Dashcam footage released Uber Volvo XC accident in AZ

[100Coogn]

Well, yea, if you step in front of a car.  Duh.  Vegas has a huge number of pedestrian deaths and almost all of them are idiots committing "misdemeanor" offenses and getting the death penalty for it.  Cross in the middle of a high speed multi-lane street wearing no lights and dark clothing, and you get what you get.

I watched the video.  I think a regular car and driver probably would have hit her even harder since most drivers drive well over the speed limit.  No lights, crossing in a very dark area of a high speed multi-lane road, stepping out right in front of a moving car...  Yea, you get death for that pretty much everywhere.  MANY fatalities every year in Vegas for that exact same situation, with regular alert drivers behind the wheel.

Sure, a "good" autodrive system should have seen that person, but was there even an IR signature from the pedestrian?  Wearing enough clothes and pushing a cold bike, she might have not even shown up in an IR scene.  Visible light... no way.  Lidar?  The pedestrian was hustling so she may have been just outside the left edge of the critical area in a frontal collision avoidance area, until she ran right in front of the car.  So...  Yea, a better system maybe could have used multi-spectral imagery or radar/lidar to see and avoid a pedestrian doing her best to run in front of a moving car, but I think it's a stretch to blame the car here since I highly doubt a regular driver could have done anything different.

I think the lesson I'm taking away is to not run across the middle of a dark high speed multi-lane road right in front of a car.

At least a pedestrian would have a chance with human driver.  Some may stop some may not.  All of these automated cars will be programmed the same.  'Run em all down!'
Now what if it's a dog/deer or your child in the road?

Coogan

[Skuzzy]

eagl, you missed a couple of points.

1)  She did not just step in front of the car.  She crossed one empty lane before getting to the lane the car was in, then crossed left to right, getting hit by the right front fender.  The camera footage has already been proven to be artificially darker than the real scene was, by several factors.  Was she wrong?  Sure, but she did not need to die because of that mistake.

2)  The LIDAR system Uber is claiming to use has a range of 500m and a 360 degree sweep.

Both the car and the safety driver failed.  A needless death was the result.

[Wiley]

eagl, you missed a couple of points.

1)  She did not just step in front of the car.  She crossed one empty lane before getting to the lane the car was in, then crossed left to right, getting hit by the right front fender.  The camera footage has already been proven to be artificially darker than the real scene was, by several factors.  Was she wrong?  Sure, but she did not need to die because of that mistake.

Failure all around in my opinion.  It's sad, but crossing a highway at night you should see vehicles for miles.

But the vehicle appeared to not react in any way, not good.

Wiley.

[Karnak]

https://arstechnica.com/cars/2018/03/leaked-data-suggests-uber-self-driving-car-program-years-behind-waymo/

It is even worse than I stated previously.  Even though Uber is using comparatively easy Arizona they average a safety driver intervention once every 13 miles.  Waymo, in comparison, is averaging a safety driver intervention once every 5600 miles in California and GM's Cruise program (the one actually bragging about using San Francisco, my bad) averages 1250 miles per intervention.

Uber is literally years behind Waymo and GM.

Also, remember the reason Uber is using Arizona and not California.  California required permits and data reporting, Uber said they didn't need permits, California put their foot down and Arizona told Uber to come do the testing in Arizona where they wouldn't be held back by bureaucracy.  It seems that Waymo and GM don't find California's permits excessively burdensome.  I suspect it wasn't the permits that Uber was trying to avoid, but rather the public reporting of data.  The 5600 mile per intervention and 1250 mile per intervention are publicly available data points, the 13 miles per intervention comes from leaked documents.

[Lusche]

I am still waiting for the promised arrival of the atomic car! 

[Brooke]

I am still waiting for the promised arrival of the atomic car! 

(Image removed from quote.)

 

[icepac]

There wasn't even any sidewalk or road from where she was coming.

Did she just walk that bike right out of that park on the left?

It reminds me of when Donte Stallworth got pounded for hitting a guy who was walking across the lanes of this causeway.

Sure, stallworth deserved partial blame since he would not have hit the guy if he weren't driving (he was buzzed) but who knows if a designated driver wouldn't have hit the guy if you put the car and pedestrian into the same sequence of events.

[Wiley]

There wasn't even any sidewalk or road from where she was coming.

Did she just walk that bike right out of that park on the left?

It reminds me of when Donte Stallworth got pounded for hitting a guy who was walking across the lanes of this causeway.

Sure, stallworth deserved partial blame since he would not have hit the guy if he weren't driving (he was buzzed) but who knows if a designated driver wouldn't have hit the guy if you put the car and pedestrian into the same sequence of events.

(Image removed from quote.)

I think the main problem people are having with it is it seems like exactly the kind of thing an autonomous vehicle should be great at detecting.  As Skuzzy said upthread, the LIDAR theoretically has a range of 500 yards.

A human might've been looking the wrong way at the moment the pedestrian became visible or been slow to react.  The way AVs have been sold to the public is that it won't get distracted, its reaction times are way better than human, and blahblahblah.  Then you see what should be a basic, incredibly simple part of the program "If there's something in the road, avoid it." and it failed spectacularly.

It doesn't matter what crosswalk was where.  The fact that there was a living obstacle in the road and the car didn't even seem to twitch is the issue.

Wiley.

[Delirium]

The populace said cars were too dangerous when they started to appear on the streets in the early 20th century. Some states required someone to walk in front of the car with a red flag for safety reasons. Eventually, people accepted the change and cars began to travel ever faster on improved roads.

Driverless cars will come onto the scene in much the same way. Sure, avoid it as long as you can much like the populace avoided elevators and opted to take the stairs, but no one will be able to stop the flood of technology (short of a nuclear war).

I don't know about the rest of the group here, but I'm looking forward to taking a road trip in a camper one day without actually having to drive the trip itself! (I know we are many years from that)

[Wiley]

I don't know about the rest of the group here, but I'm looking forward to taking a road trip in a camper one day without actually having to drive the trip itself! (I know we are many years from that)

It'll be swell when it works.  The way there is going to have bumps though.

My main problem with the above footage and the stuff Karnak posted is it appears their product isn't up to snuff, but they're basically gambling with peoples' lives in the meantime.

Wiley.

[Karnak]

Update:

Arizona tells Uber they can no longer test autonomous cars in the state:

https://arstechnica.com/cars/2018/03/uber-told-to-stop-testing-driverless-tech-in-arizona/

[FBKampfer]

Get em Google, they're vulnerable  !

[ToeTag]

If this is ever going to work pedestrian traffic and autonomous traffic has to be seperated.  Much like the road system had to be completely redone for interstate travel by modern cars vs carriages, foot traffic and sleighs. Decades away from working.

[100Coogn]

I see some similarities here..



Coogan

[Karnak]

If this is ever going to work pedestrian traffic and autonomous traffic has to be seperated.  Much like the road system had to be completely redone for interstate travel by modern cars vs carriages, foot traffic and sleighs. Decades away from working.

GM and Waymo are testing them in San Francisco and other California locations.  There are lots of pedestrians there.  Pedestrians don't seem to be an issue for them, just Uber.  With proper programming the autonomous cars should be much safer than a human driver as they don't get distracted and lidar/IR sensors work regardless of illumination.

I doubt we are decades away.  We may not even be 5 years away.  It just won't be Uber's system that triumphs.