Augmented Reality

[yesup]

Regarding the width of HUD have a look at this. It also overcomes the issue of polarized sunglasses that comes up often as another negative of a HUD.

DLP HUD has already been quoted in this forum a few times. Again, it is expensive and has not been used in any production car despite its claimed advantages. So would Tesla bet their entire Model 3 production line on one supplier (Texas Instrument)?
And the photo on that page is misleading. TI's DLP HUD maximum coverage is 12 degree angle of view - that's approx equal to a 200mm telephoto lens. And that's much smaller coverage area than the HUD graphics in that photo would suggest.

 

[3Victoria]

Answer will come within 6 weeks.

 

[calisnow]

I just want the car.

For some reason lately - I have been feeling like Tesla isn't as far as they used to be concerning technology. AP1 ( partnering with another company ) seems to be so much better than AP2 ( Tesla going solo concerning the programing ) even in February 2017. With that said...I'm wondering if Tesla would be building its own AR if they choose to go that route.

People keep forgetting AP 1 hardware was in the cars a full year before going live - and it was quite glitchy in 7.0 in Oct 2015. AP 2 is progressing much faster than AP 1 did.

Jeff, we would just have to disagree on the difficulty in AR implementation, so I would leave it as that.

So let's get back to the reality, there is still no hardware/software implementation in any production car. So if Tesla want components which are economic viable and reliable; and there would be plenty of parts available with no supplier worries for Model 3, it simply does not make sense to implement this type of unproven technology with limited benefit (both in terms of driving and chance of increasing car sales) for their base mass production models.

@yesup - probably didn't make sense to land rockets upright in the ocean, launch autopilot etc. or it didn't seem to everyone else at the time. So if you know so well what Tesla won't do - tell us what you think they *will* do then - eh? What is the spaceship display going to be?

 

[dsvick]

So let's get back to the reality, there is still no hardware/software implementation in any production car. So if Tesla want components which are economic viable and reliable; and there would be plenty of parts available with no supplier worries for Model 3, it simply does not make sense to implement this type of unproven technology with limited benefit (both in terms of driving and chance of increasing car sales) for their base mass production models.

I don't think a HUD with AR would require any different hardware than a normal HUD. Aside from the hardware that actual gets the information used in the AR part of it, which is already included, all of the AR display is software generated and then incorporated into the normal HUD.

I think you're right though in that they won't introduce it in the Model 3 first, if it's going to happen we'll see it in the S and X first. That would give them several months of testing and refining to get it right.

 

[dsvick]

So if you know so well what Tesla won't do - tell us what you think they *will* do then - eh? What is the spaceship display going to be?

It was never going to be a spaceship display, the quote was referring to the steering controls. That doesn't preclude a HUD but he wasn't talking about displays when he made that quote.

 

[Garlan Garner]

I don't think a HUD with AR would require any different hardware than a normal HUD. Aside from the hardware that actual gets the information used in the AR part of it, which is already included, all of the AR display is software generated and then incorporated into the normal HUD.

I think you're right though in that they won't introduce it in the Model 3 first, if it's going to happen we'll see it in the S and X first. That would give them several months of testing and refining to get it right.

It all depends on what type of AR you are talking about.

For instance - AR with gesture control would require additional hardware. AR with retina distinction would require additional hardware.

I wonder what people think AR is.

 

[JeffK]

It was never going to be a spaceship display, the quote was referring to the steering controls. That doesn't preclude a HUD but he wasn't talking about displays when he made that quote.

Although I would love a futuristic AR HUD, there's also the possibility that some essentials are integrated into the wheel itself.

What concerns me is also the original tweet:

The issue is that Part 2 was the FSD part. Did he actually mean part 3? Being FSD would explain the lack of a normal instrument panel (see google err waymo cars)

The more I think about it the more I consider not needing a HUD at all. I firmly believe AR HUDs will lead to safer manually driven cars someday, but if we pretend for a moment that FSD becomes perfected in the next 5-10 years, there may not be a real need for a HUD at all.

My fingers are still crossed for an AR HUD, but I'm not holding my breath.

 

[electracity]

DLP HUD has already been quoted in this forum a few times. Again, it is expensive and has not been used in any production car despite its claimed advantages. So would Tesla bet their entire Model 3 production line on one supplier (Texas Instrument)?
And the photo on that page is misleading. TI's DLP HUD maximum coverage is 12 degree angle of view - that's approx equal to a 200mm telephoto lens. And that's much smaller coverage area than the HUD graphics in that photo would suggest.

I agree with this. The 12 degree FOV is a huge issue, regardless of software capability. A HUD that illuminated secondary objects such as pedestrians would probably need to be over 45 degrees or more and have a huge projection area on the windshield.

The HUD seems likely to be a large driver display, while the center LCD is the shared display. The car probably is safely operable with the center display only. I do think the HUD will be DLP-like, but non-critical for car operation. I imagine Tesla placing an order for a half million TI DLP chipsets with an upside of millions get the cost way down.

Note that Musk said that the center display is flat oriented so that the passenger can more easily see it. This suggests that the driver of the model 3 has a high res display in the HUD.

