Waymo Robotaxi Hits a Child Near an Elementary School in Santa Monica (techcrunch.com) 167
A Waymo robotaxi struck a child near an elementary school in Santa Monica on January 23, according to the company. Waymo told the National Highway Traffic Safety Administration (NHTSA) that the child -- whose age and identity are not currently public -- sustained minor injuries. TechCrunch: The NHTSA has opened an investigation into the accident, and Waymo said in a blog post that it "will cooperate fully with them throughout the process."
Waymo said its robotaxi struck the child at 6 miles per hour, after braking "hard" from around 17 miles per hour. The young pedestrian "suddenly entered the roadway from behind a tall SUV, moving directly into our vehicle's path," the company said in its blog post. Waymo said its vehicle "immediately detected the individual as soon as they began to emerge from behind the stopped vehicle."
"Following contact, the pedestrian stood up immediately, walked to the sidewalk, and we called 911. The vehicle remained stopped, moved to the side of the road, and stayed there until law enforcement cleared the vehicle to leave the scene," Waymo wrote in the post.
Waymo said its robotaxi struck the child at 6 miles per hour, after braking "hard" from around 17 miles per hour. The young pedestrian "suddenly entered the roadway from behind a tall SUV, moving directly into our vehicle's path," the company said in its blog post. Waymo said its vehicle "immediately detected the individual as soon as they began to emerge from behind the stopped vehicle."
"Following contact, the pedestrian stood up immediately, walked to the sidewalk, and we called 911. The vehicle remained stopped, moved to the side of the road, and stayed there until law enforcement cleared the vehicle to leave the scene," Waymo wrote in the post.
The best outcome of a tough situation (Score:5, Insightful)
Sounds like the rules of the road were followed. This can happen and does happen daily.
Lucky kid!
Re:The best outcome of a tough situation (Score:4, Insightful)
This is a situation that many drivers would have encountered early in their early driving careers and get trained to anticipate. Someone emerging suddenly from behind a stationary vehicle, such as a school bus, especially in a residential area. This scenario and other similar scenarios are also part of the EURO NCAP scenarios that self-driving cars have to pass, and all self-driving algorithms are tested for before getting certified.
Re: (Score:2)
Someone emerging suddenly from behind a stationary vehicle, such as a school bus, especially in a residential area.
As far as I know, in all the countries I have driven, when a school bus stops and engages its warning lights, traffic must stop in both directions. There may be some exceptions for traffic in the opposite direction (e.g., when a divider is in the centre of the road.) This is a sensible law that accounts for young children being careless sometimes. We have all been there.
I have no doubt that Waymo and other self-driving cars can recognize a school bus and act appropriately within the law.
Re:The best outcome of a tough situation (Score:5, Insightful)
It happens frequently, yes.
But usually, the kid gets seriously injured or killed and the whole thing barely makes the local news, if at all. This is a success story and a nice demonstration what SAE 5 can actually do and do it reliably (quite unlike a human).
Re: (Score:2)
One story doesn't show reliability. It's just an indication. (A really GOOD indication, since the kid walked away, but just an indication.)
Re: (Score:3)
A "demonstration" is an indication, not proof.
Re:The best outcome of a tough situation (Score:4, Insightful)
From the summary, looks like the Waymo did well, although we'll have to learn more details to see if there are factors that aren't being told. (why did the student suddenly appear from behind a SUV? Was this in a crosswalk?) If the pedestrian just stood up and walked away after the 6-MPH accident, though, looks like it was not a serious problem.
I don't think a human driver could do better.
Re: (Score:2)
I'm pretty sure an inattentive kid routinely impacts stationary objects at around 6 MPH. It'll leave bruises and may hurt for a little while, but it's certainly not a dangerous speed to get hit at unless you're SUPER unlucky and end up under a wheel or something.
Re: (Score:2)
I'm pretty sure an inattentive kid routinely impacts stationary objects at around 6 MPH.
