Author Topic: EEVblog #1066 - Uber Autonomous Car Fatality - How?  (Read 34514 times)

0 Members and 1 Guest are viewing this topic.

Offline StillTrying

  • Super Contributor
  • ***
  • Posts: 2673
  • Country: se
  • Country: Broken Britain
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #200 on: March 24, 2018, 05:56:54 pm »
Perhaps the car was busy at the time.
"Updating Firmware, Do not switch off, brake or change lanes."
CML+  That took much longer than I thought it would.
 
The following users thanked this post: BrianHG

Offline wraper

  • Supporter
  • ****
  • Posts: 13198
  • Country: lv
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #201 on: March 24, 2018, 06:17:35 pm »

It will hopefully be a long time before autonomous vehicles get the same level of situational awareness and precise vehicle control of a good human driver...


That human driver retard died in a car crash (in same car as on the video) 4 months after this video was put on youtube :horse:. It wasn't him driving that time, but it's pretty safe to assume car was driven in the same style. Yet another Darwin award.

http://www.dailymail.co.uk/news/article-2348365/Infamous-Georgian-street-drifter-raced-packed-city-streets-viral-videos-killed-car-crash.html
« Last Edit: March 24, 2018, 06:22:48 pm by wraper »
 

Offline mikerj

  • Super Contributor
  • ***
  • Posts: 2737
  • Country: gb
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #202 on: March 24, 2018, 10:02:49 pm »
HAHA! Is this some 1960's footage taken from a 32 HP Citro├źn 2CV?

1976.  It's a fairly infamous short film called "C'├ętait un rendez-vous".  The sound track is from a Ferrari and was dubbed onto the film, the actual car was Mercedes.
 

Offline G7PSK

  • Super Contributor
  • ***
  • Posts: 3746
  • Country: gb
  • It is hot until proved not.
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #203 on: March 24, 2018, 10:16:51 pm »
Every one is talking about Lidar. This car also had Radar the pedestrian was pushing a good radar target, unless of course the bike was a fancy racing bike made from carbon fibre. surely the radar should have picked up the bike if not why not.
 

Offline taydin

  • Frequent Contributor
  • **
  • Posts: 515
  • Country: tr
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #204 on: March 24, 2018, 10:18:07 pm »
That human driver retard died in a car crash

Good riddance :) The most favorable outcome, he could have caused others to die with his reckless driving.
Real programmers use machine code!

My hobby projects http://mekatronik.org/forum
 

Offline SilverSolder

  • Super Contributor
  • ***
  • Posts: 5072
  • Country: 00
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #205 on: March 24, 2018, 10:44:29 pm »
Here's another traffic situation that might cause "issues" for autonomous vehicles....  Got to love how they don't even bother painting lines on the road, those are for amateurs.

https://youtu.be/pLUm3Q-7iZA
« Last Edit: March 24, 2018, 10:46:43 pm by SilverSolder »
 

Offline coppice

  • Super Contributor
  • ***
  • Posts: 6388
  • Country: gb
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #206 on: March 24, 2018, 11:13:00 pm »
Every one is talking about Lidar. This car also had Radar the pedestrian was pushing a good radar target, unless of course the bike was a fancy racing bike made from carbon fibre. surely the radar should have picked up the bike if not why not.
The radar on most autonomous cars is quite crude. I assume the Uber one is no more sophisticated than others which are well documented. The simplest ones use a fixed (no scanning) broad forward facing beam. The better ones use a fixed broad beam for close in detection, and a fixed narrower beam for further out detection. In both cases they will pick up a lot of echo from stationary objects outside the car's lane. Therefore they have to be made fairly insensitive to near-zero doppler objects. If you look at videos of radar guided emergency braking systems on production cars, they only stop when approaching a stationary target if the car is fairly slow. If the car approaches a much slower car, the slower car creates enough doppler shift to stand out from the stationary objects. You will find videos of, say, a car at 100kph approaching one at 50kph, and slowing nicely to keep a proper distance from the 50kph car.
 

