Remember the Uber self-driving car that killed a woman crossing the street?

Bmblbzzz wrote:

That the softare had no way of recognizing her as a person crossing the road because she wasn't at a designated pedestrian crossing shows it's not simply a software failure, it's a failure of consideration on the part of Uber. 

Which would be fine, if the driver "assisting" the vehicle had been doing their job, but they weren't. I expect Uber will throw them under the autonomous bus for this.

This actually ties back well to what someone was saying earlier about level 4 vs level 5 autonomy. The system has to be fully independent of human intervention, because whether it is or not drivers will behave as though it is. They aren't going to be paying attention just in case the system asks for their help, and to be honest that makes sense, because what is the point of a self driving car that you have to drive? It's all or nothing, which is why it needs incredibly rigorous testing, not this half arsing from uber.

vonhelmet wrote:

Bmblbzzz wrote:

That the softare had no way of recognizing her as a person crossing the road because she wasn't at a designated pedestrian crossing shows it's not simply a software failure, it's a failure of consideration on the part of Uber. 

Which would be fine, if the driver "assisting" the vehicle had been doing their job, but they weren't. I expect Uber will throw them under the autonomous bus for this.

I can't entirely agree with this point. If the human driver had been paying attention and ready and able to take control of the vehicle at any point, that would – hopefully – have avoided this death. But the fact that Uber were prepared to put the vehicle on the street without any way of identifying humans other than at specified points, betrays their machine-first attitude. I don't think it's just a case of releasing the vehicle into the wild prematurely, it's a case of not caring about the complexities of human-vehicle interactions.

Quote:

This actually ties back well to what someone was saying earlier about level 4 vs level 5 autonomy. The system has to be fully independent of human intervention, because whether it is or not drivers will behave as though it is. They aren't going to be paying attention just in case the system asks for their help, and to be honest that makes sense, because what is the point of a self driving car that you have to drive? It's all or nothing, which is why it needs incredibly rigorous testing, not this half arsing from uber.

On this point I agree. In fact we've already seen similar incidents with less advanced vehicles that don't claim to be autonomous, such as Tesla's lane-keeping and braking system. In that case it was the human driver, who wasn't driving but sitting in the driving seat, who was killed when the car's systems failed to detect a lorry ahead. 

Argus Tuft wrote:

Without autonomous ride-share vehicles,Uber is in deep trouble.They've convinced legislators in one maverick state to relax the rules enough to allow a  limited experiment. Carmakers are backing away from this as fast as they can. The public has lost interest. Look up "What happened to autonomous cars" on your search engine of choice.

Think of all the better,more useful ideas that have proved futile-Supersonic passenger aircraft,Maglev trains,flying cars-it's a long,long list.I may be wrong but I'd be ready to bet that these cars will be the next addition.

I guess widespread takeup of autonomous vehicles will create something of an issue for driving instructors...

Mungecrundle wrote:

Argus Tuft wrote:

Without autonomous ride-share vehicles,Uber is in deep trouble.They've convinced legislators in one maverick state to relax the rules enough to allow a  limited experiment. Carmakers are backing away from this as fast as they can. The public has lost interest. Look up "What happened to autonomous cars" on your search engine of choice.

Think of all the better,more useful ideas that have proved futile-Supersonic passenger aircraft,Maglev trains,flying cars-it's a long,long list.I may be wrong but I'd be ready to bet that these cars will be the next addition.

I guess widespread takeup of autonomous vehicles will create something of an issue for driving instructors...

Examiner actually,but fair point,and well spotted! As a recent retiree,I'm missing work,but I'd die of boredom before I'd go back to that job. It had it's moments though.Once at the end of my pre-test spiel,I asked the kid if he had any questions.He said "Yeah,what's a demystifier?"

Mungecrundle wrote:

Argus Tuft wrote:

Without autonomous ride-share vehicles,Uber is in deep trouble.They've convinced legislators in one maverick state to relax the rules enough to allow a  limited experiment. Carmakers are backing away from this as fast as they can. The public has lost interest. Look up "What happened to autonomous cars" on your search engine of choice.

