I didn't say you can't drive a car without being cautious.
I said you can't drive it without constantly braking the moment you detect an object, irrespective of what the object is doing. (e.g. moving into or away from the driving path).
i.e., just because an object was detected 6 seconds before impact did not mean the car ought to have started braking at that moment. It could be that the object was 200 feet away and moving away from the car's driving path, 6 seconds before impact. It'd be absolutely ridiculous to brake in that situation.
We have no information about this context, e.g. the car's data or determinations within the 6 second window. We only know it detected an object 6 seconds before impact.
It appears like the person I was replying to implied 'the braking distance was 180 feet, but the person was 380 feet away, thus uber could have prevented killing this woman had it not shut off the brakes'. In reality, the 6 second figure isn't relevant. What is relevant is the context that allowed a reasonable driver/AI to determine at a particular point in time, that the car should have slowed/braked. And we don't have that information yet. That's what I'm interested in.
The detection at 6 seconds was just of an object though, not an object moving in to the car's path. You couldn't drive a car if you had to constantly break because objects (such as people standing by the road) were being detected.
It's not clear at what point the car ascertained a collision would occur between detection 6s before and the determination that emergency breaking was necessary 1.3 seconds before.
Was there any other determination in between, and when? What I'd like to see is Uber's modelling of the woman's trajectory and the likeliness of collision across the 6 second window. That's completely left unsaid.
The average braking distance of a car is about 24m at 40mph, which is approximately the distance between the woman and the car at 1.3 seconds out. So perhaps the 1.3s figure wasn't the first moment the car determined a brake was necessary, but rather, the last moment the car could have braked to prevent a substantial collision. I want to know the first moment the car determined a brake was necessary at all. It's likely not 6s, but it's also likely not 1.3 seconds. It seems this was entirely preventable, or at least the collision impact could have been mitigated severely, had there been a braking and/or warning system in place.
Shutting off brakes on literally the only driving agent tasked with full attention is inexcusable. But that's what they did. To me that's murder. They used to have two passengers, one for tagging circumstantial data, the other to override the car when necessary and keep eyes on the road at all times. Either keep that and shut off emergency brakes from the car and put a warning system in place for the 'driver'. Or do not shut off emergency brakes. Instead they put a single person in the car, tasked to do things that kept her eyes off the road half of the time, and shut off brakes for the AI. That's insane.
>It's likely that the accident still would have been fatal, or at least the pedestrian would have been seriously injured if the safety driver was driving. Based on the video there was only a second or two to react from the time the pedestrian stepped out of the shadows, and human reaction time varies from 0.7 to 3 seconds [1]. One study found the average driver's reaction time is 2.3 seconds.
Yes, but it is difficult to state that it would have been "likely" fatal, and even sevral fractures are IMHO a big improvement over death, the whole point is not about the "average" braking time (BTW that same study gives 1.64 seconds average for "steering away" which is probably what most human drivers would have done in a similar situation, instead of braking or instead of only braking), it is the peaks that count.
I mean, it is about the (rightful) expectation that an automated system is safer (faster, more reactive) than the best you can find (not the average, the best) "in nature" among a random set of drivers, just like the state codes (correctly) take the worse case in the sample or however incereases the recommendation to three seconds.
If you play the game suggested on the same site you gave a link to and deliberately count up to three before clicking, you get:
Reaction time: 2916.00ms
Reaction speed age: 89 years old
Playing fairly, you are likely to get values in the range 300-500 ms.
They can't detect me slowing down before I start slowing down. So if it's t-4 until impact and I'm still moving at full speed, they would need to start braking now if they can't stop in 4s (assuming the worst case that I continue on my current trajectory).
That being said, I'm happy to find my assumptions about stopping time are incorrect and a car traveling at 25mph can stop in less than a second. So on busy NYC streets this wouldn't be an issue. Even at 50mph it appears that stopping time is sub 3s, so the vehicle could probably have avoided this collision if it were running a more intelligent program.
> If the car in front of you hits something and comes to an extremely abrupt stop and you don't have the stopping distance to stop, you're too close.
It is physically impossible for a car to come to an instantaneous stop for no reason. It would have to e.g. drive into a concrete wall, but if there's a wall across the freeway you should've seen that from afar.
You only really need to have enough distance to be able to notice the car ahead braking and start braking yourself.
> But why didn't it brake from t minus 6 to t minus 1.3? Looks like it detected that the car's and object's paths were converging, so why didn't it brake during that interval?
> Computers have better reaction times than humans. Forcibly braking is a simple maneuver.
The question isn't about the MHZ speed of a computer or the theoretical reaction time but the lack of reaction by the computer in the actual real world. It doesn't matter if this is a sensor failure, a software bug, or a slow reaction by a computer. The result is the same. A person died.
From the article:
*"... the lack of braking or swerving whatsoever is alarming and suggests that the system never anticipated the collision.”