Offline grumpydoc

  • Super Contributor
  • ***
  • Posts: 2821
  • Country: gb
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #207 on: March 24, 2018, 11:44:10 pm »
If you are a bit squimish about fast driving, do not watch this film.



Admittedly, the streets were fairly empty in this film.
It probably looks faster than it is in reality - for instance the section between the Arc de Triomphe and the Place de la Concorde - a distance of almost exactly 2km is covered in about 1 minute 6 seconds (1:29 to 2:35 in the video) or roughly 110kmph(68 mph) - a bit reckless and not likely to be viewed kindly by the authorities should you try to re-create the stunt but given the very light traffic conditions encountered not that extreme. I have to say that the driver seems to have an impressive disregard for red lights.

I'm not sure it would be possible at all these days though.
 

Offline Marco

  • Super Contributor
  • ***
  • Posts: 5135
  • Country: nl
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #208 on: March 25, 2018, 12:37:36 am »
The radar on most autonomous cars is quite crude.

60 GHz scanning imaging radar using the same concept as LIDAR would be nice (for a HUD display in a car, not autonomous murder machines). Neither fog nor rain would really bother it.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 12257
  • Country: 00
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #209 on: March 25, 2018, 12:39:48 am »
It will hopefully be a long time before autonomous vehicles get the same level of situational awareness and precise vehicle control

Why "hopefully"?

of a good human driver...

I wouldn't hold that up as an example of skillful driving.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 12257
  • Country: 00
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #210 on: March 25, 2018, 12:41:29 am »
It probably looks faster than it is in reality

...as noted in a previous post.
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 21140
  • Country: nl
    • NCT Developments
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #211 on: March 25, 2018, 01:12:13 am »
Here's another traffic situation that might cause "issues" for autonomous vehicles....  Got to love how they don't even bother painting lines on the road, those are for amateurs.
https://youtu.be/pLUm3Q-7iZA
That traffic moves much slower. In such situations (been there, done that) driving is much easier. Point the car in the direction you want to go and don't hit anything. If something comes up you can brake in time because you are only travelling at 10km/h.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Online Brumby

  • Supporter
  • ****
  • Posts: 10899
  • Country: au
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #212 on: March 25, 2018, 01:35:32 am »
of a good human driver...

I wouldn't hold that up as an example of skillful driving.

Neither would I.

Certainly it is an exercise of accurate vehicle positioning - but it is based on a comparatively predictable environment.  Such driving reduces the margin of safety and adding just ONE unexpected motion of another object and you are in trouble.

AV technology could be trained to drive like this - but it would not find itself being received too warmly.
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 21140
  • Country: nl
    • NCT Developments
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #213 on: March 25, 2018, 01:45:19 am »
It will hopefully be a long time before autonomous vehicles get the same level of situational awareness and precise vehicle control of a good human driver...
Most (99.9%) human drivers have ZERO situational awareness and ZERO predictability skills. Most modern cars already have ESP which does the precise vehicle control for the driver when necessary.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline SilverSolder

  • Super Contributor
  • ***
  • Posts: 5072
  • Country: 00
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #214 on: March 25, 2018, 02:08:57 am »

That traffic moves much slower. In such situations (been there, done that) driving is much easier. Point the car in the direction you want to go and don't hit anything. If something comes up you can brake in time because you are only travelling at 10km/h.

Think how many objects the computer has to track, and predict, in that scenario...  it almost amounts to a denial of service attack...
« Last Edit: March 25, 2018, 02:18:21 am by SilverSolder »
 

Offline pantagruel

  • Newbie
  • Posts: 2
  • Country: us
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #215 on: March 25, 2018, 02:11:31 am »
There was an engineer of sensors for autonomous cars on the radio the other day - very interesting discussion about this accident . Here is the link if anyone wants to listen to it .
https://www.wpr.org/shows/effects-and-ethics-driverless-vehicles
 
The following users thanked this post: SilverSolder, ez24

Offline SilverSolder

  • Super Contributor
  • ***
  • Posts: 5072
  • Country: 00
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #216 on: March 25, 2018, 02:15:09 am »
It will hopefully be a long time before autonomous vehicles get the same level of situational awareness and precise vehicle control

Why "hopefully"?


