https://www.theregister.com/2023/10/04/driverless_cruise_car_pedestrian/?td=rt-3a
Victim in critical condition
Should the Cruise car have not started moving if there was a person still on the crosswalk? This whole sad affair raises many questions.
There are some questions but "should cars start moving while a person is still on the crosswalk?" is surely not one of them.
A different question I have is whether or not the cars have transponders or other communication devices to automatically call emergency services in case of accidents. I'm assuming not because they would probably have a lot of junk calls and I doubt the company would have spent the time to create an algorithm for when to call 911 if they didn't create an algorithm for what to do if there's a pedestrian in a crosswalk.
That's one of the big downsides of these driverless cars: if a human accidentally ran over the victim, they have the capability to get out of the car to assess the situation, call 911, and offer aid to the victim. An empty car can only ever just sit there with its hazard lights on and maybe call for emergency services.
The auto-driving company should be required to have something like an on-star operator available any time the vehicle receives an impact/shock above a certain threshold and any time physical safety measures are required. The local governments should not have to pay for the externalities created by these 'disruptive technology' jerks, especially when there are literal lives on the line.
Cruise has this. I actually applied for the position after my contract doing the same thing with Waymo ended (but was unfortunately ghosted). They've got a team of people who monitor the fleets in real time, mostly just helping a "stuck" car by identifying any objects or street signs that the SDC has been confused by, so that it can proceed with its course. But they also have protocols in place for reporting any collisions as soon as they've happened, as well. Willing to bet that Cruise called emergency services before anybody on the scene even did.
A person laying on the ground in a crosswalk was likely never considered by the team to include in their training data. Those outlier situations are exactly what real world data is needed for. And the only way to properly train for most of these situations is to drive in the real world. The real world isn't perfect situations and nice lines on fresh asphalt so while base training in perfect situations is useful, it will still miss the exact same situation in a real world environment with crappy infrastructure.
Not sure what or how Cruise uses the data collected in real-time, but I can see camera visuals categorizing a person laying in the crosswalk as something like damage to painted lines, and small debris that can be ignored. Other sensors like radar and lidar might have categorized returns as something like echoes or false results that could be ignored, again because a person laying in the crosswalk is extremely unlikely. False data returns happen all the time with things like radar and lidar, millions of data points are ignored as outliers or info that can be safely ignored, and sometimes that categorization is incorrect.
Well if you know you need backup with edge cases, why isnt there a human in the car with controls?
Not only that, but no matter whether it can identify a person as a person, cars shouldn't be driving over objects that are child sized or larger.
These cars won't even drive over a fire hose laid out across the road (this was literally a test we did regularly when I worked on Waymo's SDCs). They definitely won't drive over a person (intentionally).
Inertia is a factor here. Cars can't stop on a dime. By the time the pedestrian was knocked in front of the SDC, there was likely not enough time to have possibly avoided hitting her, just due to physics.
A person laying on the ground in a crosswalk was likely never considered by the team to include in their training data
I didn't bother reading any further than this. The person was on the crosswalk when both cars started moving. Neither car should have been moving while anyone was still on the crosswalk.
That was the exact moment I called bullshit as well. You'd damn well better plan for people tripping and falling. It happens all the time, but generally is pretty minor if not exacerbated by being run over. This is like saying they didn't train it on people holding canes or in wheelchairs.
It's not about the ability to recognise someone lying in the road (although they obviously do need to be able to recognise something like that).
She was still walking, upright, on the crosswalk when both cars started moving. No car, driverless or otherwise, should be moving forward just because the lights changed.
Thats the whole point of their comment. The car did not recognize anyone was on the crosswalk because it was never trained to look for people laying in the crosswalk.
And that's fine. But if it's unable to recognize any object in the road, it's not fit for purpose. The fact that the object was a person just makes it so much worse.
Agreed. I'm not defending Cruise at all. They should have humans in the car if they are testing. Or at least a drone-style driver sitting in a room watching a camera feed. I wonder if the car thought there was just a speed bump ahead. Some speed bumps are striped similar to crosswalks. I can see situations where the autopilot can't determine if something is a speed bump or genuine obstruction (either false positive or negative).
I actually work at one of these AV companies. We definitely have training data on adults and children laying down. I'd be very very very surprised if Cruise doesn't due to all the people laying down on the sidewalks in SF. In addition, the clarity of the lidar/camera data on objects on the road is very clear. You can see the dips and potholes in the road as well as specifically see the raises of the painted lines. There's no way they weren't tracking the person.
I could see predictions on the pedestrian saying the coast is clear. Once the initial crash happens, there likely isn't enough room to stop in time even with a max break.
Without knowing what type of vehicle the first car was, it's hard to say how this played out. If it was a van or truck or something else that could've easily obstructed Cruise's LIDAR system or if the other vehicle stopped ahead of the crosswalk line, the SDC would've had little to no way of knowing that there was anybody in the crosswalk.
Can't know for sure unless Cruise releases the video to the public, which they're unlikely to do until the police do their investigation.
If your vision of the crosswalk is obstructed, you don't proceed through until it's unobstructed. That's true whether it's radar, lidar, or vision. Truck in the way? Pull up as far as you can safely see, then look and proceed if clear.
Oh, this company is not doing itself any favors.
There needs to be full transparency on these fleets rather than having governments bend over backwards in the name of trade secrets. We’ve gone absolutely too far in that direction with everything from vehicles on our streets to fracking chemicals in our groundwater.
We desperately need footage of this to make any conclusions.
The human driver hit her first, and knocked her into the neighboring lane...
So she was hit and flung into another lane...