Which isn't the same as a (relatively) stationary kid being hit by a 4000 lb object moving at 6 MPH. The impact, the transfer of momentum, will (usually) be greater.
It'll leave bruises and may hurt for a little while, but it's certainly not a dangerous speed to get hit at unless you're SUPER unlucky and end up under a wheel or something.
But, yeah, agreed.
This story serves as a good advert for Waymo, imo.
Re: (Score:2)
It all depends on full circumstances. For example, if there were a lot of kids around, and many had previously done the exact same thing when the car was further back; whether you could see the kid approaching before they were behind the SUV that blocked them, whether there was a crosswalk, etc. There are many cases where a cautious human would not have hit the kid. On the other hand, maybe not. It all depends on what else was going on.
Re: The best outcome of a tough situation (Score:5, Insightful)
I think the difference is (Score:2)
I don't know how you would implement that in software. I'm sure it could be done but can it be done in a cost-effective manner.
They Don't really teach you what defensive driving really means.
It means you're assuming that the other guy is going to do something stupid and unpredictable and you need to be ready for it.
That goes doubly when the other guy is a kid.
Re: (Score:2)
Why is it impossible to follow? I do it all the time. If I'm passing a parked vehicle that I can't see around I drive as if something is going to jump out.
Re: The best outcome of a tough situation (Score:5, Interesting)
Death in the local area (Score:2)
Mom talks about how a few years ago a high school senior was killed in a car accident because he stepped out into the road unexpectedly. While wearing a hoodie with the hood up and headphones.
The odd thing that always got my attention - it was supposedly the very first time he'd ever walked to school.
The Waymo stopped a lot faster than most humans would have.
Re: (Score:2)
If you want to stop bad things happening: Outlaw cars and outlaw children.
Re: (Score:2)
Re:The best outcome of a tough situation (Score:5, Informative)
Every speed has a minimum stopping distance where, even assuming a reaction time of 0 (impossible), if something appears in that zone without warning, you will hit it no matter what. That's just physics. You can't bring a vehicle to a halt in an arbitrarily short distance.
Somebody stepping out from behind a tall obstacle with no warning right in front of you is one such case. All you can to is try to stop as quickly as possible and hit them with as little force as you can.
The only absolutely safe speed is 0, with the parking brake engaged, and chocks under the wheels.
Re: (Score:2)
Re: (Score:2)
Nothing in the article mentions the kid exiting a vehicle. Just that he or she came out from behind a SUV.
Re: (Score:2)
Re: (Score:2)
Yes it's wise not to drive faster than your visibility permits. If a brick wall suddenly enters the range of your headlights you should be able to stop in time.
Your rule of "never overdrive your ability to stop for something that may appear in front of you without warning" is much harder. Someone can hide (intentionally or not) in front of any parked car and jump out full speed without warning if they want to. To follow your rule, you'd have to drive 5 MPH or maybe even less, any time that there are cars
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Re:The best outcome of a tough situation (Score:4, Insightful)
Since you just made up a rule that is impossible to follow unless you always drive no faster than ten miles an hour, I can't say that I attach much importance to it.
Re: (Score:2)
10 mph is way too fast for that rule. The only speed that can satisfy that rule is 0.
Re: (Score:2)
Not true. There is definitely a nonzero time it will take the kid to emerge from behind the SUV to walk in front of your car. If you can't drive slow enough for that, then don't drive so close to parked cars near a school.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
If it's not safe to merge, they did the right thing. Or should they just force their way into moving traffic?
Re: (Score:2)
If it's not safe to merge, they did the right thing.
Both times I saw this it was in Bismarck, North Dakota. With the weather conditions and amount of traffic present you could have merged an aircraft carrier successfully. The only thing un-safe about the conditions for the merge in these instances was the drivers ability to execute it.
Or should they just force their way into moving traffic?
It's always the merging driver's responsibility to yield. With free-flowing freeway traffic there are fleetingly few scenarios where stopping at the end of an onramp would ever be appropriate. The lack of ability to successf
Re: (Score:2)
You don't have to stop. But maybe it's 5 miles an hour. Again, automated vehicles should be great at this.
