I just received a software update that included auto rain sensing wipers... thought you may enjoy hearing this 
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Wow - I have them too.
Well done Tesla in 2017 too, a bit late but forgiven already.
The worlds first and only (afaik) commercial application of image based dynamic neural network auto wipers.
Still think a sensor would have been simpler but what the heck, now I have DYNNAW
(just hope there's enough processing power left for some more AP2 goodies)
I suppose when they don't behave quite as you would like them to you can explain to your passengers
- don't worry the wipers are still learning
wonder if they will ever become self aware ...
Well done Tesla in 2017 too, a bit late but forgiven already.
The worlds first and only (afaik) commercial application of image based dynamic neural network auto wipers.
Still think a sensor would have been simpler but what the heck, now I have DYNNAW
(just hope there's enough processing power left for some more AP2 goodies)
I suppose when they don't behave quite as you would like them to you can explain to your passengers
- don't worry the wipers are still learning
wonder if they will ever become self aware ...
I wouldn't worry. Neural networks aren't quite the same as other forms of programming, and not all new algorithms have any additional cost. In this case, it looks like the rain sensing is just one(of many) outputs of the same neural network that does anything. I'm going to guess that it was probably needed in that network anyway, even if not used for this purpose, in order to allow the network to learn around the distortions caused by rain.
Due to neural nets being mostly black-boxes, nobody not really understanding why it works, I believe it's best to separate them to keep them manageable. You'd not want lane keep regressions when trying to improve rain detection accuracy. I believe verygreen also confirmed a separate net for this.
At some point later, the FSD can probably add a general "I can't see, try to wash/wipe" algorithm to its self-driving mechanism.
Rain detection could probably do fine at 10fps. And their net is pre-learned. Does not require lots of proccess power.
The "not understanding how they work" bit is only partly true. We don't understand the details of *how* a given non-trivial neural network is solving its problem. We do understand fairly well how(and why) the learning process works, what kind of things can help a network learn, what architectures work well for given tasks, etc.
Separating networks is not always a good idea. They should usually be kept separate if their tasks lack any overlap. However, it's fairly common to have a network produce multiple different outputs, because learning about how solve one problem helps the network solve a different, but related problem. I work in the field and have seen multiple different networks produce the *same output feature* because learning to produce that particular feature made the network have better accuracy in its main task.
I see a path here for that to be true for the rain sensing network. The object detection outputs of the network need to be robust to distortions produced by raindrops(or broader areas of water) so the car doesn't suddenly hallucinate an obstacle in front of it. Having it explicitly learn about detecting areas of distortion will likely help keep it robust to that distortion.
For @verygreen, my understanding of what he said is that there is a separate *output* for rain level, not a separate network.
Yeah, those are good points. As long as the problems are enough related. I guess rain sensing can be if it's at a general "Something is obstructing my view", but not sure if the existing neural network is mature enough to know what to expect to see. So my guess is that it just recognizes droplets at this point, and will be improved later.
No, according to @verygreen there is a separate neural network for this - the only NN using the fisheye now anyway... The NN in question is called: fisheye_autowiper
Links here if anyone wants to research: Rumor summary: Blind-spot cameras, Rain sensing, Level 3, Big battery, Interior/HUD
Fair enough, misunderstood then. Still, Id expect that, at some point, it’ll just be part of the object recognition for the same reason as above.
DoubleACL
Member
In FL, the rain sensing wipers are working great for me first time out. This closes a big gap in features.
need to be a touch more sensitive to fine drizzle, other than that pretty good.
I wonder if one of hte NN inputs is if the driver prods the single wipe button, as this could be used increase the probability of wipe requirement, thus tuning the system to a driver's preference.
Often thought that AP is a bit "thick" in this repect, as when a driver in the same GPS location daily has to disconnect AP in precisely the same place to avoid a tree mating ritual due to some dodgy road markings, always though AP could learn this, Maybe it thinks it is a superior intellect already.
I wonder if one of hte NN inputs is if the driver prods the single wipe button, as this could be used increase the probability of wipe requirement, thus tuning the system to a driver's preference.
Often thought that AP is a bit "thick" in this repect, as when a driver in the same GPS location daily has to disconnect AP in precisely the same place to avoid a tree mating ritual due to some dodgy road markings, always though AP could learn this, Maybe it thinks it is a superior intellect already.
There are two obvious ways to accomplish what you're suggesting. The first would be to adjust the threshold it uses for rain detection slightly within some defined bounds. The problem with this is that this would always tend towards loosening the threshold entirely, because the network could only ever get affirmative input(there's no button to press for "stop wiping") and also that it may be totally spurious(splatter on a part of the windshield the camera can't see). This is also kind of a one-size-fits-all solution that would generally just make it more likely to run the wipers under all situations, rather than targeting one area of weakness.
More potentially useful would be a network that shifts its actual weights(or another internal representation) based on this input(likely some kind of LSTM recurrent network). Unfortunately, these tend to be much more difficult to train.
Those are probably why you haven't seen much movement towards what you suggest for either the wipers or AP generally.
More rain here in Palm Desert!
Finally got a longer more real world test on camera.
They perform perfectly in position 2!!!!
Everything I hoped for.
2+ minute clip Thanks Elon!
Finally got a longer more real world test on camera.
They perform perfectly in position 2!!!!
Everything I hoped for.
2+ minute clip Thanks Elon!
Raining in LA and tested the auto sensing wipers on my AP 2.5 MS. I felt it was too sensitive even on intermittent option 1. Started drizzling and the wipers activated but it immediately wiped a second time even when the windshield had no rain drops. It kept doing this and I was afraid it will eventually ruin the blades and possibly scratch my windshield so I turned off the feature. But it stopped wiping when I was under a bridge with no drizzle. Once the rain started pouring the wipers worked perfectly.
P100DHG.
Member
It finally rained in LA so now we get to join the discussion. I would say it might be overly sensitive. I am leaving mine on the first position and it wipes too often for me. I noticed a pattern. Wipe twice, pause. Wipe twice, pause. It happens in quick succession. Wipe wipe pause, wipe wipe pause. Clearly that’s a condition of the amount of rain but a wipe here and there is plenty in this rain. And if that’s not articulate or specific enough well...
Edit: @eddy009 stole the words from my mouth.
More rain here in Palm Desert!
Finally got a longer more real world test on camera.
They perform perfectly in position 2!!!!
Everything I hoped for.
2+ minute clip Thanks Elon!
Finally got a longer more real world test on camera.
They perform perfectly in position 2!!!!
Everything I hoped for.
2+ minute clip Thanks Elon!
TaoJones
Beyond Driven
You know when you're a Tesla enthusiast when.... you find yourself watching YouTube videos of windshield wiper activity.
In position 2, did you not expect more frequent wiping given the accumulation of raindrops? Iirc the release notes, position 1 was "medium" and position 2 was more... active, yes?
In any case, nice video.
In my limited testing today, the wipers are certainly operating differently than before when in positions 1 and 2. Can't yet say whether better than the old-style timed/consistent wiping, but it's not worse.
ucmndd
Well-Known Member
I've had a couple days to give these a test and in daylight they have exceeded my expectations, working quite well. Only once or twice where I thought they were wiping quicker or more slowly than I would have preferred.
They are worthless at night - don't wipe at all until something illuminates the windshield like an oncoming car or streetlamp. I have a feeling this is going to be a hard problem to solve.
They are worthless at night - don't wipe at all until something illuminates the windshield like an oncoming car or streetlamp. I have a feeling this is going to be a hard problem to solve.
Yes, very excited even though it hasn't rained yet and my car clears impending fog mist.... pretty darn impressive.
