The last Waymo leak claims that it is more than just 'progress', but that 'rain is solved', incidentally: https://thelastdriverlicenseholder.com/2018/09/14/waymo-plan...
These sound like edge cases but there are places where they occur very frequently and local human drivers, for all their faults, do have a way of dealing with them.
I'd say that's pretty calculable, and has a solution, slow down to a speed that the tires can shed water.
Perhaps most telling about your qualification is the misplaced risk; the light precipitation with a little oil slick is likely more dangerous, mostly because you don't slow down for it.
Anecdotally, an even harder problem is when water exceeds 4", negotiating with traffic so you won't float away or get water in your intake.
In fact, they will have a lot more prior experience than any human driver in the world.
While it's quite fun if you pin up a cat dressed up in an elephant costume and surely amuses a few colleagues, a self-driving car is not something that should confuse an Elephant for a Cat.
ML is IMO not reliable enough for use in self-driving cars, in complicated situations the driver should take over.
Similarly, my 6 year old is also not reliable enough to pilot a car.
Both will change as they mature.
[0] I mean, there are going to be cases, but the number of people currently driving in horrific weather conditions because it would be inconvenient not to is too damn high.
On I95 in the DE-NY area, it's more like 60 mph in heavy rain. Or even in moderate snow. Everyone knows that they could never stop. But they're counting on others' inability to stop ;) But occasionally, there are massive chain-reaction pileups :(
Humans can and do drive safely in these conditions, however, and it's a valid concern to point out that Waymo's West Coast-centric models might not yet be trained to handle these cases.
(To be clear, I don't doubt that machines will eventually surpass humans in whatever weather conditions, if they haven't already.)
I've seen people do all sorts of foolish things. When the rain is so hard that I can't see more than 15 feet, I pull over and stop.
It's certainly not an unanswerable question. It's just quite hard to verify that they answer you have is a sufficiently correct one.
Why? It seems to me that estimating the confidence of a prediction by the quality of measurements is not an infrequent machine learning task. It'd be more complex for a driving algorithm than calculating the confidence interval of a linear regression, but it's in the same theme.
Self-driving cars would be putting lives at risk if they just pulled over and shutdown on every summer thunderstorm.
Can confirm, just spent 30 summer days in Florida. People don't pull over and stop, they just put on their "hazards" and slow down, sometimes to 25-30 MPH depending on the intensity of rain. This happens every single day.
If you're in the center lane and traffic is heavy on both sides of you, while your windshield (or camera) is overwhelmed with water, you need to keep moving with traffic for a bit -- or you get rear-ended.
https://en.wikipedia.org/wiki/List_of_countries_by_traffic-r...
> Also snow seems to be solvable, as there will probably similar patterns that the sensors get in their signals.
Real snow is not just a software problem. Fancy rear facing sensors near bumper? I hope they function through a couple inches of frozen muddy spray. Front windshield? Sometimes needs scraping. Lidar on roof? Does it work through the big heap of snow on it?
Still, for a wide range of weather conditions the autonomous car will just say a big NOPE and refuse to drive anywhere. Which is why it must always have manual controls. The autonomous car that could take me to the city through the worse conditions I experience every winter, won't have to take me in it because that car is probably clever enough to do my job as a software developer too.
But I guess when it's your product; you change your mind.
*This is not an attack to alkonaut.
Though I'm sure SV's real solution to this is to have perfect weather (ie. "can't reproduce; closing bug").
It seems that developers for self driving cars are dealing with a very different, and in some ways more difficult, environment than aircraft auto-pilots. While cars only have to deal with two dimensions instead of three and cars are still less complex machines compared to planes there are a lot more challenges. Unregulated and not controlled environment, traffic ranging from cars to trucks to motorbikes to bikes to pedestrians to playing children. Less room for sensors in the car, changing road condiitions, animals, you name it. Not a simple problem to solve.
And with a near perfect track record of similar systems in the aerospace sector the bar for acceptance hangs pretty high. Add to that the fact that pilots are trained in the use of these systems while the average joe / jane in a car is not. And the car should be able to drive without driver in the end.
So, I could imagine developers of self-driving cars could learn a thing or two from aerospace. But that is just my opinion.
Yes, sometimes you can pull over and sometimes you just have to put the blinkers on, stop in the traffic lane, and hope for the best--even though it can be very dangerous. One challenge is that, with full autonomy, the vehicle needs to be able to very quickly deal with a once-in-10-years scenario in the least bad way--both for its occupants and for any other traffic on the road.
Surprising that Waymo has figured out rain but not things like left turns. Different problem spaces I'd imagine, but still surprising.
The waymo cars have trouble with getting hit while making unprotected left turns because they drive very cautiously. Except for the time they hit a bus, because they somehow thought a bus would swerve to avoid them (although the then Google vehicle was being overly cautious and trying to avoid an inconsequential object in the road). All in all, I would rather the waymo style of overcaution and immobility than the Uber style of over confidence and ignoring of potential obstacles. However, I would currently still prefer mediocre to average human drivers on the road with me over a waymo.
From what I've read here and in the linked article, a Waymo car would be unable to pass a driving test here in Germany. Consequentially, they shouldn't/couldn't be allowed unsupervised on German roads, just like a learner driver. (In Germany driving schools are mandatory for a license and their cars are fitted with dual controls for the instructor).