Re: (Score:2)
What about the rule to never overdrive your ability to stop for something that may appear in front of you without warning?
What about the humans that basically follow this rule like "never"? The SAE-5 guidance system was careful and reacted fast. A human might not have been and would definitely have reacted slower.
Re:The best outcome of a tough situation (Score:4)
Have fun never driving above 1 mph in a city environment.
Re: (Score:2)
That applies for the things you can reasonably see...
If something jumps into your path of travel randomly from a total blind spot, there's generally not much people can do.
Basically you'd have to drive at no more than 3 miles an hour if there are woods near the road or parallel parked cars or anything that could conceivably obscure someone that could jump out at you.
Re: (Score:2)
You slightly overstate the case for "woods", as the trees are generally at a slight distance from the road. (Of course, deer jump considerably faster than a kid does. It depends on the speed of the obstacle as well as your stopping distance inclusive of reaction time.)
Re: (Score:2)
When passing an obstruction like a tall van, there is actually a point where the 'safe' speed forward is under 1 MPH when you include reaction time. Do you pass parked tall vans at less than 1MPH on a routine basis?
Followup, if yes, do other drivers frequently throw eggs, coffee, and trash at your car?
Re: (Score:2)
Meaning, the "driver" was not in control, considering its speed and the environment. It was reckless.
So what you're saying is that Waymo have finally achieved autonomous driving to human standards?
If your goal is to not have any contact then there is no safe speed at all to pass any obstacle.
Re: (Score:2)
So, in your opinion, 17mph is recklessly fast to drive while passing parked cars. What would you think it an appropriate speed? 10? 5? 1?
I would actually like transportation to work on reasonable time scales, so I don't think your viewpoint is the least bit rational or sane. It is no different than "you might trip if you walk, so walking is recklessly dangerous." At some point, you need to take into account factors other than maximum safety.
You do realize that parked cars are on most streets in every city?
Re: (Score:2)
It was reckless.
LOL! Ok. I'll play along: A couple years ago I was travelling through an intersection at the speed limit. Good visibility, clean and dry roads, no conditions that would warrant reduced speed. Someone in the opposing left-turn-lane thought he saw his light turn green so just blindly accelerated into the intersection. I had enough time to involuntarily pucker my asshole up, move my foot to the brake pedal, and steer to the left enough to hit his back wheel instead of his passenger door. T-boned him doing
Ideal (Score:4, Insightful)
Re:Ideal (Score:4, Informative)
The average human driver would have been texting while speeding and would have run the kid over.
Re: (Score:2)
Re: (Score:2)
Is that mean, median, or mode?
I accept what I believe is your point, but the expression is...less than optimal.
Re: (Score:3)
Could I please see a reference demonstrating that the word "average' equates to claiming it applies to EVERY SINGLE ONE?
Re: (Score:2)
Could I please see a reference demonstrating a strawman EVERY SINGLE ONE is strawmaning while typing and that EVERY SINGLE ONE would have written a strawman? Has no poster ever strawmaned with non-strawmaners?
Re: (Score:2)
Yep, according to the article:
Waymo said in its blog post that its “peer-reviewed model” shows a “fully attentive human driver in this same situation would have made contact with the pedestrian at approximately 14 mph.” The company did not release a specific analysis of this crash.
Re: (Score:2)
Indeed. This is a resounding success. The engineers will be very happy with the performance of this thing they built.
Had a human been driving, this would not have gone so well. It would likely barely have made the local news though. And do you know what is the biggest danger for kids near a school? Parents driving their kids. They are tired, stressed, distracted, etc. Good luck.
Re: (Score:2)
It's not necessarily the case that it can react to a human pedestrian faster than a human driver can. A human might be able to determine that the thing is indeed a pedestrian and not a puff of smoke or a plastic bag or a toy balloon faster than the car can. Once it has computed an adequate probability that the thing is a pedestrian, yes, then it's faster, but not necessarily before then.
