From the article:
- Running through yellow emergency tape and ignoring warning signs to enter a street strewn with storm-damaged electrical wires, then driving past emergency vehicles with some of those wires snarled around rooftop lidar sensors.
- Twice blocking firehouse driveways, requiring another firehouse to dispatch an ambulance to a medical emergency.
- Sitting motionless on a one-way street and forcing a firetruck to back up and take another route to a blazing building.
- Pulling up behind a firetruck that was flashing its emergency lights and parking there, interfering with firefighters unloading ladders.
- Entering an active fire scene, then parking with one of its tires on top of a fire hose.
* Robotaxi violates traffic laws like blocking a fire engine? Fine of $10,000
* Robotaxi drive into an active crime/fire scene? $100,000 fine
* Robotaxi actively hinder emergency services (ex parking on a hose)? Confiscate the vehicle.
Corporations are the same anywhere, until a problem becomes too expensive, they don't do anything about it. So sure, let them drive anytime anywhere, but if they make egregious mistakes the cost is not "oh darn, an minimum wage 'automation engagement specialist' will have to drive out to the car and move it", it will be a "We were fined $1 million last night due to interactions with emergency services".
This is certainly in San Francisco’s tradition. Given the economic depression those policy’s are putting it in, one might wonder if chasing away this industry, too, is smart in the long run.
Yikes.
If a robotaxi gets in the way of your fire appliance, chop it up.
If only there was, perhaps, a similar effective, and cathartic response to these things.
I don't know if massive fines are the way to go, because they could cripple that technology, but firefighters should have free reign to shove those vehicles around when they misbehave, or some way to signal them to get the hell out of the way.
There was a case in the news where someone was exceeding the posted speed-limit by 150kph on an undivided highway (i.e. just a stripe between you and the traffic going the opposite direction) and got a 6 month suspension.
I was on the jury for a case where the defendant was involved his third DUI, fled the scene of the accident, and still had his license.
Uber is clearly a different problem, but a much bigger one in the macro.
There are a quarter million vehicle crashes in the DataSF Fire Department Calls For Service. Self-driving cars will prevent these. It's the systematically better way to go.
That's a bold claim. Do you have evidence to support it? I'm not talking about results from controlled or ideal driving conditions, I mean evidence showing that self driving will definitely be better at the conditions those quarter million accidents occur in.
Perhaps there's some research I'm unaware of but as far as I know the best we can say is that self driving might end up being safer for general usecase driving, but that there are wide disagreements about how likely that chance is.
It’s not even close.
Still wouldn’t get into a Waymo for another 3-5 years, but the quality gap is huge.
It should be completely banned until such time as there exists a comprehensive testing regime for validating that the automated system functions in all the scenarios it might be expected to face including emergency vehicles, construction, pedestrians, bad weather, and damaged sensors.
There is sufficient evidence that these are not ready for prime time.
The fire chief decided to have fire drills. Whenever a cop saw a car with diplomatic plates parked in front of a hydrant, they would call the fire department, who would come out and perform a drill. Upon seeing the vehicle blocking the hydrant, they would practice breaching through the vehicle to attain access to the hydrant. The vehicle would not survive in a drivable state.
The Soviets complained, but stopped blocking fire accesses.
EDIT: the firechief could solve this by putting truck or police cruiser style bullbars on their vehicles and "carefully nudging" the miscreant vehicles out the the way.
This makes me feel a human operator probably would be no good at taking control ... however a computer that has access to more data (eg data from other vehicles), or that has more control (of other vehicles, of traffic lights, of pedestrian crossings, etc.) or more capabilities, might be far more successful than a person?!
He drove off, and someone else got the spot.
In practice humans frequently neglect to signal in instances where it is legally required (I don’t know if it is strictly legally required in California when parallel parking, but it’s certainly good common sense). Human drivers know that people don’t signal and account for that.
In this particular scenario the reverse lights would be a crystal clear indication of the driver’s intent, even if they failed to signal.
I don’t want to assume the intent of your comment, I don’t know what you’re getting at..
I suppose the drawback to this strategy is that real harm has to happen first and that could easily involve loss of life or limb, but perhaps the threat of that would be enough to motivate the robo-taxi providers to fix the problem.
