WTB custom AI programmed to be rude to vehicles displaying the wrong allegiance.
News Article: Self-Driving Uber Car Kills Pedestrian in Arizona
- Thread starter Bamaro
- Start date
Sure, I think that's absolutely right. And like I said, I think it also introduces a host of other problems elsewhere.
The real question here (which there is no way to really answer) is what are the odds that this person would have been hit if a normal driver was at the wheel and when you consider how many autonomous vehicles are on the road compared to how many normal ones, this doesn't look good. Thats a very high percentage of fatalities.
The question is best answered by statistics. How many accidents per mile are there for autonomous vehicles vs human-operated ones?
I think the link that Audub supplied a page back or so illustrates that it could NOT have been avoided if a human was at the wheel.
don't know that I am quick enough to even matter and I will likely be doing 80 in a 70 anyway
Police have released video of the crash. I don't want to link it but it should be easy to Google it. Pedestrian is in a dark section of street, wearing BLACK, walking a bicycle across a street with traffic coming and is not even looking toward the traffic. Maybe the car should have detected and reacted better, but this pedestrian is the biggest kind of doofus.
Video on the BBC site, at 18 sec the pedestrian is not visible, at 21 secs she's been hit.
Human driver would have hit her just the same.
Guys, you have to remember that video cameras do not see light as good as our eyes. Take a video with you phone in low light and I guarantee you that your eyes can see much more in low light then your camera. That video is not representative of what the drive could or could not see.
Plus, these self driving cars have LIDAR and other radar sensors that do not depend on visible light to detect objects. This radar system should have detect the object moving across the street with time to react. Seems like the software failed to react since the car didn't attempt to brake. I would love to see the data the system collected to see what the sensors and LIDAR did or did not detect.
Also, this video shows me a driver who appears to be looking down at a phone instead of monitoring this experimental system. The driver is literally looking down a second before the crash. That's negligence in my book. Would the drive have time to react if they were actually paying attention? Well, we will never know because, they cared more about staring at their phone than paying attention. Another case of a driver using a cell phone and killing someone IMO.
Plus, these self driving cars have LIDAR and other radar sensors that do not depend on visible light to detect objects. This radar system should have detect the object moving across the street with time to react. Seems like the software failed to react since the car didn't attempt to brake. I would love to see the data the system collected to see what the sensors and LIDAR did or did not detect.
Also, this video shows me a driver who appears to be looking down at a phone instead of monitoring this experimental system. The driver is literally looking down a second before the crash. That's negligence in my book. Would the drive have time to react if they were actually paying attention? Well, we will never know because, they cared more about staring at their phone than paying attention. Another case of a driver using a cell phone and killing someone IMO.
Last edited:
That's a leap, you have no idea what she's looking down at, it could just as easily be a part of the systems she's supposed to be monitoring. The victim was oblivious to the vehicle, so even if the driver was distracted by a phone or something else she shouldn't have been looking at the victim still has at the very least an equal share of the responsibility.
After watching the video a couple of times IMO the human eye would pick up on the fact that there is sometime there before the jaywalker came into view of the headlights. The jaywalker momentary blocks a light source in the background which may clue the average diver driving a conventional car that there is an object moving across the road. Our eyes are good at detecting motion at night. The video camera in the car was set to provide the best view for 100 feet or so ahead of the car. I doubt the AI can come up with a 3D model of the world like we are able to build our in our mind based on a momentary blockage of distance light and other clues. That must have been some kind of new stealth bike. You would think the car's radar would have picked it up.
As far as the human driver behind the wheel, it is very difficult to stay alert and focused when you are just monitoring a machine that almost never messes up. It's like a pilot trying to stay focused on a long cross country flight when the autopilot does all the flying and navigating.
As far as the human driver behind the wheel, it is very difficult to stay alert and focused when you are just monitoring a machine that almost never messes up. It's like a pilot trying to stay focused on a long cross country flight when the autopilot does all the flying and navigating.
Latest threads
-
Final Home Midweek for Alabama Brings Samford to The Joe- Started by Diamond Tide
- Replies: 0
-
-
-
-
-
I'd like to hear about some actual prayer requests that have been answered the way it made sense to answer them.- Started by Labrador
- Replies: 15