-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Any details on headline - Arizona pedestrian is killed by Uber self-driving car
- Thread starter brkaus
- Start date
-
- Tags
- Autonomous Vehicles
Thank you for pointing that out to me! Woah, that street really isn't that dark and the visibility is great.
Seattle Tom
Member
Tam
Well-Known Member
Not so fast apparently!
There are still family members, mother, father and son, coming forward:
More family members of woman killed in Uber self-driving car crash...
Probably not a fair comment without knowing the circumstances. The victim had had a few years there with drug arrests, 6 or more I think, and may have had other issues like mental. Many times people try to help out family members but especially when drugs are involved that isn't always possible. From what I've read and even seen in news stories many homeless prefer to live on the street over chosing some shelter or even their family home where interpersonal issues come into play. We also don't know if some of the family might have been checking in on her and taking her things or food she needed. I did read a comment from a friend of hers that she had been attempting to get a job recently.
Tam
Well-Known Member
Nvidia announced partnership with Uber in January but I guess it does not want to take credits for this accident.
Not Our Self Driving System, Says NVIDIA After Fatal Uber Crash
Nvidia CEO Jensen Huang:
"Uber does not use NVIDIA DRIVE technology. Uber develops its own sensing and drive technology, Uber developed their own sensing and driving applications and software."
Not Our Self Driving System, Says NVIDIA After Fatal Uber Crash
Nvidia CEO Jensen Huang:
"Uber does not use NVIDIA DRIVE technology. Uber develops its own sensing and drive technology, Uber developed their own sensing and driving applications and software."
Tam
Well-Known Member
Interesting Opinion criticizing weaker Arizona Autonomous Vehicle rules.
Also, it cites a statistic of 1 annual fatality per 100 million miles of human driving VS 10 million miles of Autonomous Vehicle driving.
That means currently, human is 10 times safer than Autonomous Vehicles are and Autonomous Vehicles are 10 times worse than human is!
Updated: link corrected
Ducey's Drive-By: How Arizona Governor Helped Cause Uber's Fatal Self-Driving Car Crash
Also, it cites a statistic of 1 annual fatality per 100 million miles of human driving VS 10 million miles of Autonomous Vehicle driving.
That means currently, human is 10 times safer than Autonomous Vehicles are and Autonomous Vehicles are 10 times worse than human is!
Updated: link corrected
Ducey's Drive-By: How Arizona Governor Helped Cause Uber's Fatal Self-Driving Car Crash
Last edited:
Tam
Well-Known Member
Sorry for the wrong link. Here it is as corrected:
Ducey's Drive-By: How Arizona Governor Helped Cause Uber's Fatal Self-Driving Car Crash
Tam
Well-Known Member
Report: Software bug led to death in Uber’s self-driving crash
According to 2 anonymous sources, the sensors were detecting the pedestrian fine but the software classified her as "false positive."
It's a choice for programmers to choose between a more dangerous way of providing smoother Uber experience VS more cautious and safer way but getting complaints for Tesla phantom brakes.
According to 2 anonymous sources, the sensors were detecting the pedestrian fine but the software classified her as "false positive."
It's a choice for programmers to choose between a more dangerous way of providing smoother Uber experience VS more cautious and safer way but getting complaints for Tesla phantom brakes.
Tam
Well-Known Member
Another interpretation with the same conclusion:
Uber’s self-driving software detected the pedestrian in the fatal Arizona crash but did not react in time
That means both sensors and software detected the pedestrian fine but the settings from software programmers have allowed such accident to happen.
If they didn't set their software to such settings:
It sounds like Uber has chosen smoothness and comfort over safety.
Uber’s self-driving software detected the pedestrian in the fatal Arizona crash but did not react in time
That means both sensors and software detected the pedestrian fine but the settings from software programmers have allowed such accident to happen.
If they didn't set their software to such settings:
It sounds like Uber has chosen smoothness and comfort over safety.
Last edited:
Maybe smoothness, but I'm still going with the leading and trailing wheel spokes (along with small tube cross section) caused the rain/ snow filter to activate thus 'seeing' through the pedestrian to the road beyond.
Tam
Well-Known Member
That's a good explanation and theory.
However, according to the articles, there's nothing wrong with the sensors.
It's the immatured Uber software that is not advanced enough to classify pedestrian/bicycle as important obstacles to avoid.
LIDAR can see 3-dimensional. It can measure that a pedestrian/bicycle has thickness for software engineers to write a code to avoid hitting it and not just a paper thin flat object that software engineers to write a code to run over it.