Today, young guys hop up their cars by tuning engines, suspensions, brakes, etc...

In the future, the youngsters will download illegal software that makes their autonomous cars drive like our skilled, but misguided, (and dead) Georgian driver... in addition to tuning the engine, suspension, brakes, etc. ...   

« Last Edit: March 25, 2018, 02:18:53 am by SilverSolder »
 

Offline ez24

  • Super Contributor
  • ***
  • Posts: 3082
  • Country: us
  • L.D.A.
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #217 on: March 25, 2018, 03:26:38 am »
My two cents from reading links posted by members:

- I read that Arizona does not allow employers to ask if the person was a  felon -  so the felon issue is off the books.  But I bet it will be back on because of all the damage this has caused.

-  I think the police video is "fishy".    I think the police chief called the governor first and they discussed the video and it was "released" in the effort to show it was not Uber's fault because of the "warm" relationship between the governor and Uber.  I think the video is a misrepresentation of the lighting and the police knew this.  Money talks.

-  Even though Arizona is in the desert, there are fishy things going on there.
YouTube and Website Electronic Resources ------>  https://www.eevblog.com/forum/other-blog-specific/a/msg1341166/#msg1341166
 

Offline Bud

  • Super Contributor
  • ***
  • Posts: 4755
  • Country: ca
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #218 on: March 25, 2018, 04:36:27 am »
Quote from: nctnico
Most (99.9%) human drivers have ZERO situational awareness and ZERO predictability skills.
What do you guys smoke  there in Netherlands?
Facebook-free life and Rigol-free shack.
 

Online Monkeh

  • Super Contributor
  • ***
  • Posts: 7098
  • Country: gb
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #219 on: March 25, 2018, 04:53:12 am »
Quote from: nctnico
Most (99.9%) human drivers have ZERO situational awareness and ZERO predictability skills.
What do you guys smoke  there in Netherlands?

I have no idea, but again, they kill almost half as many people on the roads, so they're doing something right.
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 14942
  • Country: us
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #220 on: March 25, 2018, 05:20:49 am »
Quote from: nctnico
Most (99.9%) human drivers have ZERO situational awareness and ZERO predictability skills.
What do you guys smoke  there in Netherlands?

I have no idea, but again, they kill almost half as many people on the roads, so they're doing something right.

Seems I read that cars are extremely expensive there, so it's quite possible that a much smaller percentage of the population owns a car. Those that do are likely to be better trained and drive more carefully.
 

Online Monkeh

  • Super Contributor
  • ***
  • Posts: 7098
  • Country: gb
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #221 on: March 25, 2018, 05:33:13 am »
Quote from: nctnico
Most (99.9%) human drivers have ZERO situational awareness and ZERO predictability skills.
What do you guys smoke  there in Netherlands?

I have no idea, but again, they kill almost half as many people on the roads, so they're doing something right.

Seems I read that cars are extremely expensive there, so it's quite possible that a much smaller percentage of the population owns a car. Those that do are likely to be better trained and drive more carefully.

Households owning cars:
Canada: 84.4%
USA: 95%
UK: 73%
Netherlands: Approximately 75%

Fatalities per 100,000 population:
Canada: 6.0
USA: 10.4
UK: 2.9
Netherlands: 3.4

So you have a mild point, but it doesn't make up the numbers.

