Discussion of sensor suites

[AnxietyRanger]

[Moderator note (bmah): I moved a number of posts discussing sensors from this thread, because they're not specific to a given software version and seemed to deserve their own thread:

2017.50.3.f3425a1 is out! ]

The rain on that video was a bit borderline. I would say a rain sensor would have done better, but also that in moderate rain for it me it was not that bad. But the wiping did kick in very late often, compared to how a usual rain-sensor in my experience would have done. At times the water accumulation on the windshield did get to borderline dangerous before it wiped. Had I not been testing it, I would have switched to manual mode actually.

Interesting thought, by the way, could the camera heater hurt its ability to see rain. More reasons for a redundant sensor?

 

[Joe F]

Interesting thought, by the way, could the camera heater hurt its ability to see rain. More reasons for a redundant sensor?

How about Tesla put in 5 rain sensors, because you know, the first 4 just might fail. Or, how about 10 extra cameras? Two steering wheels? Now there's a car some here would have nothing to find fault with, I'm sure.

 

[AnxietyRanger]

How about Tesla put in 5 rain sensors, because you know, the first 4 just might fail. Or, how about 10 extra cameras? Two steering wheels? Now there's a car some here would have nothing to find fault with, I'm sure.

I'd be happy with 360 degree radar coverage.

And an ignore list.

 

[Joe F]

And an ignore list.

Can't: BALANCE. Newbies easily misled.

 

[AnxietyRanger]

Can't: BALANCE. Newbies easily misled.

I return to my analogy of the dysfunctional family's holiday dinner. At some point during the dinner Uncle Bob and Cousin Violet will always revert to the same old debate, misrepresenting each other's position while doing so. Things settle down to a civil level after some nudging, but underneath the same misunderstanding and mistrust persist, only to surface again next time.

That's an Internet forum on a good day.

But I do think both Uncle Bob and Cousin Violet would benefit from 360 degree radar (maybe even lidar) coverage and one sun/rain sensor to complement their AP2 suite. Radars make great and relatively easy blind-spot detectors, we would have had that feature for a year already (probably longer had Tesla launched the radars as AP1.5 with Model X/facelift Model S), and ditto with the rain-sensor. (Lidar would provide great additional crash protection and darkness redundancy.)

Instead, we still don't have proper blind-spot detection and, well, the rain-detection took a long time and isn't very good yet. (Luckily blind-spot monitoring should be next, we shall see how good? Rumor summary: Blind-spot cameras, Rain sensing, Level 3, Big battery, Interior/HUD)

Beyond the transition period, the additional technologies would provide (or would have provided, as the case may be) a layer of redundancy, being a different type of technology with different inherent benefits and different operating condition limits compared to cameras.

Now, of course I don't expect rain-sensors to return to Teslas, but one can hope for more radars.

 

[mongo]

I return to my analogy of the dysfunctional family's holiday dinner. At some point during the dinner Uncle Bob and Cousin Violet will always revert to the same old debate, misrepresenting each other's position while doing so. Things settle down to a civil level after some nudging, but underneath the same misunderstanding and mistrust persist, only to surface again next time.

That's an Internet forum on a good day.

But I do think both Uncle Bob and Cousin Violet would benefit from 360 degree radar (maybe even lidar) coverage and one sun/rain sensor to complement their AP2 suite. Radars make great and relatively easy blind-spot detectors, we would have had that feature for a year already (probably longer had Tesla launched the radars as AP1.5 with Model X/facelift Model S), and ditto with the rain-sensor. (Lidar would provide great additional crash protection and darkness redundancy.)

Instead, we still don't have proper blind-spot detection and, well, the rain-detection took a long time and isn't very good yet. (Luckily blind-spot monitoring should be next, we shall see how good? Rumor summary: Blind-spot cameras, Rain sensing, Level 3, Big battery, Interior/HUD)

Beyond the transition period, the additional technologies would provide (or would have provided, as the case may be) a layer of redundancy, being a different type of technology with different inherent benefits and different operating condition limits compared to cameras.