Re: (Score:2)
Re: Ideal (Score:2)
Re: (Score:2)
Good for you!
Unfortunately, the world is not made up of just you, and plenty of assholes don't give a hoot about other people's kids. They're not even aware they're near a school because they're responding to a text message while driving.
Re: (Score:2)
That may be, but my comments weren't about evaluating people's driving skills. The point was that Waymos will be safer if they have enough intelligence to take the situation into account and drive more conservatively in higher risk situations. This was a response to RobinH's comment that
The car was probably following the posted speed limit precisely.
In my view following the letter of the law may be too fast.
Re: (Score:2)
Re: (Score:2)
In you attempt to state what a theoretical human driver might ideally do, you described exactly what the Waymo did. It slowed down to 17mph, under the speed limit, in the situation. 17 mph is not fast. 17 mph is right in the range I would deem appropriate. It's 8 mph lower that the School speed limit where I lived in all the cities I've lived in the US. The lowest school speed limit on a public road I've seen is 15mph.
Every Waymo does this, and far, far less than half of human drivers would, most would be s
Re: (Score:2)
I'm accustomed to 20 in school zones and I have seen 10. I guess in LA it varies between 15 and 25 so you have a fair point. I'd be interested in hearing if it was going 17 specifically because it was in a school zone during drop-off hours, and if they typically go under 20 in that situation.
I'd also be interested to know how the crossing guard was involved, and how Waymos work with crossing guards generally.
I have no doubt that this is all big news because it was a Waymo, and I'm sure many human drivers
First Hand Experience? (Score:2)
Would anyone here that has had first hand experience, riding in a Waymo, care to share what it's like? Speeds, distances, traffic, general feeling, comfort level..?
Re: (Score:3, Interesting)
Re: (Score:2)
I've only taken one ride, while on vacation in San Francisco
At many major conferences in San Francisco one of the (unofficial) things to do for first time visitors is ride in a Waymo (for the experience?). The number of Waymo's outside the Moscone Center (the major conference facility in San Francisco) can be amusing.
Re: First Hand Experience? (Score:4, Interesting)
I've ridden in one. It was just like riding in any car, but a little disconcerting to have the driver's seat empty while the car did its thing. Maybe a little smoother and more predictable than the average Uber.
Given how many Waymos there are on the road and how little trouble weâ(TM)ve heard about - despite breathless coverage of every incident, - Iâ(TM)d say they seem to be doing better than human drivers.
Re: (Score:2)
they seem to be doing better than human drivers.
This is why I never understood the doom and gloom about how self-driving cars could never work. "Better than human drivers" is an *extremely* low bar.
Re: First Hand Experience? (Score:2)
very conservative, defensive driving. surprisingly normal feeling
Re: (Score:2)
very conservative, defensive driving. surprisingly normal feeling
You mean, completely the opposite of a NYC Taxi driver? To be fair, I have had completely uneventful rides in NYC in taxis, but I really remember those that were a bit more, um, exciting.
Re: (Score:2)
I rode a few while on a trip to Phoenix and it was way better than the average Lyft or Uber ride I've taken.
It drove defensively, I got to choose my own music and climate control, I didn't have an driver tapping on their phone to plan out new rides, or high, or bragging about how they've been driving since 5am, or complaining about how they've been driving since 5am.
This is coming from someone who is typically pretty skeptical of big tech bullshit and a fan of labor. I still use the cashier when there is a
Re: (Score:2)
But the tech here is so good that its going to destroy all of those uber drivers (assuming it can sort of handle weather -- I can't speak to that)
Some Uber/Lyft drivers are trying to keep Waymo out of their locations precisely because they see what is coming (others just figure they will find some other gig job).
The prices for Waymo are still a bit higher than Uber/Lyft, so the price conscious may continue to use Uber/Lyft for some time, but Waymo can pretty much decide what to charge should they choose to undercut the alternatives (Uber is limited as to how little they can charge, since the drivers need some revenue).
