Self-Driving Shuttle Involved In Crash Two Hours After Debut (www.cbc.ca)
New submitter Northern Pike writes: Las Vegas roll out of new driver-less shuttle spoiled by human error. It sounds like the shuttle did what it was designed to do but the human semi driver wasn't as careful. "The shuttle did what it was supposed to do, in that it's (sic) sensors registered the truck and the shuttle stopped to avoid the accident," the city said in a statement. "Unfortunately the delivery truck did not stop and grazed the front fender of the shuttle. Had the truck had the same sensing equipment that the shuttle has the accident would have been avoided." The self-driving shuttle can transport up to 12 people and has a attendant and computer monitor, but no steering wheel and no brake pedals. It relies heavily on GPS, electronic curb sensors and other technology to make its way.
I wonder if the shuttle doing the right thing was what the human driver expected.... maybe their algorithms are incompatible.
5 out of 6 people enjoy Russian Roulette & 6 out of 7 Dwarfs are not Happy
Had the truck had the same sensing equipment that the shuttle has the accident would have been avoided.
If the shuttle had the same sensing equipment as the truck has the accident would have been avoided (ftfy). A human would have laid into the horn as the truck got closer to alert him hes about to hit someone. A human would also have seen the truck backing in and yielded a larger room for error. An alert human may also see the situation that they could quickly back up a bit before the truck hit them. (per article trucker was cited for illegal backing (up?). This isn't ready in my opinion, but a nice alpha test though.
The victim self-driving shuttle bus didn't try to back away from being run over. According to reports, it couldn't for unspecified reasons. (I speculate that the autonomous logic or arrangement of sensors didn't adequately cover "going into reverse.")
Someone up-topic asked about sounding a horn. I haven't heard any press reporting that the autonomous vehicle tried.
Either case (if true) represent a difference between how the self-driving logic reacted and how a human driver would probably have. This tells me unless an autono-car can do everything a human driver can, at least as well as a human driver (admittedly a low bar), it shouldn't be on the streets. There will always be corner conditions; they have to be handled as well by the robot as they would be by a human.
Welcome to the Panopticon. Used to be a prison, now it's your home.
1) Robotic vehicles need a horn - and additional logic to handle when to sound it.
2) Robotic vehicles would benefit from the addition of a mechanical arm with a mechanical middle finger - for these sorts of post-accident situations.
#DeleteChrome
How many comments have I seen on Slashdot asking if all edge cases have *really* been tested? Well it turns out everyone was right in this case. I mean, was this AI tested on real streets at all? It's hard to imagine a car on the road for more than a month wouldn't have had a truck pull out in front of it unexpectedly a couple times. It doesn't matter how fast the AI brain is, this is a case where anticipation may have helped. I just feel bad for the truck driver. Yes, he was in the wrong but a lot of times driving a big truck you have to maneuver this way and rely on other cars working with you a little bit.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
The truck driver was in the wrong, yes, but self driving cars can't just check out if in an unexpected situation. They still have to try to avoid an accident like a human would. You want to live in a world where self driving vehicles run people over because they were crossing the street illegally?
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
The point is that they have to include in their parameters the possibility that humans may break laws. If this is so confusing for them that hey 'just stop' they will very definitely do more harm than good.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
Had the truck had the same sensing equipment that the shuttle has the accident would have been avoided."
If the shuttle had a human driver the entire incident would never have happened because the half-assed excuse for 'AI' they keep trotting out can't actually THINK.
A large part of the fact that I've managed to avoid accidents for so long is the fact that, as a human, I understand how other humans are likely to act and react.
The problem with AI drivers is that humans only loosely follow the rules of the road; their actions are driven multiple influences, and understanding what another human is likely to do in any given situation requires being a human being. For example, consider the following:
I'm sure there are dozens of other similar cases, but you get the point. AI might understand, in the nominal sense, how to drive a car. What it can't understand is what other drivers are likely to do.
The society for a thought-free internet welcomes you.