Think of all the better,more useful ideas that have proved futile-Supersonic passenger aircraft,Maglev trains,flying cars-it's a long,long list.I may be wrong but I'd be ready to bet that these cars will be the next addition.

I guess widespread takeup of autonomous vehicles will create something of an issue for driving instructors...

Not quite sure if that's intended as a real point or is just a throwaway  jokey comment.

But I can't believe that's a job that anyone is going to call for a national-strike to preserve.  I doubt anybody will care about that issue, not even driving-instructors.

And I agree with Argus above.  There are as many over-hyped 'next big things' that didn't pan out in tech as there are in rock music.  I think there's a fair chance self-driving cars will be "the Stone Roses second album" of technology.

Maybe you'll get automated lorry convoys on US highways or some motorways.  I'm not holding my breath when it comes to real urban transport.

Also, what bothers me is the hidden agenda strongly alluded to in this sort of disengenuous article - from someone who is a proponant of AI, yet, significantly, admits we'll have to rearrange ourselves for its convienience, rather than the reverse.  Funny how infrastucure changes are a huge stumbling block if it's for bicycles, but apparently can be easily achieved if it suits the commercial interests behind AVs.

https://www.wired.com/2016/03/self-driving-cars-wont-work-change-roads-a...

AlsoSomniloquism wrote:

The reason I picked up the footage maybe being doctored is that only one street light (up high on the right of the Uber one) in the distance is on. According to both the correspoding videos showing the accident site taken a couple of days later, there should have been at least two more in view closer and to the left. It is just a black area. So indicates doctoring or those lights might not have been on. I would be surprised if it was just the settings on the camera as the lights on the building and the previously mentioned light in the distance are visible if fuzzier. 

I'm sorry for coming across as callous and / or totally victim blaming as the accident should never have happened if the "driver" was "driving" or if Uber had programmed better. However I still don't understand what the pedestrian was expecting to happen when she crossed there. Any car approaching would have had to take avoiding action of some sort. However, If the lights were out, it is easier to misjudge the position or speed of a car just from headlights so maybe it was that. If the car changed lanes it might have been that. Unfortunately we will never know from her side and that is the big tragedy here. 

She may have been drunk or suffering from mental issues or just simply pre-occupied with whatever was going on in her life. She wasn't threatening or violent and just simply wanted to get across the road. Maybe she would have been sorry for causing a few seconds of inconvenience to a driver.

hawkinspeter wrote:

She may have been drunk or suffering from mental issues or just simply pre-occupied with whatever was going on in her life. She wasn't threatening or violent and just simply wanted to get across the road. Maybe she would have been sorry for causing a few seconds of inconvenience to a driver.

Did you have the same understanding for the Kim Briggs in the Alliston case? Or the ped knocked down on  the crossing who then sued the cyclist? Because when you look at what happened in all of these cases, they are all pedestrians crossing the road and getting hit by a vehicle that was on the road. Or where you more on the side of the cyclist on these, and apportioning some fault for the incident on the pedestrian to some extent?

AlsoSomniloquism wrote:

hawkinspeter wrote:

She may have been drunk or suffering from mental issues or just simply pre-occupied with whatever was going on in her life. She wasn't threatening or violent and just simply wanted to get across the road. Maybe she would have been sorry for causing a few seconds of inconvenience to a driver.

Did you have the same understanding for the Kim Briggs in the Alliston case? Or the ped knocked down on  the crossing who then sued the cyclist? Because when you look at what happened in all of these cases, they are all pedestrians crossing the road and getting hit by a vehicle that was on the road. Or where you more on the side of the cyclist on these, and apportioning some fault for the incident on the pedestrian to some extent?

Kim Briggs was quite a tragic case as her death could easily have been avoided. I don't have much sympathy for Alliston as he didn't show any remorse after the event, though I still think his punishment was excessive compared to similar incidents involving drivers and pedestrians.

