Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
Here's the video of the accident, I thought the LIDAR was supposed to see things like that. The car was going 40 mph in a 35 mph zone and didn't slow down at all:
You do know...a human being was at the wheel. These self driving cars are early prototypes designed to help them learn how to drive. The human being failed here, and folks are blaming the vehicle for not fixing the human error. That driver could have pressed the brake.
Are they safer than human driven cars? I would argue that the statistics say that Ubers version is not, but that Waymo (google) is. If you look at the data from California especially you can see that Uber is 3 years behind Waymo in how well their vehicle drives.
Which stresses even more that this was a human failure. The driver should have hit the brakes.
And you know what? The police who reviewed the video of the incident have stated that they do not believe the driver could have.
We'll see what the NTSB has to say. Tempe Police seem to take issue with what was printed earlier in the SF Chronicle. Here's a snippet from yesterday's Phoenix New Times (Glover is a Commander with the Tempe PD):
Glover downplayed Moir's statements, saying some were taken "out of context" by the Chronicle. The chief disagrees with the Chronicle's headline, "Tempe police chief says early probe shows no fault by Uber," Glover said. It's too early to say which party was more responsible, he said, and Tempe wants to "wait to reserve judgment."
With all due respect, I don't believe the standard here should be "what could a human have done". An AV should easily be able to detect an object the size of a human/bike combo on an empty street, day or night. Identifying it is another matter, but certainly the car should try to avoid an object the size of a human no matter what it is. So again, let's have all the details in the report. Were there other cars on the street at the time? How fast and in which direction was the pedestrian walking? How far away was the car when the woman stepped onto the roadway?
Here's the video of the accident, I thought the LIDAR was supposed to see things like that. The car was going 40 mph in a 35 mph zone and didn't slow down at all:
They showed the video of the human driver while all this was going on and said human was looking down most of the time until right at the time of impact.
Why is said human paid to be in the car?
We'll see what the NTSB has to say. Tempe Police seem to take issue with what was printed earlier in the SF Chronicle. Here's a snippet from yesterday's Phoenix New Times (Glover is a Commander with the Tempe PD):
Glover downplayed Moir's statements, saying some were taken "out of context" by the Chronicle. The chief disagrees with the Chronicle's headline, "Tempe police chief says early probe shows no fault by Uber," Glover said. It's too early to say which party was more responsible, he said, and Tempe wants to "wait to reserve judgment."
With all due respect, I don't believe the standard here should be "what could a human have done". An AV should easily be able to detect an object the size of a human/bike combo on an empty street, day or night. Identifying it is another matter, but certainly the car should try to avoid an object the size of a human no matter what it is. So again, let's have all the details in the report. Were there other cars on the street at the time? How fast and in which direction was the pedestrian walking? How far away was the car when the woman stepped onto the roadway?
If you watch the video the person was almost across the roadway so it wasn't "just stepped in front of" or some fast movement. It was a dark section of roadway that on the video looked hard to see in but video is not human eye quality. Didn't matter because the human driver was not paying attention like she was PAID to do.
I'd also think that the car should be able to detect way better than any human as we don't have radar...
If you watch the video the person was almost across the roadway so it wasn't "just stepped in front of" or some fast movement. It was a dark section of roadway that on the video looked hard to see in but video is not human eye quality. Didn't matter because the human driver was not paying attention like she was PAID to do.
I'd also think that the car should be able to detect way better than any human as we don't have radar...
Exactly. This wasn't a dart out case. The woman was right in front of the car. I would think the sensors would have picked that up and braked. For some reason, they did not.
Yes, this woman appears to have been jaywalking. In a situation where you were comparing human negligence (or the negligence of a driver vs. the pedestrian) it looks like you'd have to put more blame on the pedestrian. However, that misses the point here. Self driving cars are being pushed largely on the notion that they can drive more safely than humans can. Therefore, the comparison about how a human would react is inappropriate.
Sure, this is just one accident. However, there haven't been that many miles driven yet by self driving cars. I'm not impressed by the fact this technology could not detect a woman in the middle of the road way in time to stop. Why didn't the radar pick her up? She had crossed most of the road way before being struck?