There haven’t been many cases of actual harm. This article pretty much lists out situations which could cause harm but didn’t, due to factors out of the operator’s or fire department’s control.
It's difficult to see domains where AI can really improve productivity without having major drawbacks.
I'm not against research on AI, but as long as science cannot define what intelligence really is, I guess AI will not make major advances.
[1] https://arstechnica.com/cars/2022/02/gm-seeks-us-approval-to...
> The fire chief said each robotaxi company offers training to help deal with “bricked” vehicles. > “We have 160,000 calls a year. We don’t have the time to personally take care of a car that’s in the way when we’re on the way to an emergency,” she said.
I understand that lag can be an issue, but if speeds are limited to 5-10 mph, that should be less of a problem.
And then do real world staged validation.
Actually why are extensive real world mock scenarios not running 24/7? If a car does something bad, do mock scenarios around it.
Please tell me they are doing these things because it’s crazy if they aren’t.
now, like anything else, does test/staging translate to production? Not even remotely. Automomous vehicle manufacturers claim to be in a monitoring/production phase to compare to their internal testing/staging. And over time we'll normalize Autonomous driving and the associated risks.
We used to drive all the nails in the world with a hammer precariously aiming at your own hand.
eg1: What about a direct line to a 24/7 operations center with cruise?
eg2: Or how about a proximal access console controller that allows them to take control of the the vehicle (with occupants consent if occupied) .
This neednt be a total ban on AVs
It would be useful if vehicles with sirens/emergency lights/hazard lights could broadcast position and identity in a similar way to ADS-B. Make data open and allow third parties to forward, store and re-broadcast that data. Make the consumer of the data (such as a self driving car) responsible for how it uses that data.
Some of these are likely just a lack of training data.
However, I agree that there needs to be a better self awareness of "Cruise has insufficient training data about this exact situation, now we fall back into a safer discovery mode, where we make reasonable guesses and also ask for the help of an operator"
Second of all, if that is not fast enough, then humans are insufficient too. You cant just magically get people to act (a lot of the time they'll do the wrong thing and fail to even recognize what situation they're in at the moment) Notice how bumper to bumper traffic is pretty bad for EMS, people do not know how to react.
This will have to be added somehow.
inclusion of people with disabilities
improved transportation options for the disadvantaged
reduction of greenhouse gases; and
passenger safety
But self-driving cars get away with it because we haven't figured out how to hold them responsible yet.
What are you supposed to do with a car with no humans around causing a problem for people? Key it? Break the windshield? Wait around for a tow truck to move it? Revoke an entire fleet’s license to drive in the State of California over a specific incident in San Francisco? Or wait until the fleet has tallied up enough incidents for the State to say “that’s enough”?
We probably do have to validate driverless vehicles to a higher standard than the average driver, and maybe a higher standard than the best drivers. We want correctness and accountability, because driverless or manned, an automobile mishandled is a death trap.
EDIT: How about this? The State can pass a law that gives immunity to anyone who breaks into a misbehaving unmanned autonomous car in order to disable the autonomy and take control of it and move it out of the way. No, I wouldn’t limit this to emergency personnel either. Anyone with a Driver’s License can do it. Cruise and Waymo can then decide whether to make this process easier or they can replace a lot of broken windows.
if we were still using Horse drawn carriages for travel, and someone invented the first Horseless Buggies today it would be banned and never allowed to advance at all. The amount of death and injury from the inception of the automobile would never be allowed in a new industry today. We have rationalized and assimilated the everyday human automobile into our lives, but refuse to accept any risk for something new
As you lay there bleeding out would you think "I have contributed to human progress - this is good"
I believe there are ways to advance self-driving without these problems. Why can't they have a 1800 number staffed 24/7 for fire departments to contact?
I know this was tongue in cheek, but it underscores a problem I have with this article title.
The problems, despite the article's title, usually don't involve robotaxis vs trucks, but robotaxis vs active firefighting scenes. The article I read earlier this year referenced cars running over hoses and inching towards firefighters after being blocked by an active firefighting scene.
https://www.forbes.com/sites/bradtempleton/2023/01/31/san-fr...