Now, of course I don't expect rain-sensors to return to Teslas, but one can hope for more radars.

I agree, with other sensors, Tesla could have more feature parity now. But they are aiming for down the road where they have all the features with minimal hardware/ cost.

Lame analogy warning:
Oars make a boat move, but won't get your fleet across the ocean (usually), and even if it does, so many oarsmen to feed...

 

[Joe F]

Beyond the transition period, the additional technologies would provide (or would have provided, as the case may be) a layer of redundancy, being a different type of technology with different inherent benefits and different operating condition limits compared to cameras.

Now, of course I don't expect rain-sensors to return to Teslas, but one can hope for more radars.

Perhaps, however if one is a visionary, and can envision what others can not, and is given time to assemble a team that share and execute that vision, great things can be accomplished. Who would have thought landing a rocket on a floating platform in the ocean was even remotely possible?

Personally, I would not want to be tooling down the highway in a bizarre Frankensteinan contraption. Perhaps in the future when devices are less expensive and can be integrated into a vehicle such that there's no impact to COD, or does not blemish its aesthetics, then and only then would I embrace the idea as being something I would consider.

Those blinded by their own disbelief are often left behind, floundering in their entrenched position, holding out for the vision to ultimately fail, only to fail in their uneducated position on that which they know next to nothing.

IMHO.

 

[AnxietyRanger]

Who would have thought landing a rocket on a floating platform in the ocean was even remotely possible?

I would have seen nothing impossible about that. I would have recommended using redundant sensors, though.

Perhaps in the future when devices are less expensive and can be integrated into a vehicle such that there's no impact to COD, or does not blemish its aesthetics, then and only then would I embrace the idea as being something I would consider.

Radars can be integrated completely invisibly.

We shall see where the Tesla Autopilot suite eventually evolves.

 

[AnxietyRanger]

But they are aiming for down the road where they have all the features with minimal hardware/ cost.

I'm not actually sure about that. I think they are aiming at having the minimal sensor suite to start developing, testing and fleet deploying their self-driving software.

Once that software matures and the prices go down, the suite they use in future cars may actually be far more comprehensive than it is today. I expect we are far from the optimal suite even for Tesla, let alone for others.

We shall see. Easy to tell if I was wrong, of course.

 

[arcus]

Instead, we still don't have proper blind-spot detection and, well, the rain-detection took a long time and isn't very good yet. (Luckily blind-spot monitoring should be next, we shall see how good? Rumor summary: Blind-spot cameras, Rain sensing, Level 3, Big battery, Interior/HUD)

As far as blind spot detection, I think the probing frequency and feedback on the dash have improved substantially over the last few software releases to make it reliable. If anything, I would just replace the "halo" rings with a proper car icon instead.

Let's see what the activation of the additional cameras bring!

 

[Rouget]

The lack of additional radars is indeed quite puzzling... but after all George Hotz thinks he can have FSD cars with only one camera and one front radar, so.

By the way, if blind spot detection is next, usually a visual signal appears on the rear-view mirror that an object is there, so any member would be nice enough to dismantle one of them to see if there's one? <3

 

[mongo]

It happens to me almost every time on a 60 mph divided highway. The car definitely "sees" the traffic ahead as supported by the icon on the dash. It will apply the brakes at first, then accelerate like crazy, then slam the brakes but without me having confidence that it will follow through and actually stop.

Depending on set up (pulse rate/ length) a strong return can alias to a multiple of the pulse rate, but I don't expect that under these distances. I wonder if there is a dead band/ hand off issue between the "car directly in front" radar zone, and the "two or more cars ahead" radar zone.

 

[Joe F]

Radars can be integrated completely invisibly.

And, to that point, lidar can't.

[snip]

(Lidar would provide great additional crash protection and darkness redundancy.)

[snip]

Beyond the transition period, the additional technologies would provide (or would have provided, as the case may be) a layer of redundancy, being a different type of technology with different inherent benefits and different operating condition limits compared to cameras.