I note this woman was wearing a dark shirt, this was not at an intersection, and it is somewhat dark before the collision occurs. This is exactly the sort of situation where the radar ought to have detected the woman and stopped.
I also have to ask if the vehicle was going 40 miles an hour, why was this case on a street that had a 35 mile an hour speed limit?
I also have to ask if the vehicle was going 40 miles an hour, why was this case on a street that had a 35 mile an hour speed limit?
I'm not up to speed (pun intended) on my autonomous vehicle programming just yet, but does anybody know how the cars are programmed to know what the speed limit is and what speed to go?
Do the vehicles use visual cameras (what about at night in dimly lit areas) to see the speed limit signs and adjust according?
Obviously driving over 10 or 20 miles in and around a city and on and off freeways, the speed limit could be changing a half a dozen times or more over that distance. I could see where they program the vehicle to go say a max 5 miles over a posted speed limit, but there's no radio signals beaming out to the car, telling it this is the posted speed limit for this section of road.
Do they already know ever single posted speed limit for every section of road in a given metro area and where they change and it's fed into the computer/gps/mapping system? What about if it's a construction zone and the speed limit all of sudden changes and that hasn't been inputted into the program yet?
I think there's a speed change along that stretch, from 35 to 45 as you travel north. To me, it's not so much the speed that's the issue but rather the complete absence of any action from the car.
I don't know for certain, but I expect the speed limits are part of the map database.
The computer can make the kid versus adults calculations easier than biological units can. And it doesn't tunnel vision towards an individual event while doing it. We just have to think of the possible events to prepare it. Be it a kid or moose expected behaviors versus a deer brought up in a sister thread. We are just in the beginning phases of finding those events and programming for the robotic takeover of the public roads
And all of that should have been done prior to allowing this to be foisted on the public! After looking at the video it appears that this was not some one in a gazillion chance scenario. It would seem that something logical slipped through the cracks or there was some system failure. Either way it's still too early to put this technology into the public domain.
As for the driver it was obvious she was in the driver's seat but paying attention much less than necessary compared to a driver in full control of a non-autonomous vehicle. At this stage of development/testing that lack of attention should not have happened.
Think of all the effort that goes into sending a crew into space. Apollo 11 is a perfect example. The crew had all kinds of back-up systems and back-up to the back-ups. The same effort needs to be applied to this technology as you are putting a hell of a lot more of humanity at risk. I mean I can't even get a hard drive to last more than a couple of years.
New item on the new car's sticker: MTBF. Hopefully no one gets killed before the 'F'!
If you watch the video the person was almost across the roadway so it wasn't "just stepped in front of" or some fast movement. It was a dark section of roadway that on the video looked hard to see in but video is not human eye quality. Didn't matter because the human driver was not paying attention like she was PAID to do.
I'd also think that the car should be able to detect way better than any human as we don't have radar...
Thanks, I hadn't seen the video when I wrote that last message. Yup, initial reaction is this doesn't look good for Uber.
I'm not up to speed (pun intended) on my autonomous vehicle programming just yet, but does anybody know how the cars are programmed to know what the speed limit is and what speed to go?
Do the vehicles use visual cameras (what about at night in dimly lit areas) to see the speed limit signs and adjust according?
Obviously driving over 10 or 20 miles in and around a city and on and off freeways, the speed limit could be changing a half a dozen times or more over that distance. I could see where they program the vehicle to go say a max 5 miles over a posted speed limit, but there's no radio signals beaming out to the car, telling it this is the posted speed limit for this section of road.
Do they already know ever single posted speed limit for every section of road in a given metro area and where they change and it's fed into the computer/gps/mapping system? What about if it's a construction zone and the speed limit all of sudden changes and that hasn't been inputted into the program yet?
Also would the car know the default speed limit in a state like New York where it's 55 on unmarked state and county roads?
Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.
Detailed information about all U.S. cities, counties, and zip codes on our site: City-data.com.
Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