See: https://firematic.com/trucksnew/levittown19/2.jpg - 1 3/4" attack line on the right with plumbing, and 5" LDH (large diameter hose) supply line on the left.
https://www.firehouse.com/operations-training/hoselines-wate...
That has some "how much pressure can be loss to friction and still provide the necessary water rate at the other end." Bends in the hose will increase the amount of friction and thus reduce the flow rate.
Beyond instructing emergency responders "if it doesn't stay away, just smash its windows" (mentioned in the article, at n=1 scale)...I wonder if robotaxis have some convenient central depot, which might find its driveways block by emergency street repairs, or some such.
Montreal has no qualms going for it: https://www.youtube.com/watch?v=jCT74Szg_sA
$10,000 seems fair. A blocked fire engine could only lead to delay firefighters to get where they’re going. What’s the worst that could happen? A few building burn down? Somebody has to wait to receive medical intervention?
In the scale of things that’s a small price to pay to uh, eventually avoid small talk with Uber drivers. /s
You'd have to add at least 2-3 zeros before any action would be taken, and the action would likely be something else then what you are expecting. These things just don't happen often enough for this to ever be addressed without the original point of piercing the corporate veil and making management personally responsible for these crimes.
But at that point you'd likely still end up with the dynamic of people getting paid to become the fall guy
Assuming they are all the lowest "class" of incident with a $10k file each, that's $390,000 . A not insubstantial fine but maybe not huge impact. Though we know several of these incidents (because they were in the news) were of a more severe class. So lets assume that's half a million in fines at least.
So a million dollars a year in fines just to operate at limited times and in limited locations. Open that up to a lot more cars in a lot more places? Those fines will add up to be quite substantial.
What do you do with vehicles without human controls? Mandate human controls?
If an ostensibly normal car were to block an emergency response, would its occupants be off the hook if they all denied having driven the vehicle to its current location, and said they have no ability to move it out of the way? I doubt it. They would be at the very least charged with something.
The same thing should go for alleged robotaxis—“But we don't control the car we're sitting inside of!”... not the city's problem. The car you were in was blocking an emergency response.
Would this be unfair? Yes it would. Would this make people fearful of using robotaxis? Absolutely yes it would. And would that put economic pressure on the robotaxi operators to fix these issues as fast as possible? Damn right it would.
I have never run into this where the human driver didn’t get the hint when I remained in front of them with my reverse lights on. They have somewhere to be so sooner or later they go around me (backing up if necessary). From my understanding, the driverless car in the OP just sat there forever.
When roads are closed, I see people trying to maneuver around cones and get through as if the rules don't apply to them.
The self driving cars are much safer drivers than a very high percentage of SF drivers.
We have a liability model in place. Fine the company
We need to stop allowing computational agents (and thus those who deploy them) to escape blame as if they don't exist, or as if their behavior is just nobody's fault. Computational agents need to be viewed as bona fide actors (as people are), with the actions of the computational agent being considered the willful actions of whomever deployed it.
Put 10 different human drivers in the same situation, you'd expect a variance of decisions being made. Put 10 robocars in the same situation, you'd expect 1 decision being made. Is this not how they are being tested? Am I crazy to assume this?
Now, if 12 points loses your license, you have 10 cars, and rack up 120 points across the fleet, it seems obvious that on the whole the fleet is performing worse than a fleet of human drivers we would otherwise ban.
Not an easy problem, to be sure. On one hand, companies can't be allowed free reign to clog up the roads and cause havoc without consequence. On the other, if we NIMBY the development too much, we will be stuck behind countries that don't. In my opinion, we're near the right balance but probably need to take some short term action to make it clear to companies that a remedy for the current behavior around emergency vehicles needs to be priority 0.
Edit: also, it's the same as a recall to me. If one part is deployed to thousands of cars, all of them are recalled. Same applies here for me.
(You may want to sort this out before inserting potential random road blocks into a city.)
Either way, none of those address the new problem of robotaxis. If any normal human did this type of thing, you could go scream at them or honk your horn enough for them to figure it out. If it isn't programmed into the robotaxis logic it won't do anything as shown. Or possibly do the wrong thing.
https://www.ch.ch/en/vehicles-and-traffic/how-to-behave-in-r...