The Brushnett-Hazeldean case (presumably that's the one you're referring to) was initially presented as a poor innocent cyclist who got caught out by the pedestrian crossing whilst looking at her phone, but the facts turned out to be somewhat different (the cyclist deliberately rode at the bunch of pedestrians crossing the road). In the comments on the stories here, I advocated that the cyclist was a fool for not getting legal advice and I then suggested that his best course of action was to go for bankruptcy. I didn't give the cyclist or the pedestrian much sympathy as neither was blameless (obviously, I'd have preferred that they hadn't collided at all). In the end, the cyclist got landed with an excessive bill due to his ignorance of court/law procedures and I think the 50/50 result did actually make sense.

I don't think cyclists are always blameless or drivers are always at fauly, but if you're in charge of heavy machinery then you must take responsibility for your actions and recognise that you are bringing significant danger into the lives of others. What bugs me is the double standards whereby drivers can claim "I didn't see her until too late" or "I didn't expect her to cross the road there" and they pretty much get let off due to juries all being drivers and all thinking "I could have done that".

AlsoSomniloquism wrote:

hawkinspeter wrote:

She may have been drunk or suffering from mental issues or just simply pre-occupied with whatever was going on in her life. She wasn't threatening or violent and just simply wanted to get across the road. Maybe she would have been sorry for causing a few seconds of inconvenience to a driver.

Did you have the same understanding for the Kim Briggs in the Alliston case? Or the ped knocked down on  the crossing who then sued the cyclist? Because when you look at what happened in all of these cases, they are all pedestrians crossing the road and getting hit by a vehicle that was on the road. Or where you more on the side of the cyclist on these, and apportioning some fault for the incident on the pedestrian to some extent?

That seems a weak  kind of argument, as it ignores what was actually said about those cases. 

ktache wrote:

Can anyone please explain why making an automobile self driving would result in a paradigm shift in car ownership, so that no one will own their own status symbol and instead be prepared to be driven around in whatever was on offer?

Is there any evidence for this assumption?

It does seem to be regularly trotted out by self driving car advocates.

Or is it perhaps a smokescreen to cover the idea that, rather then paying for expensive parking, owners will just have their cars drive around for a few hours, or for longer periods drive home and then back again when needed, say for the length of a working day?  So causing massively more congestion.

I think it will work in much the same way as streaming music and films destroyed the ownership of physical CDs and the status symbol of a record/CD/film collection.

Initially self driving cars will be expensive and will therefore be status symbols themselves, as the technology matures and the price drops a lot of people will realise that they can summon a high end self driving car for all their driving needs for the same price as owning a fairly mundane car.

So they get the status at no cost.

Eventually self driving cars will become so common that they will be far far cheaper to use than your own car.

Added to that the convenience and lack of stress and you've got a lot of reasons to abandon car ownership altogether.

ktache wrote:

Can anyone please explain why making an automobile self driving would result in a paradigm shift in car ownership, so that no one will own their own status symbol and instead be prepared to be driven around in whatever was on offer?

Is there any evidence for this assumption?

It does seem to be regularly trotted out by self driving car advocates.

Or is it perhaps a smokescreen to cover the idea that, rather then paying for expensive parking, owners will just have their cars drive around for a few hours, or for longer periods drive home and then back again when needed, say for the length of a working day?  So causing massively more congestion.

Cars spend well over 90% of their time sat still doing nothing. There's far too many of them taking up space doing nothing. If they're self driving they can go where they're needed as and when and you need fewer of them. That's the idea. You get one to take you to the shops then it picks someone up leaving town. That's why uber want in on this - it's their exact business model without the expensive human element.

ktache wrote:

Can anyone please explain why making an automobile self driving would result in a paradigm shift in car ownership, so that no one will own their own status symbol and instead be prepared to be driven around in whatever was on offer?

Is there any evidence for this assumption?

It does seem to be regularly trotted out by self driving car advocates.

Or is it perhaps a smokescreen to cover the idea that, rather then paying for expensive parking, owners will just have their cars drive around for a few hours, or for longer periods drive home and then back again when needed, say for the length of a working day?  So causing massively more congestion.

I would contend that a choice of non car ownership on the premise that you can rent one at a moments notice or have your stuff delivered is already widespread. Certainly among a younger generation of city dwellers who are very comfortable using Uber, Deliveroo and even bicycle hire schemes.