Inclement weather is a chal
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
It was a nice ride but it wasn't allowed to go on the highway so the trip took longer.
Waymo's can now transport passengers using (at least some) highways (I think it is still the case that it is a slow roll out for now, and you need to request access to highway routes, but the approvals tend to come quickly).
Waymo has always tended to roll out new capabilities initially to a small group (in Phoenix, non-employees needed to get "early rider" approval), and expand as they gain additional experience. Waymo has been reportedly doing highway routes for their employees for a while now, and now
Wrong headline (Score:5, Insightful)
Shouldn't the headline actually read, "Waymo vehicle saves child from serious injury"? Because that's pretty much what happened.
A human driver probably would have hit that kid and knocked him 20 feet, or even run him over. Instead, the child gets up and walks away after doing something incredibly dangerous.
Re:Wrong headline (Score:4, Insightful)
Indeed. But the media is not after reporting accurately and sensibly. They want sensationalism.
Also, had that been a human driver, the dead or severely injured kid would barely have made the local news.
Beautiful Bot Bashing Brings Bountiful Buyers (Score:2)
But bots flubbing up is better click-bait. There's an anti AI movement going on. For example, people who post AI political memes tend to get voted down on social media sites, with replies saying "get that AI slop outta here!" even if well-rendered, yet they don't complain nearly as much about bad Photoshop or MS-Paint.
Re: (Score:2)
Yeah. That part of Asimov's predictions is developing.
And if that had been a human driving... (Score:2)
The child would probably severely injured or dead now. But that would not make the news.
Face it: Regarding safety, most humans cannot compete with self-driving.
Re: (Score:2)
A human should understand that in a school zone, especially at pick up and drop off times, you can't just pass a parked vehicle without paying attention to the possibility of someone small walking out into the road. This is why school busses have those little stop signs that swing out. This is exactly the kind of subtle detail that you can't make an A.I. system anticipate. 17MPH is too fast to be going in that situation.
Re: (Score:2)
Which is exactly what the Waymo did. it had slowed down to 17mph before detecting the child and hitting the brakes. 17mph is an entirely reasonable speed for the exact situation you describe. It's 8mph lower than the school speed limit. I would be shocked if that's not well under the average speed of human drivers in that same time and place.
At some point, you need to acknowledge that things that are out of sight are actually not visible to the car, regardless of who is in it. A very short period of visibil
Re: (Score:2)
I have no idea why people here think
That is just the thing. All the negative comments come from people that do not think. This is a nice demo how SAE-5 actually is a lot better than the average (or below-average) human drivers. And it is so consistently and over all instances, meaning there are no below-average vehicles and none that "have a bad day". Instead these are mindlessly claiming that this vehicle should have done better and insinuating this shows a defect in the tech. It does not. Everything worked as intended and it worked well.
Wha
Re: (Score:2)
Yes, "a human should". But quite often a human does not. And a human has far slower reflexes. An SAE 5 vehicle, on the other hand. will have consistent performance, very fast reflexes and even after such a non-event, will get updates making a serious accident even less likely. All not there for a human driver.
You are trying to argue something that has no substance. Does not make you look smart.
Re: (Score:2)
You're probably right, but I feel you are arguing from too small a sample. OTOH, the sample publicly available is highly biased. I think the only reasonable choice is to say "Provisionally current SAE5 vehicles are safe enough to allow full use...but with strong continued monitoring.". And also to recognize that not all SAE5 vehicles are equivalent.
Re: (Score:2)
Face it: Regarding safety, most humans cannot compete with self-driving.
Especially when 93% of drivers think they possess an above-average ability to drive. That's a lot of Dunning Kreuger piloting around multi-thousand-pound blocks of steel. https://www.thewisedrive.com/t... [thewisedrive.com]
Re: (Score:2)
I bet more than 93% of AI developers think they are above average coders. Do we get to put them in jail when their code kills someone? A human who severally injures or kills a child in that situation would be going to jail.
