AP2.0 Cameras: Capabilities and Limitations?

[lunitiks]

Are you saying the new HVAC sensor is also wired to wiper action in the manual?

In your previous post you said it was wired to BCM, but is there some diagram showing a connection to wiper action through there?

(Other than theoretically both being connected to the same BCM of course...)

English isn't my language so I'm sorry if my posts are a little confusing

What I'm trying to say is that the new HVAC sensor is connected (only) to the BCM but on the diagram page where the HVAC sensor is shown, the BCM only refers to wiper functionality.

You see the BCM does a lot of stuff, and it's referenced on a bunch of pages. But each page concerns different aspects of the car, and the BCM info is therefore different on each page - excluding irrelevant signaling. For example, on the page that deals with Air Suspension, the only I/O shown for the BCM deals with ride hight. Reading this page only, you wouldn't have a clue that the BCM also is a hub for wiper and headlight action: These functions are listed on other pages.

So the peculiar thing is that the HVAC sensor is listed on the page dealing with windshield wipers and washer pump. That's all

 

[AnxietyRanger]

What I'm trying to say is that the new HVAC sensor is connected (only) to the BCM but on the diagram page where the HVAC sensor is shown, the BCM only refers to wiper functionality.

Understood.

However, it could also only mean that a non-rain sensing HVAC sensor replaced the rain sensor when rain sensing moved to the APE, and manual makers did not fully redo that page of the manual (given that we know the manual page on AP1 did indeed cover rain sensing).

Maybe the old Rain Sensor was sort of Rain + HVAC Sensor, but now all that is left is an HVAC Sensor...

Just speculating. I do think your theory warrants consideration and further research.

 

[bjornb]

I agree with @AnxietyRanger, I still believe that the sensor we can see on AP2 cars is a light sensor, and not for detecting rain.

I also agree with the posts above that it should be possible to detect rain especially on the lower part of the fisheye image, but I dont understand why Tesla went to this step when they instead could have continued to use an (inexpensive?) off the shelf component..

BTW I also mentioned months ago in this thread a different (older) patent regarding camera based rain sensor:
AP2.0 Cameras: Capabilities and Limitations?

 

[J1mbo]

Note: auto-high beams is US-only at the moment as well.

Suggests a legal issue rather than a LHD / RHD issue

 

[Joerg]

First post guys...
I recently noticed on our AP2 X a kind of partial working rain sensing. It was raining quite heavily, I switched the wipers on (not interval, full), then the rain stopped but I didn't switch off the wipers or set it back to interval - kept it running. Then after 3-4 minutes the wiper 'automatically' went to interval mode. So still wiping, but at least not constantly.
I was pretty suprised and that was the first time I've seen that (FW 28.c528869).

 

[JeffK]

First post guys...
I recently noticed on our AP2 X a kind of partial working rain sensing. It was raining quite heavily, I switched the wipers on (not interval, full), then the rain stopped but I didn't switch off the wipers or set it back to interval - kept it running. Then after 3-4 minutes the wiper 'automatically' went to interval mode. So still wiping, but at least not constantly.
I was pretty suprised and that was the first time I've seen that (FW 28.c528869).

Nice, that at least sounds like what the owners manual describes

Someone needs to try this out with a hose.

 

[AnxietyRanger]

Nice, that at least sounds like what the owners manual describes

Someone needs to try this out with a hose.

Well not really, @Joerg said the wipers were put into full instead of auto?

 

[JeffK]

Well not really, @Joerg said the wipers were put into full instead of auto?

hmm.. then it shouldn't do that.

 

[AnxietyRanger]

hmm.. then it shouldn't do that.

Agreed, sounds a bit weird... but I guess anything is possible

 

[SpotfireY]

@stopcrazypp that's interesting but seems to be a really old paper... Too bad it's just an abstract. Do you have a link to the actual doucment?

So essentially it would rely on light emission the same way as traditional rain sensors (except here the receiving end is a 'multi purpose' video camera).

I just read the paper and it's essentially describing the same thing as the Delphi patent. There's also someone from Delphi quoted as a co-author... So this migth actually be the paper for exactly that delphi patent. It uses a dedicated portion of the camera's FOV together with dedicated optics and a light source to detect the rain. Note the divider and the extra lens at the bottom.