You're massively underestimating human response time to visual cues and also the distance needed to stop a vehicle traveling at 35-40mph. It takes a full quarter of a second to respond to a visual stimulus on average, and more than that to also move your foot and depress a brake pedal. By that time the car was less than 50 feet from the pedestrian. At 40mph braking distance is about 80 feet in good conditions. There is absolutely no way a human driver could have avoided this accident assuming the same visual distance and dynamic range as the camera. Best case, the car may have slowed down a bit before impact.
> To be clear the driver is absolutely partially at fault
I'm... not so certain. Why? The autopilot had likely exhibited proper behavior every time that the vehicle had passed that particular section of road prior, and if the driver was paying full attention to the behavior of the vehicle, he would only notice the problem around the 5 second mark.
Five seconds, if you have no reason to be concerned about the vehicle's behavior, is not much time - especially if you consider that alert drivers are recommended to give themselves a minimum of 4 seconds of reaction time (i.e. follow a vehicle by at least 4 seconds).
> which part of the fatalities happened > 130kmph?
Usually (not always) you realize that something bad is going to happen so you are going to brake. Let's assume there is an obstacle in front of you that is 150m away. After a reaction time of 1 second, and with continuous braking over 100m you are able to get from 180kmh to 100kmh at time of impact, but if you had started from the generally suggested 130kmh, you would have stopped 40m before the target (assuming -8.83m/s^2 braking acceleration).
This was the point I was trying to say, but I think I didn't communicate it well.
I would like to know just how fast these cars can react. I'm betting they could be actually slowing the car before a human is even aware something is wrong, but I would like to have some numbers to cite.
> every other car could slam on their brakes at any time
Or veer out of the way of an obstacle you can't see, leaving you to either similarly veer or collide with it. IOW, do not assume you only need to account for your own reaction time; rather assume the vehicle in front of you can instantly stop moving and you must account for both your reaction time and your vehicle's stopping distance.
My gripe with this argument is that even some humans can have a reaction time to at least start braking under these circumstances, I don't mean all drivers but as an example: I raced go-karts (just as an amateur but took some lessons and also some racing classes in cars with 300 hp) and sometimes my reflexes while racing were surprising, even for myself, narrowly avoiding collisions at 100+ km/h in tight circuits. With that said I'd have thought that sensors and computers could be much faster than even my trained reflexes.
> In other words, these drivers did not leave enough distance to react and brake.
Again, recommended following distance is ~250-300 feet when stopping & reaction time can be 350-400 feet. This is OK when the car in front of you brakes, because you'll both be decelerating with a similar profile and you'll have more than 250-300 feet to stop. It is not OK when the car in front of you stops substantially faster than braking with no warning because it has hit an unknown obstacle.
Right. Because once you have AI the laws of physics no longer apply. The braking distance with an instantaneous reaction time at just 40mph is still about 20 feet.
At the absolute bare minimum, hit the brakes and reduce the impact. An attentive human driver would have at least started to hit the brakes. The software should have had plenty of time (I'm estimating a full second) to do something productive.
> Still, you would think that there would be some threshold where the car decides, "Hey, this stationary obstacle is right in front of me. I should slow down"
I think there is, it's just that by the time it reaches that threshold, the distance to obstacle is less than the stopping distance for the car at 60+ mph. A human making that judgment might steer around the obstacle, but AEB systems don't have that option.
I don't know my cars that well, but an alert driver will take at least 300ms to begin reacting. That's over 8m at 60mph (not sure why we're combining unit systems here). So I'll stand in front of the Model S in this particular scenario and hope to get away with 'mere' broken legs/pelvis. Give the human more time and is probably take the human, not because that's necessarily the more reliable option, but because its distribution of outcomes is more well known to me.
> When the vehicle ahead emergency brakes or stops you have the growth of size of the target to gauge by. That's not much movement, until the last moment when it's far too late to evade.
You see the brake lights up, and as long as you're alert and keep some distance you're going to be alright. If you're not alert and/or driving too close you hit it.
Well, yea, actually it would. Because the vehicle would detect the movement long before I do. And even in that situation, hit the brakes is the right answer.
I didn't say you can't drive a car without being cautious.
I said you can't drive it without constantly braking the moment you detect an object, irrespective of what the object is doing. (e.g. moving into or away from the driving path).
i.e., just because an object was detected 6 seconds before impact did not mean the car ought to have started braking at that moment. It could be that the object was 200 feet away and moving away from the car's driving path, 6 seconds before impact. It'd be absolutely ridiculous to brake in that situation.
We have no information about this context, e.g. the car's data or determinations within the 6 second window. We only know it detected an object 6 seconds before impact.
It appears like the person I was replying to implied 'the braking distance was 180 feet, but the person was 380 feet away, thus uber could have prevented killing this woman had it not shut off the brakes'. In reality, the 6 second figure isn't relevant. What is relevant is the context that allowed a reasonable driver/AI to determine at a particular point in time, that the car should have slowed/braked. And we don't have that information yet. That's what I'm interested in.
reply