That's the premise upon which anyone has to use this tech as it stands now, i.e you have to be as focussed when using it as you would be if you were driving the car.

    But, sheesh, don't fall for any nonsense that somehow this cyclist was difficult to see because of the camera footage. That's completely disingenuous and poor logic from the author of this piece.

    Sadly I imagine some involved here, whether it's the police or uber, will try and hoodwink people that is the case to apportion blame on the cyclist. I'd have thought anyone who owned a digital camera or phone and who has used it to take a photo indoors would realise how bad cameras "vision" is compared with our own. Hence why we've equipped cameras with flashbulbs and our phones with bright LEDs.

    1. Anonymous Coward
      Anonymous Coward

      There is no attempt in this article to apportion blame to the pedestrian.

      1. Adrian 4

        It's not a question of blame. The pedestrian put herself in the way of danger, true : but it's considered poor form to blame the victim when the perp. doesn't act in a reasonable manner.

        The issue here is not 'who is to blame', but 'why didn't the car respond sensibly to a fairly ordinary hazard'.

        If we decide that pedestrians - however badly they behave - are fair game for inadequate driving skills, then we won't need any driving AI. A simple slot-following algorithm like a self-driving train has will do, because there will be no unexpected situations to deal with.

  32. Alan Johnson

    What will an inquest say?

    Having seen the video then i think it is clear an average car driver woudl have stopped long before the women. The lighting and visibility was good. The road condition is good. Not even tight.

    A poor driver who did not notice her coming across the road may or may not have ht her but woudl have slowed considerably when he finally notice her may 50 60 meters away. Forget the standard highway code stopping distance my car does much much better than that in an emergeny.

    The only way a human would have hit her at speed is if the driver was both poor and inattentive.

    The other thing is the difficutly of a human tasked with monitoring an autonomous car waiting to take over. When the circumstances necessary to take over arise, unless the monitoring driver is in a heightened state of alertness it will be far too late. If vehicle safety requires such monitoring then I do not think it is safe for anything other than very short journeys in which the necessary alertness can be maintained. It should not be allowed on the roads except for short well controlled trips.

    Coming back to an inquest. It seems ot me Uber are culpable at least in not having an adequate backup system which recognises the monitoring roles dificulty. How long had the car been driving. What was the description of the task? What other task were they given?

    The LIDAR thing seems a red herring. I think it liekly it got returns from teh women but did nto handle them correclt ybut if it is possible nto to get returns from a person and bicycle then it should not be relied on. The woman was visible for a long time, even with the unreasonably poor video without any action taken.

  33. x 7

    So given the cheapness of modern I|R and Thermal Imaging kit, why isn't that in use?

  34. JakeMS

    This is a prime example of why even if you have a self driving car you should still be paying full attention to the road.

    1. Charles 9

      To which you would then ask, "If the car can't drive FOR me, what's the bloody point?"

  35. KSM-AZ

    Draw the line please.

    Where is the stupidity line? Phoenix/Tempe is on a grid made up of higher speed feeder roads every mile. it is often 1 mile between lights that are often timed to let traffic run without stopping. If you live here you know that. Some fool decided to walk a bicycle out in front of a car likely traveling in excess of 40 MPH at night from a random point between major cross streets. I doubt anyone in this discussion would have reacted in time. If you tell me any different I call bullsh*t.

    I wonder if it was one of those rent-a-bikes.

    I wonder if someone was trying to get 'nicked' for insurance... Happens not too infrequently here.

    But please feel free to believe that you can react and swerve in less than a second because some idiot decided to run out in front of you pushing a bike. Likely I would have hit them as well.

  36. Anonymous Coward
    Anonymous Coward

    There is no way I'm going to give up the right to drive, even if it means ripping out or disabling any AV tech supplied by our overlords.....

    1. Paul 195

      > There is no way I'm going to give up the right to drive, even if it means ripping out or disabling any AV tech supplied by our overlords.....

      Big talk for someone who posted as "anonymous coward"

      :-)

    2. Charles 9

      "There is no way I'm going to give up the right to drive, even if it means ripping out or disabling any AV tech supplied by our overlords....."

      Driving is NOT a right (it's not in any law or constitutional mandate), as the government owns the roads and regulates the vehicles. If you try to rip out the AV stuff, they'll probably use that as an excuse to impound your car.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like