Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

Register Log in
  • Congratulations waross on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

2 dead in Tesla accident "Noone wasdrivingthe car" 15

Status
Not open for further replies.
MartinLe

MartinLe

Civil/Environmental
Oct 12, 2012
394
DE

“no one was driving” the fully-electric 2019 Tesla when the accident happened. There was a person in the passenger seat of the front of the car and in the rear passenger seat of the car.

the vehicle was traveling at a high speed when it failed to negotiate a cul-de-sac turn, ran off the road and hit the tree.

The brother-in-law of one of the victims said relatives watched the car burn for four hours as authorities tried to tap out the flames.

Authorities said they used 32,000 gallons of water to extinguish the flames because the vehicle’s batteries kept reigniting. At one point, Herman said, deputies had to call Tesla to ask them how to put out the fire in the battery.
 
Replies continue below

Recommended for you

Sort by date Sort by votes
@JRB, I've never considered these as software "failures" but rather, they are design and systems engineering failures. While Tesla has done a lot of impressive design, they have specifically eschewed lidar, which would have eliminated many issues associated with pure image processing, namely, "Is that a solid object that I'm about to hit?" and "Are those vaguely identifiable objects on a collision course with me?" This is very basic navigation and collision avoidance, but we obviously see at least three system designs where these decisions aren't even remotely embedded in the decision trees.

In the Florida collision with the semi trailer, as with the Uber collision with a pedestrian, everything is an instantaneous decision and the system does not make use of history, i.e., I saw something at time1, and then something else at time2, but moved, a human would think, "Oh, I can't figure out what the hell it is, but it's moving into my path," while both Autopilot and Uber's system say, "I see thing1", "I see thing2", "Oh, crap, thing3 is in front of my path." I've seen this sort of blind allegiance to the image processing without any attempt to connect the dots in time, even at my job, and it was a failure to not do systems engineering, nor address what the end objective of the image processing is for, which is not even remotely done when it finds objects, that's just the start of the process. In the Florida Autopilot case, lidar would have shown that the white side of the semi trailer was not sky, and therefore an collision was imminent, maybe.

In a case in San Jose, Autopilot lane followed by following only one set of lane markers, because the second set was unrecognizable, resulting in following the lane into a gore point. That would have been marginally OK, had the collision prevention cannisters been properly maintained, but there had been a prior accident, and the cannisters were empty, allowing the car to collide with the metal barriers. So, Tesla decided to not augment the lane following with map information, nor any information about the other cars, both of which would have indicated that the road was curving away from where Autopilot was steering toward. Again, lidar would have told Autopilot a collision was imminent, maybe.

I added "maybe" because it's not clear that Autopilot is designed to consider multiple sources of information; ironically, "sensor fusion" is much older that Tesla, but basic idea is what humans do to integrate information streams from their eyes and ears into a coherent story. Autopilot doesn't appear to be designed that way, because even without lidar, it has other information that it ignores to concentrate only on lane following. That makes it pretty stupid and not worthy of consideration, at this moment in history.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert! faq731-376 forum1529 Entire Forum list
 
Upvote 0 Downvote
@spsalso, that decision making typically is not part of an NN design; NNs are still pretty simplistic. If you look at a Tesla display, it shows individual objects detected and identified by the NN, so that means the NN is simply a front end processor, collecting data for some decision making system that may, or may not, be an NN.

There isn't much to interview or deconstruct in an NN, there's multiple collections of stored coefficients that are the product of training the NN, but it's not immediately obvious, nor subsequently obvious, why the coefficients are why they are and what they do in the processing. Unless there's a complete log of EVERYTHING, it may not even be possible to deconstruct the decision process

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert! faq731-376 forum1529 Entire Forum list
 
Upvote 0 Downvote
The only way they are going to get something like to work and not even then I would trust a autopilot unless it had some other safety system like laserscanners or such.
It is to combine the NN learning with brain wave research and eye movement.
Mainly letting people look at films from people driving cars and at the same time letting NN learn which objekts in the moving pictures that are important via the observers brainwaves, eye movements and what they are fixating on combined with speed brake and steering wheel movements of the vehicle.

/A


“Logic will get you from A to Z; imagination will get you everywhere.“
Albert Einstein
 
Upvote 0 Downvote
Astounding!

The whole POINT of being in a car is to constantly change one's location with respect to the surroundings. Hence the auto-drive should always be comparing surroundings over time.

Which absolute moron(s) at Tesla did not know this?

OR

Which absolute moron(s) knew this but discounted it?



spsalso
 
Upvote 0 Downvote
GregLocock,

This does not surprise me. I see no problems with robot cars in good weather and driving conditions. I don't how we and the police are going to signal them. My GPS shows off signs on the road quite well, so robots would be aware of them, as long as they are not new.

pointless_signs_ilvjqq.jpg

I do not see how a robot would detect black[ ]ice on the road ahead, possibly not a problem where you are, but definitely one where I am. Robot cars may have to be kept off the roads in bad weather.

