Given Cruise has managed to embarrass CPUC, et. al. government agencies with their maneuvers here, it feels like things could get really ugly for them - uglier than being forced off the road. Embarrassed government agencies often have a nasty habit of overcompensating after the fact, coming down like a bag of sledghammers on the source of the embarrassment. This may soon get even more interesting.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Cruise
- Thread starter hooty
- Start date
Cruise attempts to blame an individual car called Panini. I know that really they are giving out details, but it almost seems like they are trying to make it sound that the fleet software isn't the problem, just that car. Bad Panini!
The incident occurred on October 2, 2023 at 9:29 PM PDT in San Francisco. A Cruise AV named Panini, operating in driverless autonomous mode... The Nissan Sentra then tragically struck and propelled the pedestrian... The AV ... still made contact with the pedestrian. The AV detected a collision, bringing the vehicle to a stop; then attempted to pull over to avoid causing further road safety issues, pulling the individual forward approximately 20 feet
Also Cruise is kind of attempting to blame "California regulations" for their car dragging the pedestrian:
After a collision, Cruise AVs are designed to perform a maneuver to minimize the safety risks to the extent possible within the driving context. This is called achieving a minimal risk condition, and it’s required under California regulations and encouraged under Federal AV guidance. The specific maneuver, such as coming to an immediate stop, pulling over out of lane of travel, or pulling out of traffic after exiting an intersection, is highly dependent on the driving context as well as the Cruise AV’s driving capabilities in the moment.
stopcrazypp
Well-Known Member
Yeah, Cruise pretty much proved CPUC wasn't doing their job and the people and city governments that fought and said the recent approval for expansion was inappropriate were correct. Given CPUC's revoking of the permit, it also shows they could have taken action in the earlier approval, but just didn't.
That's a misleading characterization given stopping in lane is also an acceptable minimal risk condition, something that Cruise commonly does, as others pointed out (but the one time it would be appropriate, it doesn't).
OxBrew
Active Member
Not going to happen though, given our culture.
And each car is going to cost a fortune.
This idea assumes at its core that we replace private car ownership with at best subscription or co-op car access, or municipal fleets of public access cars.
I mean, I'm down with that, but I don't think Joe Blow F150 owners will get anywhere near that "socialist ideal."
diplomat33
Average guy who loves autonomous vehicles
Cruise decides to pause driverless across their entire fleet as they assess next steps.
Sounds like a management change is coming to Cruise...
Cruise pauses all driverless operations after collisions with pedestrians, permit suspensions in California
The orders of suspension followed a barrage of incidents since Cruise received approval to conduct around-the-clock robotaxi service in San Francisco.
www.cnbc.com
"Cruise will keep running its autonomous vehicles with human safety drivers behind the wheel, supervising the drives, the company also said on Thursday."
Cruise learning same lesson Uber learned the hard way. You can't game the system, either your product works or it doesn't. There can be no in-between at L4.
diplomat33
Average guy who loves autonomous vehicles
Very possibly and it might be needed to help rebuild public trust. In fact, I would not be surprised if Kyle gets fired over this.
Moreover, could GM shut down Cruise altogether? I think it is possible. Ford shut down Argo because Argo was not producing results and they did not want to continue losing billions each year. GM's financials are terrible. I think Barra recently said that GM is pausing their EV program. Now, with Cruise losing their permits, I could see GM deciding that they can't afford to keep Cruise anymore.
stopcrazypp
Well-Known Member
Or they may absorb it into the main company to help with L2/L3 products. That was Barra's original goal, but Cruise leadership didn't want to go that direction, so currently the collaboration is much more limited and Cruise is still operating fairly independently.
diplomat33
Average guy who loves autonomous vehicles
Just a couple months ago, I would have said it was unlikely but now with everything that has happened with Cruise, I think it is a definite possibility. It would allow GM to cut their losses and leverage Cruise's tech into Ultra Cruise. So I can see a lot of advantages. I definitely think Kyle's days as CEO of Cruise are numbered.
More likely, they would try to find a buyer rather than shut it down. A functional L4 system, even with flaws, has a lot of valuable IP.