...directly in front of a Cruise autonomous vehicle (AV) that was driving around by itself with no-one on board. The self-driving car then ran her over and came to a stop on top of her body, turning on its hazard lights. Her leg was pinned down by the back tire.
So it stopped Like others have mentioned, driving over someone under a car can cause more injury than not moving. Was she screaming to move forward/move back? Was she flung in a way that a human driver could have stopped in time?
If this was just a hit and run and there was no footage like what was provided by the ai car it could be that the victim or her family would be 100% on the hook for the medical bill. It will be interesting to see if the perpetrator is found and the footage surfaces with details so we can get some answers.
edit: From another article
The initial impact was severe and launched the pedestrian directly in front of the AV.
I'm hoping that the car has multiple cameras recording so maybe the hit-and-run driver will be caught and prosecuted. That said, unless I misunderstood the article, I think the driverless car (DC) didn't do a horrible job here. It sounded like the victim was struck and flung in front of the DC and it stopped (unfortunately on top of her). I don't know if I could have reacted better. The article wasn't clear but it read like the car contacted the police and the police instructed it to remain where it was, which is what I would have done if I were driving a car. The DC was then lifted off the woman by emergency personnel. We can't expect DCs to be magically perfect. Just like we don't expect people to be perfect. A DC is only as good as it's programming. Hopefully this incident will be studied and if a better solution is found that can be integrated into the DCs operations. I really feel bad for the woman here. I don't know but even if she shouldn't have been walking no one deserves that. Let's hope the hit-and-run driver is caught.
"I didn't cause those injuries, it was the driverless cars" might actually work here...
maybe. eh. the car company probably has better lawyers.
Yeah but the focus will be on the driverless car, I’m not saying they are blameless, just that the general heat and attention will probably not be on them now
Im not sure that is a problem.
Autonomous vehicles are still mostly half baked, and the question of liability of who gets the blame hasn’t even preheated the oven.
The reality is company’s like waymo are using their cars in SF precisely to harvest training data because they can’t finish it without real world data- the physical driving a car is easy; interacting with humans is not.
One of my son’s coworkers was just killed in a similar incident. Woman hits a pedestrian, she freaks out and calls her boyfriend instead of emergency services, boyfriend arrives and runs over the injured pedestrian ensuring he was dead.
They are unsure which vehicle actually killed him.
What have the general public got to do with it? This is a criminal offence, not a viral bit of gossip.
The general public is irrelevant here. If it weren't for the self driving car the hit and run wouldn't even have made the news. If it all this makes it worse for the driver.
And I don't think there is a wrong party to blame. Both are equally fucked up.
If it weren't for the self driving car the hit and run wouldn't even have made the news.
That is exactly the kind of thing I am talking about.
Perhaps he's implying that due to the second car it may obfuscate if the first driver's actions would have been lethal or just left the person injured. He'd probably rather be tried for a hit and run resulting in injury than a hit and run resulting in a death.
”The driver of the other vehicle fled the scene, and at the request of the police, the AV was kept in place.”
Hopefully not while still on top of her!
If you get stuck under a car the fire dept is going to come lift it off of you. They aren't going to try and drive it off that would almost surely cause further injuries.
Oof for that edge-case, in every sense. I hope the victim recovers with no long-term consequences. Truly horrific.
Reminds me of someone I knew that would keep a pint of Jack in his trunk. He drove drunk constantly and it was there for when if he ever got into an accident he was prepared to run out of the car, pop the trunk, and pound the bottle in front of all the witnesses.
Can't prove he was drunk at the time of the accident.
With the right legal team you can kick the police in the balls and piss on their car and get away with it.
I know I’ve read of at least one successful case where the person fled the scene and went home, then claimed he was drinking at home. Honestly, though, there’s so many things that factor into whether an individual gets arrested or released that we’d need more examples to differentiate between just letting someone go and This One Simple Trick Judges Hate.
With a half competent lawyer it could.
It's up to the police to show you were driving while intoxicated. You have witnesses corroborating that you were drinking after the accident. Any field sobriety, or blood test they give you would be worthless because it would be after that.
I'm sure someone has tried this before somewhere.
"I wasn't drinking and driving, officer. And I'll prove it by drinking out of this open container!".
This kind of thing has been repeated amd handed down for like a century. But I've never ever heard of anyone actually doing it, much less having it work.
That sounds like a terrible legal defense. Yes - I had alcohol in my car, and I was pounding it at the time of the accident. But trust me, I was totally sober when I actually hit that person.
It's up to the police to prove you were drunk while driving. Normally that's not hard since they can show that between them getting to you, field so riery testing, and taking you in there's no way for you to have had a drink before they take a blood test for instance. But if you break that chain, there isn't a good way to prove that it wasn't from after the incident.
Ah but you didn't just have alcohol in your car. That's totally legal, otherwise you would never be able to drive home from a store with alcohol. You even have witnesses stating you got it from your trunk, so even if it was already open, it could not have been within reach while driving. Which is a component of most open container laws.
It’s up to the police to prove you were drunk while driving.
Actually, its up to the prosecution to prove you were drunk while driving. And that standard is 'beyond a reasonable doubt', which I'm pretty sure 'pounding liquor after an accident to have plausible deniability on your insobriety' would make an easy argument to meet that threshold.
The cops will take you either way and let a judge decide what to do with you.
You might not be able to prove it, but anyone willing to chug alcohol in front of witnesses to have that kind of plausible deniability can easily be assumed to have already been drunk to start with. That just doesn't seem like it would hold up to the 'reasonable doubt' standard...
It was just protecting her like a dog would.
Whose a good car!? That's right, you're a good car!