Removing the driver is completely logical for many of those same use cases.

Add in the rise of electric vehicles and the fact that many town dwellers will not have access to practical on street charging, or space to store a vehicle on their own property and buying into a car share scheme could be both practical and cheaper for many people.

Sure the wealthy will probably behave exactly as you describe but isn't that what chauffeurs already do?

Personally I have too many cars and motorcycles which I don't really use, but I'm of an older generation and I prefer cycling for my short work commute. My children have very little interest in learning to drive, they don't see it representing the same freedoms as I did at the same age. As I get older I'm absolutely looking forward to being driven around in an autonomous vehicle and am very open to the idea of shared ownership schemes if they were available where I live.

hawkinspeter wrote:

How about we skip autonomous cars and go straight to autonomous flying cars? I reckon it's easier to get flying cars to miss each other (and birds and planes) up in the air and it'd free up the roads for cyclists/walkers/joggers/street parties.

I approve of this idea. Especially the street parties. 

Rich_cb wrote:

The Uber fatality was the result of the failure of two systems. 1. The AI system. 2. The human driver. Using this example to demand the end to the development of system 1 in order to continue using system 2 indefinitely seems a bit short sighted. The market leader in driverless cars has already launched a completely driverless taxi service in Phoenix, Arizona. They have also clocked up 10 million driverless miles without, AFAIK, any deaths or serious injuries.

One Taxi Service in one city-After how many years -And what busy metropolis were the 10 milion miles clocked up in? Don't hold your breath waiting for this miracle.

Argus Tuft wrote:

One Taxi Service in one city-After how many years -And what busy metropolis were the 10 milion miles clocked up in? Don't hold your breath waiting for this miracle.

Ten years from founding to launching a completely driverless taxi service.

If you think that's slow progress then you clearly have no understanding of the challenges involved.

The miles were clicked up in many places but mainly silicon valley and Phoenix, Arizona.

Rich_cb wrote:

Argus Tuft wrote:

One Taxi Service in one city-After how many years -And what busy metropolis were the 10 milion miles clocked up in? Don't hold your breath waiting for this miracle.

Ten years from founding to launching a completely driverless taxi service. If you think that's slow progress then you clearly have no understanding of the challenges involved. The miles were clicked up in many places but mainly silicon valley and Phoenix, Arizona.

I do have an Idea of the challenges involved.That's why level 5 autonomy  on any road may never happen.

Argus Tuft wrote:

I do have an Idea of the challenges involved.That's why level 5 autonomy  on any road may never happen.

 

Do we need level 5?

If urban/suburban journeys and motorway journeys are autonomous then that's the vast vast majority of car journeys covered and the vast majority of benefits already delivered.

If you understand the challenges involved why do you consider achieving level 4 autonomy in a decade to be poor progress?

Rich_cb wrote:

Argus Tuft wrote:

I do have an Idea of the challenges involved.That's why level 5 autonomy  on any road may never happen.

 

Do we need level 5? If urban/suburban journeys and motorway journeys are autonomous then that's the vast vast majority of car journeys covered and the vast majority of benefits already delivered. If you understand the challenges involved why do you consider achieving level 4 autonomy in a decade to be poor progress?

Just to shift the subject a little,do you see these cars as being single occupant vehicles,or will they pick up passengers along the way?

Argus Tuft wrote:

Just to shift the subject a little,do you see these cars as being single occupant vehicles,or will they pick up passengers along the way?

I think we'll see a mix, smaller single occupancy vehicles will have a place but most cars will probably be roughly the size of cars now but split into 4 dividable sections, if you're travelling as a group the divisions will be open, if you're travelling individually your section will be private.

You'll be able to pay more to go straight to your destination without any stops, slightly less to stop once en route, less again to stop twice etc.

You could even pre book the pick ups so you can share lifts too and from work etc.

You didn't answer my previous question, do you think level 5 is necessary? I think level 4 will deliver almost all of the benefits far earlier, I'm not sure level 5 will ever be needed.