What is the IP and engineering R&D relationship between Cruise Robotaxi and GM super/Ultra Ctuise driver assustance features? Are they largely separate efforts, such that GM can afford to lose the entire CA subsidiary without negative impact to their consumer vehicle product lines?
Since most of these Robotaxi rides seem to be 1 person rides I would think that you would need just as many Robotaxis as private cars. These companies want to keep wait times down. Keeping wait times down means more cars. Most households have 2 cars. If you and your Spouse/partner work and have to go in different directions to work you would need 2 Robotaxis.
Tam
Well-Known Member
It sounds like Cruise also performed brake checks at least in 2 rear-end accidents:
www.techtimes.com
Two Self-Driving GM Cruise Crashes, US Auto Safety Agency Investigates
GM faces investigations after its Cruise caused two new coalitions.
www.techtimes.com
1) Cruise lied and misrepresented the actions of their car; after the collision it dragged the pedestrian. So any ban is entirely their fault due to this step alone
However, going forward.
2) We haven't seen the video of the accident, but it is possible that the car wasn't able to tell that the pedestrian was under the car. Now you might suppose that it should be able to account for the missing 'object' that it hit and deduce that "I no longer detect the object to the sides, in front, or behind :: the object must be underneath me". That would require it to be able to do that kind of reasoning, and also in theory be able to do deductive math in other types of accidents "there were four objects all around me and now there are three so one of them is probably underneath me".
Is this kind of deductive reasoning possible? Or, just to stop & never move after 'hitting something'.
or
3) There must be extended sensor coverage underneath the car too. Some kind of 360° radar, say a little dome mounted in the middle that can detect the presence of an object underneath the car. Also, some kind of audio analysis of screaming, bystander yells, physical thumping from underneath or on body panels might lead additional clues. This kind of problem was inevitable, what is the AV industry's solution?
4) Slow creeping of a vehicle, or start/stop behaviour is an ominous thing. Is the AV about to take off? We don't know. It's clearly confused about something, but what is it going to do? I would like to see an externally obvious decision to be communicated. However it's done. ie: "I am about to drive forward over this fire hose", "I have decided it is safe to pull over", "Does anyone have a problem with that?". "Yell out, or tap once on my side to tell me not to do that". IDK. But these driverless AV's need to be capable of communicating to 'us' and also capable of responding to 'our' communication to them. In real time, not "Please call the number on the side of the vehicle (if there even is one), tell them the location, name & plate number, wait for an operator to help you". Real time. ie: THE CAR, itself.
However, going forward.
2) We haven't seen the video of the accident, but it is possible that the car wasn't able to tell that the pedestrian was under the car. Now you might suppose that it should be able to account for the missing 'object' that it hit and deduce that "I no longer detect the object to the sides, in front, or behind :: the object must be underneath me". That would require it to be able to do that kind of reasoning, and also in theory be able to do deductive math in other types of accidents "there were four objects all around me and now there are three so one of them is probably underneath me".
Is this kind of deductive reasoning possible? Or, just to stop & never move after 'hitting something'.
or
3) There must be extended sensor coverage underneath the car too. Some kind of 360° radar, say a little dome mounted in the middle that can detect the presence of an object underneath the car. Also, some kind of audio analysis of screaming, bystander yells, physical thumping from underneath or on body panels might lead additional clues. This kind of problem was inevitable, what is the AV industry's solution?
4) Slow creeping of a vehicle, or start/stop behaviour is an ominous thing. Is the AV about to take off? We don't know. It's clearly confused about something, but what is it going to do? I would like to see an externally obvious decision to be communicated. However it's done. ie: "I am about to drive forward over this fire hose", "I have decided it is safe to pull over", "Does anyone have a problem with that?". "Yell out, or tap once on my side to tell me not to do that". IDK. But these driverless AV's need to be capable of communicating to 'us' and also capable of responding to 'our' communication to them. In real time, not "Please call the number on the side of the vehicle (if there even is one), tell them the location, name & plate number, wait for an operator to help you". Real time. ie: THE CAR, itself.
EVNow
Well-Known Member
If the car needs to do this and bystanders need to help, its not really L3/L4. Its just another greedy company taking free help from others to make money.
Similar threads
