2022 Volvo XC90 to ship with SAE Level 3 "capable" LIDAR

[Daniel in SD]

Not sure why they're bothering when Tesla is already shipping Level 5 hardware and should have robotaxis by the end of this year.

The software needed to complete the autonomous-driving capability will be added to the vehicles at a later date via over-the-air update.

Sound familiar?
Anyway it would be nice to see a manufacturer actually make an SAE Level 3 vehicle that works somewhere.
On the plus side, the LIDAR unit looks sleek enough. Reminds me of the roof scoop on a mid engine rally car.

2022 Volvo XC90 to Get Autonomous Highway Pilot Hardware, Lidar

 

[MXLRplus]

We will see. Volvo is owned by Zhejiang Geely of China. If they get it to work, this doesn't bode well for the US automakers and AV companies. China probably already has copies of most US AV source code.

 

[diplomat33]

Looks like FSD with lidar is coming!! The clock is ticking for Tesla to deliver camera-only FSD.

On a more serious note, this is an interesting development because the anti-lidar side has argued that lidar is too expensive for consumer cars and that lidar is a crutch anyway since camera-vision is enough. If we start seeing lidar on consumer cars with reliable FSD, before camera-only FSD, that will destroy that anti-lidar argument.

 

[Uncle Paul]

If LIDAR can truly be installed in such a small nacel, and at a commercially acceptable price it will become a valuable Self Driving tool.

Right now, they are much larger than shown and too expensive and power consuming to install at a price consumers would pay.

IF it can be done with cameras only, that will give Tesla a significant cost advantage.

 

[LowlyOilBurner]

There was talk of this about a year ago on SwedeSpeed (Volvo board). Apparently, the SPA2 XC90 will also be fully electrified with a 400+mile range. By 2022, they should be on the android based infotainment system as well. Right now, the only OTA updates SPA/CMA cars get is maps. It will be interesting to see what they roll out down the road. The system mentioned in the above article is reportedly supposed to cost “in the five figure range” also.

 

[Daniel in SD]

I just want someone to release a Level 3 system. Lots of promises from Audi, Mercedes, and now Volvo but no one has actually delivered.

 

[banned-66611]

Looks like FSD with lidar is coming!! The clock is ticking for Tesla to deliver camera-only FSD.

On a more serious note, this is an interesting development because the anti-lidar side has argued that lidar is too expensive for consumer cars and that lidar is a crutch anyway since camera-vision is enough. If we start seeing lidar on consumer cars with reliable FSD, before camera-only FSD, that will destroy that anti-lidar argument.

This is very bad news for Tesla. If they are beaten to market then the revenue Musk seems to be relying on from robotaxis and the like will dry up fast. Even if they get there but the lidar systems are more reliable they will quickly lose their market share.

It looks like the picked the wrong technology because they both underestimated the time it would take lidar to get small and cheap, and the difficulty of building a self driving system using just cameras.

 

[diplomat33]

This is very bad news for Tesla. If they are beaten to market then the revenue Musk seems to be relying on from robotaxis and the like will dry up fast. Even if they get there but the lidar systems are more reliable they will quickly lose their market share.

It looks like the picked the wrong technology because they both underestimated the time it would take lidar to get small and cheap, and the difficulty of building a self driving system using just cameras.

Well, Tesla is not quite out just yet. IMO, the legacy automakers being slow and risk averse has been a great advantage for Tesla. Audi was going to do L3 and then cancelled it, supposedly because of regulatory push-back in Europe. For all we know, Volvo could cancel the XC90 or cancel the autonomous features. Or maybe Volvo dumbs down the FSD features or has software issues and delays the software updates that make it autonomous. It's happened before.