--
JHG
 
Upvote 0 Downvote
The key for robot cars to function well is communication between the cars. Ideally, only one robot car will have to sacrifice itself to the black ice but it will report it to the others.
 
Upvote 0 Downvote
GregLocock said:
AAA tests various driver assist technologies in the simulated rain. Not very good.
Click to expand...
Big surprise. I don't know if these systems are the products of a bunch of game programmers used to working within a bounded universe that they themselves define, but they might as well be.

"Schiefgehen wird, was schiefgehen kann" - das Murphygesetz
 
Upvote 0 Downvote
One of my cars which uses a camera and not radar gives me a warning that it is not active in heavy rain in terms of collision avoidance. I turn the lane keeping off as it annoys me so maybe that wouldn't work either.

It is quite distracting to have this yellow warning symbol in heavy rain.

The radar one doesn't seem to be as affected. It's failing is when a motorcycle isn't in the middle of the lane it doesn't always pick it up...

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
 
Upvote 0 Downvote
The break system could detect black ice, at least after one break even if it isn't a full stopp.
Or even something measuring the temperatur difference between the road and the air temp.

“Logic will get you from A to Z; imagination will get you everywhere.“
Albert Einstein
 
Upvote 0 Downvote
Slowing to 30 km/h and stopping before ever going down a steep hill whenever the surface temperature is between -4 and 0 C ... isn't a viable option!
 
Upvote 0 Downvote
Not sure I follow you, who sade anything about 30 km/h?
Do you not have winter tires / studded tires. ;-)


“Logic will get you from A to Z; imagination will get you everywhere.“
Albert Einstein
 
Upvote 0 Downvote
I was driving with friends a while back. It was 1am, dry, and cold. We were cruising at 70mph and passed a car that was upsidedown in the trees next to the highway. Emergency responders were there so we thought nothing of it and proceeded on our way. Very shortly afterwards we were sliding across the highway at 70mph. My friend driving had managed to gather everything up and regain control. It was that moment we realized we had hit the same patch of black ice that had sent the other car in to the trees.
 
Upvote 0 Downvote
If a self driving car is driving where black ice can occur maybe it should not be allowed to go faster then a speed so it can stop before hitting something, what speed that is a guess depends on what kind of breaks and tires it has.
If it can't detect these things on it's own maybe the speed should always be limited and not just when it is necessary.

I have been driving on black ice twice this week going to work no fancy breaks and only summer tires, but here it is a no brainer since I had to get rid of the ice from the windshield before going, but this is just the way it is this time of the years so "people" know.
And of course I have to adjust my driving and speed to that fact.

“Logic will get you from A to Z; imagination will get you everywhere.“
Albert Einstein
 
Upvote 0 Downvote
RedSnake,

In nasty weather, I make a point of touching my brakes to see what the traction is like. This tests only the road I am on at the moment. Out in the distance, black ice looks... black, for some reason or other. I cannot see LiDAR spotting this, although it might look like a big hole in the road.

Perhaps we can discuss the algorithm for driving a car out of a snow drift. This is a challenge for lots of human drivers.

--
JHG
 
Upvote 0 Downvote
I haven't mentioned any LIDAR sensors.
My suggestion/thought was that, for example a laser temperatur device can measure the road temperatur in combination with another temp sensor measuring the air temp maybe in combination with a moisture sensor and or in combination with ABS breaks automatically testing the friction between the tires and the road.

Road temp > 0,5 C, Air temp > +5C equals slippery road.

I spånade as we say in Sweden, I guess that is almost the same as pondering. [ponder]

“Logic will get you from A to Z; imagination will get you everywhere.“
Albert Einstein
 
Upvote 0 Downvote
Status
Not open for further replies.

Similar threads

MJCronin
  • Locked
  • Question
"Boiler Tank" (sic) explodes in Mexico Tequilla Plant 9
2
Replies
38
Views
93
1503-44
1503-44
JohnRBaker
  • News
28 cars crash during a single accident at a NASCAR event...
Replies
10
Views
93
drawoh
drawoh
drwebb
  • Locked
  • Question
Cruise Service Suspended by GM in SF After (ancilliary) Pedestrian Accident 2
Replies
7
Views
39
IRstuff
IRstuff
X4vier
  • Locked
  • Question
Bridge "Los Grillos" collapse on Colombia. Progressive cantilever bridge... 3
2
Replies
32
Views
33
ACtrafficengr
ACtrafficengr
Alistair_Heaton
  • Locked
  • Question
car exploded 2
2
Replies
38
Views
39
LittleInch
LittleInch

Part and Inventory Search

Sponsor

Back
Top