On the other hands, for years, other companies like Waymo have been able to recognize and classify many objects including pedestrians and bicyclist below:
Last edited:
That's a good explanation and theory.
However, according to the articles, there's nothing wrong with the sensors.
It's the immatured Uber software that is not advanced enough to classify pedestrian/bicycle as important obstacles to avoid.
LIDAR can see 3-dimensional. It can measure that a pedestrian/bicycle has thickness for software engineers to write a code to avoid hitting it and not just a paper thin flat object that software engineers to write a code to run over it.
On the other hands, for years, other companies like Waymo has been able to recognize and classify many objects including pedestrians and bicyclist below:
My thinking is the application processing of the LIDAR, not the sensor itself/ raw output. Assuming rain/ snow handling is done outside the raw point cloud, the Uber SW could ignore the region of the pedestrian due to the near/far return noise of hitting the spoke/ rims. In your image, the bicyclist tires are barely there and spoke not at all. Excessive squelch could blank the region.
Just a guess.
A sample size of 1 is hardly meaningful, and there is some question whether Uber was testing the system without LIDAR input. They would save a lot of money if they could drop the LIDAR.
If you count automobile occupant deaths and injuries, right now AV is winning. It doesn't get drunk and run into trees very often.
Preliminary report from NTSB is out for this crash: https://www.ntsb.gov/investigations/AccidentReports/Reports/HWY18MH010-prelim.pdf
This is a good reminder that folks operating experimental stuff on the roads shouldn't be allowed to more-or-less entirely self-regulate. The way Uber set up its test cars in Arizona was absolutely insane, and bordering on criminal:
"The vehicle was factory equipped with several advanced driver assistance functions by Volvo Cars, the original manufacturer. The systems included a collision avoidance function with automatic emergency braking, known as City Safety, as well as functions for detecting driver alertness and road sign information. All these Volvo functions are disabled when the test vehicle is operated in computer control but are operational when the vehicle is operated in manual control.
According to Uber, the developmental self-driving system relies on an attentive operator to intervene if the system fails to perform appropriately during testing. In addition, the operator is responsible for monitoring diagnostic messages that appear on an interface in the center stack of the vehicle dash and tagging events of interest for subsequent review.
* * *
According to data obtained from the self-driving system, the system first registered radar and LIDAR observations of the pedestrian about 6 seconds before impact, when the vehicle was traveling at 43 mph. As the vehicle and pedestrian paths converged, the self-driving system software classified the pedestrian as an unknown object, as a vehicle, and then as a bicycle with varying expectations of future travel path. At 1.3 seconds before impact, the self-driving system determined that an emergency braking maneuver was needed to mitigate a collision (see figure 2). According to Uber, emergency braking maneuvers are not enabled while the vehicle is under computer control, to reduce the potential for erratic vehicle behavior. The vehicle operator is relied on to intervene and take action. The system is not designed to alert the operator.”
This is a good reminder that folks operating experimental stuff on the roads shouldn't be allowed to more-or-less entirely self-regulate. The way Uber set up its test cars in Arizona was absolutely insane, and bordering on criminal:
"The vehicle was factory equipped with several advanced driver assistance functions by Volvo Cars, the original manufacturer. The systems included a collision avoidance function with automatic emergency braking, known as City Safety, as well as functions for detecting driver alertness and road sign information. All these Volvo functions are disabled when the test vehicle is operated in computer control but are operational when the vehicle is operated in manual control.
According to Uber, the developmental self-driving system relies on an attentive operator to intervene if the system fails to perform appropriately during testing. In addition, the operator is responsible for monitoring diagnostic messages that appear on an interface in the center stack of the vehicle dash and tagging events of interest for subsequent review.
* * *
According to data obtained from the self-driving system, the system first registered radar and LIDAR observations of the pedestrian about 6 seconds before impact, when the vehicle was traveling at 43 mph. As the vehicle and pedestrian paths converged, the self-driving system software classified the pedestrian as an unknown object, as a vehicle, and then as a bicycle with varying expectations of future travel path. At 1.3 seconds before impact, the self-driving system determined that an emergency braking maneuver was needed to mitigate a collision (see figure 2). According to Uber, emergency braking maneuvers are not enabled while the vehicle is under computer control, to reduce the potential for erratic vehicle behavior. The vehicle operator is relied on to intervene and take action. The system is not designed to alert the operator.”
Those three facts together -- wow what a stupid way to test their system.
I think the term is negligent. Negligence: Carelessness which causes harm to life.
Similar threads
- Article