Now, of course I don't expect rain-sensors to return to Teslas, but one can hope for more radars.

It would also have increased the cost and complexity that Tesla believes isn't required for their end goal, despite how fervently you feel otherwise.

 

[AnxietyRanger]

And, to that point, lidar can't.

Sure. That said, the first production Lidar in a consumer car (well, this side of 1990s Japan), in the new Audi A8, looks about the same as the front radar in an AP1 Model S. Not really an aesthetical issue IMO.

Anyway, I think AP2 suite's redundancy would have been vastly superior even with corner radars... or even just rear corner radars. That would allow it avoid some of the pitfalls of side blockage (e.g. adjacent cars) and rear-view camera blockage (e.g. dirt) with more confidence, and provide the ability to see through the car behind it or on adjacent lanes for fast-approaching rear vehicles...

 

[AnxietyRanger]

It would also have increased the cost and complexity that Tesla believes isn't required for their end goal, despite how fervently you feel otherwise

Perhaps, but that is irrelevant in a conversation about an optimal suite or my wishes for the suite. Tesla has their priorities and not all of them align with the priorities of all car buyers or users - say, me - hence my opinion that differs from their choice.

 

[mongo]

Sure. That said, the first production Lidar in a consumer car (well, this side of 1990s Japan), in the new Audi A8, looks about the same as the front radar in an AP1 Model S. Not really an aesthetical issue IMO.

Anyway, I think AP2 suite's redundancy would have been vastly superior even with corner radars... or even just rear corner radars. That would allow it avoid some of the pitfalls of side blockage (e.g. adjacent cars) and rear-view camera blockage (e.g. dirt) with more confidence.

I understand the desire for multiple data sources. But from a cost/ complexity POV, if you have two sensors, and only one is needed, then why have the other?
If you need radar due to camera failings, and radar can operate with occluded camera, then the radar should be the only sensor since the camera is not adding anything. In the case where data is required from both, then both are needed. If, however, the camera's additional data is always required and the radar is there as a band-aid, then the radar fails to provide sufficient data, and the system fails anyway.

 

[Joe F]

Sure. That said, the first production Lidar in a consumer car (well, this side of 1990s Japan), in the new Audi A8, looks about the same as the front radar in an AP1 Model S. Not really an aesthetical issue IMO.

No, that's an effectiveness issue. Mounted where it is, and given its limited range, it's a failure IMHO.

 

[Joe F]

Perhaps, but that is irrelevant in a conversation about an optimal suite or my wishes for the suite. Tesla has their priorities and not all of them align with the priorities of all car buyers or users - say, me - hence my opinion that differs from their choice.

Actually, and my apology to the thread denizens, this departure from the OP title is irrelevant. Tesla does have their priorities, and our feelings how they proceed are irrelevant. Vote with your wallet.

 

[AnxietyRanger]

Tesla does have their priorities, and our feelings how they proceed are irrelevant. Vote with your wallet.

We may be in the wrong thread, but I don't think our opinions or feelings are at all irrelevant on a Tesla owner's club forum. That's a big part of why we're here, to share our ownership experience?

 

[AnxietyRanger]

I understand the desire for multiple data sources. But from a cost/ complexity POV, if you have two sensors, and only one is needed, then why have the other?
If you need radar due to camera failings, and radar can operate with occluded camera, then the radar should be the only sensor since the camera is not adding anything. In the case where data is required from both, then both are needed. If, however, the camera's additional data is always required and the radar is there as a band-aid, then the radar fails to provide sufficient data, and the system fails anyway.

I'm not even sure if redundancy is the right word here. I'm talking about adding capabilities that currently don't exist, e.g. seeing through a car behind you or on your sides. Radar can offer this, no camera can. However no radar can see everything we need to see to drive, either. Hence the value of the combo.

The added benefit of course is additional security in case of sensor failure or blockage.

A bit similar with the rain sensor. Since it is actively lit and looks at the surface of the windshield, it has night-vision properties no camera has. It could have offered a useful second opinion (while also speeding up the transition).