But I do agree with you that it is looking like it could be bad news for Tesla. Yes, it does look like Elon overestimated how quickly camera based FSD would take. From all his promises, it seems like Elon though that camera-only FSD was already going to happen by now because of the power of machine learning and data. And yes, I agree that lidar got smaller and cheaper much faster than it took for camera vision to get ready. So right now, it is certainly looking like lidar will be the answer.So yeah, Tesla is running out of time to prove their approach will work.

In fact, I think Elon's latest promise of robotaxis next year has more to do with keeping up with what other companies are promising than about the software being ready. It is unlikely that Tesla will actually have the software to do robotaxis next year. But other companies are poised to have robotaxis next year so Elon is promising the same deadline in order to keep up with them.

 

[mikes_fsd]

Tesla is running out of time to prove their approach will work.

by that metric, Google+Waymo have been out of time for a half decade or so.

Until someone delivers usable FSD to the public, they are not "out of time".
Do I wish it would move faster? Heck yeah, I do!

 

[diplomat33]

by that metric, Google+Waymo have been out of time for a half decade or so.

Until someone delivers usable FSD to the public, they are not "out of time".
Do I wish it would move faster? Heck yeah, I do!

I get your point. But the reason I put the onus on Tesla is because Tesla is the one claiming that they can do FSD with just cameras and is trying to beat the competition to market with a camera-only FSD. Tesla's entire FSD strategy is dependent on achieving camera-only FSD in a timely manner.

 

[mtndrew1]

Looks like FSD with lidar is coming!! The clock is ticking for Tesla to deliver camera-only FSD.

On a more serious note, this is an interesting development because the anti-lidar side has argued that lidar is too expensive for consumer cars and that lidar is a crutch anyway since camera-vision is enough. If we start seeing lidar on consumer cars with reliable FSD, before camera-only FSD, that will destroy that anti-lidar argument.

This is an anticipated L3 system years in the future.

Vision must be solved for L5, LIDAR or not.

I remain unconvinced that Tesla will ever do that but it must happen for L5. What I wouldn’t give for proper L3 though. Sigh.

 

[mikes_fsd]

I put the onus on Tesla

To me, it is much more straight forward.
ANYONE - be it Waymo, Zoox, CommaAI, or Tesla - that provides FSD (L4 or higher) to the general public (without screening or other authoritarian preconditions) to be used when they wish and where they wish, wins.
I do not care if they have LiDAR or not.
The barriers to entry for FSD are high already (you MUST solve computer vision to reach L4+).

 

[banned-66611]

Well, Tesla is not quite out just yet. IMO, the legacy automakers being slow and risk averse has been a great advantage for Tesla. Audi was going to do L3 and then cancelled it, supposedly because of regulatory push-back in Europe. For all we know, Volvo could cancel the XC90 or cancel the autonomous features. Or maybe Volvo dumbs down the FSD features or has software issues and delays the software updates that make it autonomous. It's happened before.

If Volvo release L3 in 2022 they will be ahead of Tesla. Tesla is not going to get to L3 in the next 2 years.

I think people underestimate other manufacturers. They are risk averse but not behind. Aside from people like Waymo who are way out in front, take for example Nissan. They have an L2 system that is hands off and uses HD maps with radar. It does lane changes too. Is that ahead of Tesla? The hands off part certainly is, and so are the HD maps. It's been on sale in Japan since last year.

 

[diplomat33]

If Volvo release L3 in 2022 they will be ahead of Tesla. Tesla is not going to get to L3 in the next 2 years.

I think people underestimate other manufacturers. They are risk averse but not behind. Aside from people like Waymo who are way out in front, take for example Nissan. They have an L2 system that is hands off and uses HD maps with radar. It does lane changes too. Is that ahead of Tesla? The hands off part certainly is, and so are the HD maps. It's been on sale in Japan since last year.

Agreed. They are not behind Tesla. It just looks that way because they don't release beta stuff like Tesla does.

For example, other auto makers could easily release the same or better "traffic light response" feature that Tesla just released. They don't do it yet because they believe it would be unsafe and risky to release such a feature as beta.