I'm sure Tesla has worked at developing full augmented reality in the windshield. But besides the practicality, the potential for liability in accidents is significant. If the HUD can't be seen due to glare, who is liable in an AP accident?

So I think the HUD is more likely to be on the S3XY end of the continuum and less of a critical safety system. How software upgradable the initial HUD will be will be interesting. The HUD hardware would have been locked in last year. The software will fall short of what Musk wants on initial release, of course.

 

[JeffK]

The car probably is safely operable with the center display only.

I'm sure Tesla has worked at developing full augmented reality in the windshield. But besides the practicality, the potential for liability in accidents is significant. If the HUD can't be seen due to glare, who is liable in an AP accident?

What might you not see in a HUD due to glare that would lead to an accident?

I totally agree it wouldn't be for critical systems but probably more for information:
what is autopilot doing?
navigation
highlighting potential threats/hazards: animals, pedestrians, cars, etc.
Maybe speed and some indicators (these don't need to be seen 100% of the time to drive safely)
etc

 

[S3XY]

DLP HUD has already been quoted in this forum a few times. Again, it is expensive and has not been used in any production car despite its claimed advantages. So would Tesla bet their entire Model 3 production line on one supplier (Texas Instrument)?
And the photo on that page is misleading. TI's DLP HUD maximum coverage is 12 degree angle of view - that's approx equal to a 200mm telephoto lens. And that's much smaller coverage area than the HUD graphics in that photo would suggest.

The 12 degree angle of view is mentioned on the main page but if you dig further (see bottom of page 7) they say that 20 degrees is possible with a higher resolution display source.

 

[S3XY]

I don't think a HUD with AR would require any different hardware than a normal HUD. Aside from the hardware that actual gets the information used in the AR part of it, which is already included, all of the AR display is software generated and then incorporated into the normal HUD.

I think you're right though in that they won't introduce it in the Model 3 first, if it's going to happen we'll see it in the S and X first. That would give them several months of testing and refining to get it right.

Agreed. And at the M3 final reveal they could announce an AR/HUD and that it will be installed in the S and X starting now. That would give them a few months of tweaking in those models before the first M3's are delivered.

 

[electracity]

The 12 degree angle of view is mentioned on the main page but if you dig further (see bottom of page 7) they say that 20 degrees is possible with a higher resolution display source.

A pedestrian 45 degrees from the center of of the divers vision would require a HUD with a field of view of more than 90 degrees. This is why augmented reality today is done with display close to the eye.

Even windshield HUD highlighting an object blocking the roadway is problematic in hilly environments. The road surface may be above the display if the car is oriented downhill.

I don't believe Tesla can use HUD for safety unless it is near 100% reliable. They need to be able to blame the human driver for AP accidents.

 

[JeffK]

Even windshield HUD highlighting an object blocking the roadway is problematic in hilly environments. The road surface may be above the display if the car is oriented downhill.

I don't believe Tesla can use HUD for safety unless it is near 100% reliable.

Considering drivers only have a pedestrian warning in maybe a handful of production cars on the road. I think even if it's not 100% reliable then it's better than nothing especially if it's combined with an audible alarm. Even at 1% reliability it still might actually save lives.

 

[Garlan Garner]

Although I would love a futuristic AR HUD, there's also the possibility that some essentials are integrated into the wheel itself.

What concerns me is also the original tweet:

The issue is that Part 2 was the FSD part. Did he actually mean part 3? Being FSD would explain the lack of a normal instrument panel (see google err waymo cars)

The more I think about it the more I consider not needing a HUD at all. I firmly believe AR HUDs will lead to safer manually driven cars someday, but if we pretend for a moment that FSD becomes perfected in the next 5-10 years, there may not be a real need for a HUD at all.

My fingers are still crossed for an AR HUD, but I'm not holding my breath.

What is an FSD?

 

[Garlan Garner]

There needs to be a thread titled: What do you think AR is?

 

[MorrisonHiker]

What is an FSD?

Full Self Driving

 

[electracity]

The problems with getting AP2.0 back to mobileye functionality show how far Tesa is from FSD.

 

[Garlan Garner]

The problems with getting AP2.0 back to mobileye functionality show how far Tesa is from FSD.

I would have to agree.

Mobileye is still better than anything Tesla has come up with on their own. I wonder if Tesla thought it would be easy to implement or duplicate mobileye.

 

[dsvick]

It all depends on what type of AR you are talking about.

For instance - AR with gesture control would require additional hardware. AR with retina distinction would require additional hardware.

Gestures and retina tracking are just a way to interact with the AR. You can have AR and certainly just a HUD without either of them.

There needs to be a thread titled: What do you think AR is?

I think we've pretty much established that.

 

[dsvick]

The problems with getting AP2.0 back to mobileye functionality show how far Tesa is from FSD.

I would have to agree.

Mobileye is still better than anything Tesla has come up with on their own. I wonder if Tesla thought it would be easy to implement or duplicate mobileye.

Considering that the hardware that is providing the data, the type and amount of data, as well as the processor itself and probably a lot of the software have all changed from what was used with AP 1, it is not surprising at all that it would take some time to get it operating at the same levels. Why would you think it wouldn't?