Re: (Score:2)
Do we get to put them in jail when their code kills someone?
Probably. Idk. That's for the lawyers to sort out, and just like addressing any other new technology, that will take time. I suspect the answer to that will follow existing frameworks for negligence and culpability.
A human who severally injures or kills a child in that situation would be going to jail.
Bullshit. Assuming you're sober, attentive, and driving reasonably and prudently for the conditions you're not likely to be seeing the inside of a jail cell.
Re: (Score:2)
Yep. Drivers routinely overestimate their skills massively and hence do unsafe things. It is so common that yet another bad driver killing or seriously maiming somebody does not make the news. It is just business as usual.
Re: (Score:2)
Negative spin. (Score:2)
Swerve (Score:2)
Also, not to disparage the AVs, I look off to the sides of the road too, to predict upcoming trouble (such as seeing the kid before he even walked between the SUVs). I almost crawl when driving in tight areas where children are around.
Awesome! (Score:2)
17 mph self driving car immediately detects kid and does everything in its power to stop and lessen damage.
Contrast that with a human driver, who rarely can go below 20, typically drive 30+, who wouldn't immediately detect the child, but once they did would have a 2+ second reaction time of panic. If they hit the brake instead of the accelerator (which many do in panic situations, swearing they fully put the brake pedal to the floor), they'd have hit the poor kid with much more force.
These are the types of
No data (Score:2)
We have no information about the accident except the press release from the legal department of a company.
Maybe the accident happened that way, maybe it didn't.
Re: (Score:3)
Re: Robots vs people, child hit vs not (Score:2)
Re: (Score:3)
Well, you got lucky. There was a ton of warning signs and you saw and understood them. Let me know when that is the average case and the average performance of a human driver.
Also, if that kid had been killed or severely injured by a human driver, it would barely have made the local news.
Re: Robots vs people, child hit vs not (Score:2)
Most human drivers perform below their own capabilities. Luckily we can hold software to a higher standard without much pushback or whining or execuse making. Well plenty of people outside the industry are willing to white knight the self driving industry by inventing arguments and excuses for what is often a solvable problem.
Re: (Score:2)
I've been in that situation except I didn't hit the kid.
You've been in a SIMILAR situation. Were all of the cues that you had in your situation available to the robot? How do you know it's not programmed to absorb the same data you did and take the same action?
Re: (Score:3, Insightful)
Re: (Score:2)
Re: (Score:2)
Most of the times I've had balls roll out, no kid followed. However, I've had kids run out in front of me many times. Usually even if they came from invisibility between parked cars, I observed their actions before then and was ready.
Re: (Score:2)
You seem to be seeing things that are not there. Probably because you are not looking at this rationally, you only look for factors that you can then hold against them. And you do not seem to mind stretching the truth to its breaking point.
That is the typical irrational behavior of somebody deep in fear. Get over it. Accept that self-driving is already much safer then than the average human driver. And remember that way below average human drivers are also allowed on the streets.
Re: (Score:2)
I'm not sure it it, in general, much safer. It certainly appears to be so in this case, though. (And remember that "self-driving" is not a unitary term. There are multiple implementations out there. So it's not just that this is a special case...which every single data point is.)
OTOH, most of the stories are more about inconveniencing people rather than injuring people. So maybe is is much safer. It's hard to tell, because it definitely makes mistakes that people wouldn't make, like stopping in the mi
Re: (Score:2)
I have never been convinced that safety is at all being taken into consideration in the design of these things.
That is because you have no clue how the research is done. Safety is the overriding factor in the design of self-driving tech (well, except at Tesla, but they are an exception and they have the problem that their big-ego owner mistakenly thinks he is an engineer), because the people working on it know that they will be held to much, much higher standards than human drivers. Oh, and look, that is happening.
In the case at hand, a human driver would likely have bad injured or killed that kid. That would not ha
Re: (Score:2)