Here's the conclusion (emphasis mine):



Their test results seem to be quite impressive.



Now let's recap: their system with extra hardware ends up doing general sensing better but it's also easier confused and obstructed than the traditional rain sensors. Also, Tesla isn't using any specialized hardware for the rain sensing portion. Granted, things coud be very different because the researchers didn't use a wide angle camera with a very wide field of view like Tesla does (their cam seems to be similar to the traditional AP1 camera or the AP2 medium range cam). This last point might help since we already concluded that there might be a chance of some parts of the windshield actually being in focus. But I feel like without the additional light source rain sensing in dark environments will definitely be a problem.

 

[croman]

People are mistaken about a great many things. The first day that I got the update with the Tom Tom GPS, I thought AP2 was reading speed limit signs but it was just that much more accurate about its position vis-a-vis the speed limit signs.

It rains a lot in Chicago and I run that fw and I've seen nothing of the sort as described by OP. Still not saying its not possible, but its not something I believe without substantiation.

As far as the Delphi patent mentioned earlier, there is only 1 independent claim (claim 1). The rest are dependant and some are dependant on another dependant claim (claim 3). So it might not actually be something that Tesla needs to contend with because I do not believe that Tesla uses the particular lens that is claimed in Claim 1 (therefore their method is distinguishable from Delphi's). I will look in public PAIR later to look at the prosecution history, including any office actions relating to this application.

 

[stopcrazypp]

Note: auto-high beams is US-only at the moment as well.

Suggests a legal issue rather than a LHD / RHD issue

Interesting point. I do wonder how much of this is engineering and how much of this is legal wrangling over patents.
I guess that is a risk of going with a custom solution instead of off the shelf parts.

 

[AnxietyRanger]

Regarding the wiper speed: could it be drive speed related. Slower wiper when stopped for lights?

Can't remember what mine does...

 

[AnxietyRanger]

But I feel like without the additional light source rain sensing in dark environments will definitely be a problem.

Rain in the dark indeed seems like another potential issue.

We shall see...

 

[JeffK]

Rain in the dark indeed seems like another potential issue.

We shall see...

I'd imagine there are lights somewhere like the car's own headlights, oncoming headlights, ambient light, etc.

 

[lux_cars]

I gotta be amazed at all the attention /posts the rain sensing wipes gets..... Folk's really want their rain sensing wipers!

I think that's false psychology, though.

Some of us are genuinely interested in the technical details, merits of policies and whatnot. It is a hobby. I know it seems silly for those not so inclined, but rarely is it because e.g. we "want their rain sensing wipers".

I genuinely want good functioning rain-sensing wipers! If my car is going to drive itself, it sure as hell should wipe its own windshield! It's 2017, we shouldn't have to do these little things manually LOL.

 

[AnxietyRanger]

I'd imagine there are lights somewhere like the car's own headlights, oncoming headlights, ambient light, etc.

Perhaps the dynamic range of the camera is enough, who knows. It is an interesting point, still.

 

[croman]

Looked into Delphi patent more. I'm not a patent attorney, so I'm only of so much use. The invention relies on a more sophisticated camera than I think is in our cars. It has the ability to sense full spectrum of color and near infrared as far as I can tell. I doubt our cameras can do that. Further, they disclosed an even older patent from Bosch in Germany. This one seems more applicable as it is the invention that actually does visual rain sensing.

 

[BigD0g]

I just want to see some freaking evidence of this S curve Elon refers to in the self driving advancements section, I don't see any curve, I see a very very gradual ramp so far....

 

[JeffK]

I just want to see some freaking evidence of this S curve Elon refers to in the self driving advancements section, I don't see any curve, I see a very very gradual ramp so far....

S curves happen in nearly all manufacturing.

When you look at the beginning of the S curve ... Since this one only started only last month it does appear linear, but it's not. You simply don't have enough data points yet.

As far as unofficial data points, we've seen VINs over 100. It took a month to produce 50 units but they've obviously done more than all of last month's models in the last two weeks. This puts them on track for 100+ units in August. If September production is more than 200 units, then you have your S curve beginning its logistic growth curve.