Self-drive Uber identifies woman as false positive (which inc. plastic bags)

The text accompanying that video includes the following statement:I wonder whether similar Dashcam footage would stop a human driver being charged with causing death by dangerous driving?

If you can't see far enough ahead to stop safely - you are driving too fast for the conditions is how I was taught to drive - but then, I don't drive for Uber.

ps - God only knows what the woman was thinking when crossing where she was, on a dark road.

Like I said, other dashcam footage of the same area at night shows it as being much brighter as there are lights all around the road,

I think it's just a camera with poor vision at night on the Uber.
 
The text accompanying that video includes the following statement:I wonder whether similar Dashcam footage would stop a human driver being charged with causing death by dangerous driving?

If you can't see far enough ahead to stop safely - you are driving too fast for the conditions is how I was taught to drive - but then, I don't drive for Uber.

ps - God only knows what the woman was thinking when crossing where she was, on a dark road.

There was a human in the driving seat of the vehicle, I'd hope he/she is charged. The woman shouldn't have been in the road but then again she didn't jump out on him/her, she was pushing a bicycle.

Then again the US places more emphasis on the priority of the motorist vs the pedestrian relative to the UK and has offences such as jay walking etc..etc...
 
I watched the film multiple times and from it I had a hard time seeing her until it was far too late. I haven't seen any other footage of the accident to indicate other levels of visibility.
Without knowing how well all the cameras were tuned for light sensitivity it's ultimately hard to say.
Its a terrible shame someone lost their life, but crossing in dark/darkish conditions with no lights, in the middle of the road with traffic travelling at speeds...? I don't think it's fair to call it Ubers fault or a flaw of the tech.
 
If you can't see far enough ahead to stop safely - you are driving too fast for the conditions is how I was taught to drive - but then, I don't drive for Uber.

So, How fast should one drive on an unlit road when using dipped headlights?

(Any unlit road, including DC and Motorways)

It is a tricky one really, To stop within the 25 yard or so range of dipped headlights, you would have to be driving at 30 or below. However If you were seen doing this on a motorway at night you would probably get prosecuted for dangerous driving.

Rightly or wrongly, Most night driving is done on faith. by Assuming that obstacles are themselves lit or that the vehicle 100 yards ahead has "Proved" that the road between you and the vehicle ahead is clear.

Which is why human drivers often find themselves making the same mistakes as the robot one did.
 
I watched the film multiple times and from it I had a hard time seeing her until it was far too late. I haven't seen any other footage of the accident to indicate other levels of visibility.
Without knowing how well all the cameras were tuned for light sensitivity it's ultimately hard to say.
Its a terrible shame someone lost their life, but crossing in dark/darkish conditions with no lights, in the middle of the road with traffic travelling at speeds...? I don't think it's fair to call it Ubers fault or a flaw of the tech.

It could have reacted inhumanly fast if given instructions to emergency brake. Also I believe these systems detect more than visible light so what you can see in a terrible quality dashcam is a red herring.

Given that the human staring at the floor managed to react more than the car (which didn't) it's safe to say someones work on collision detection was not fit for purpose. Who else to blame but Uber?
 

Looking at this video, would many people see her in time? It could have swerved I suppose?
I think that's the same video that has been questioned in regards to if it had been adjusted prior to release as other people have taken video of the same road at the same time of night and it's been much lighter.

Also most of the other self driving car companies use multiple sensor types, from memory Uber were deliberately trying to use less comprehensive sensor coverage than the other companies experimenting with the systems.

There are also reports that this isn't a one off failure, as people who lived on the routes the self driving cars were allowed to use (including residential streets) have stated they'd noticed the cars didn't seem to respond to people in the road (as you tend to get on residential streets), in the same manner they did for things like rubbish bins.
 
So, How fast should one drive on an unlit road when using dipped headlights? . . .
I believe that the current Highway Code advises:
Drive at a speed that will allow you to stop well within the distance you can see to be clear.
I accept that this is something of a cop-out and probably rarely applies in the real world.

. . . Which is why human drivers often find themselves making the same mistakes as the robot one did.
Amen to that!


Back with the autonomous Uber, I suspect (but don't know) that their collision avoidance system(s) do not depend on the same visual acuity as would be the case with a human. I imagine that they use something like radar, infra-red, motion detection or whatever - none of which would depend on lighting.
 
Compare the lighting to this one:


Well, the second picture has the camera car with the lights on full beam rather than dipped and there is another vehicle coming the other way also with lights on.

So it isn't really a fair comparison..

Oh and PS,

I don.t know if this is the case here.

But time of day matters,

Some places in the US switch off the street lighting over night to save money. so a street may be lit early evening but not later. What time was this accident?
 
Normally I'm all for self driving tech, accidents and fatalities are inevitable, however this one shouldn't have happened. The pedestrian was obviously at fault, you're responsible for your own safety, and she should have been looking for and reacting to oncoming danger, however I think uber and the safety driver were the worse offenders here.

The car should have had no problem identifying and stopping for the pedestrian, however uber are well known to be lagging in their self driving tech, recently they were reported to have on average a disengagement (a safety driver intervention) every 13 miles, compared to every 5600 miles with Waymo. They may have been testing reduced sensitivity to try and reduce their disengagement rate, which could have led to this accident, though we won't know until the investigation is complete.

The safety driver is also clearly at fault, what was she looking down at for the ~5 seconds before the accident? If she was looking at her phone then it's surely criminal negligence or worse. Lots of people have been arguing that you can't expect safety drivers to stay attentive for hours at a time, however with such a high disengagement rate you'd want them on their toes at all times, if they were failing to stay focused then uber should have implemented methods to reduce errors (Japanese train drivers are an excellent example of this, they work in pairs, calling out all speed signs and warnings, it's nearly impossible for them to miss anything or lose concentration).

The video uber released is a blatant attempt to shift blame, the poor brightness and contrast on the video is intentionally there to make it look inevitable. Additionally, they showed the distracted safety driver, another (pretty low) attempt to shift the blame. Actually surprised they didn't get in trouble with the NTSB, who generally forbid public release of evidence while an investigation is ongoing.

I wouldn't be sad to see uber abandon their self driving ambitions, an impatient, callous company is not one you want in charge of dangerous machines.
 
Her shoes first appear at 7 seconds in and impact is during the 8th second. Assuming the video captured is accurate to what a human would see, there’s no way anyone here wouldn’t have run her over too. Hell, reaction time and the time it takes to push the brake down is valuable hundredths of a second that you don’t have to spare here.
 
Her shoes first appear at 7 seconds in and impact is during the 8th second. Assuming the video captured is accurate to what a human would see, there’s no way anyone here wouldn’t have run her over too. Hell, reaction time and the time it takes to push the brake down is valuable hundredths of a second that you don’t have to spare here.

Bit of a dubious conclusion, a human driver isn't constrained to whatever the camera sees, they have an entire front window to look out of.
 
Back
Top Bottom