Perhaps Americans and Canadians could learn to drive more carefully? Again, we all seem to be doing something right.

ps. my car is worth less than a new iPhone.
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 14942
  • Country: us
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #222 on: March 25, 2018, 06:12:40 am »
Those numbers alone don't tell the whole story. 75% of households own cars, but how many have multiple cars and how many miles do they drive? Do teenagers drive cars around? Are drivers there better trained? Are the roads different? Cultural aspects? There are countless possible factors.
 

Offline maginnovision

  • Super Contributor
  • ***
  • Posts: 1960
  • Country: us
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #223 on: March 25, 2018, 08:14:09 am »
Using the above numbers it is interesting to note the correlation between fatalities and %household ownership. However there is no cause, population density isn't mentioned, age of vehicles, %of vehicles with current required safety equipment... As a former BMW technician, and before that all makes/models, I can say alot of people drive cars that feel like death traps. It's also important to note that smart roads will never happen on a large scale due to price and the ease with which they could be tampered with.
 

Offline Decoman

  • Regular Contributor
  • *
  • Posts: 161
  • Country: no
Re: EEVblog #1066 - Uber Autonomous Car Fatality - How?
« Reply #224 on: March 25, 2018, 08:26:07 am »
There was an engineer of sensors for autonomous cars on the radio the other day - very interesting discussion about this accident . Here is the link if anyone wants to listen to it .
https://www.wpr.org/shows/effects-and-ethics-driverless-vehicles

Wow, I stopped listening after a minute. I couldn't stand the improper attitude I thought this guy had, who sort of implies that one would normally think that this death is "devastating" to the industry. But, you guys go ahead.
"Q: (...) What are you thinking about it?"
"Well first of all.. the loss of life is.. a.. it is a.. devastating thing.. and not just to the autonomous vehicle industry, but to everybody.. and ah.. we really need to scrutinize why something like this has happened.. (...)"

This opening statement has elements which is the reverse of what I would expect as a sensible reaction in the aftermath of a traffic death, regardless of it being about new technology. The victim's family would have to come first (unless you don't care about human life), then secondly anyone else having been directly impacted by this traffic death, then well after that would be the autonomous vehicle industry; however this guy seemed to imply that one would be thinking about the car industry by mentioning it first, which I found very odd because doing so is also suggestive, as if being an authority on reacting to things. Also, the very idea of this traffic related death being a 'devastating thing' to "everybody" is another red flag in my brain, which to me seem to be what one would hear from someone being insincere, as if making exaggerations and avoiding responding in a sensible manner. It is ofc entirely possible that the guy is not used to talking, nor, writing, which I find difficult to believe given how this podcast is supposed to be discussing the ethics of driverless vehicles, with someone that is apparently involved professionally with working with things to some degree, and I would have expected more precise statements in this regard. And there are other flags for me: the phrase "we really need to scrutinize" is another type of exaggeration, that doesn't lend well to sensible arguments, which rely on an emotional reaction, but not logic, as the point about this "need"is not explained at all. And then, the phrase "why something like this has happened" seems yet again to be this aversion to making sensible arguments. This might ofc just another dumb north american thing, like when they say "you need to do this" in movies, or variations of such expressions. Such points are not only illogical, because of how the meaning of 'need' seems suspect, when needs tend to be a very personal thing, especially when referred to as a specific thing as such (someone having a need), they would in real life be potentially insulting or intimidating/commanding, because of how such expressions can only be understood as an abbreviation of some longer argument, but which is never made if not made explicitly clear in the same sentence, or immediately after, or specifically referred back to later on. Another contention to the use of "need" in language, is that whenever a point is made about anyone's "need", then if the meaning of a "need" imply there being the very same need for everybody, then that is something imo deeply insincere and offensive in a way, as it dehumanizes people in making the notion of such needs non-personal, as if you had no say in it perhaps, or that a particular 'need' is imparted onto you by others, which again would be unreasonable on the merit of it being non personal in the first place.
« Last Edit: March 25, 2018, 08:38:06 am by Decoman »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf