Dream Drive upgrade

[HC_79]

That’s the video I used to convince my father to get a GT, like mine! I think it’s safe to say that hands off/eyes off are way off the table at this point!

I’m more annoyed about the silence than anything. Mine was obviously bundled into the GT but for those who paid $10K it’s a shady thing to do with no explanation.

 

[coma24]

wait, is DD pro standard on all GT's? I didn't realize that was the case.

 

[Bobby]

wait, is DD pro standard on all GT's? I didn't realize that was the case.

Everything was standard on GTs in the very beginning. Later on, items were made optional.

 

[Batteryman]

I’m more annoyed about the silence than anything. Mine was obviously bundled into the GT but for those who paid $10K it’s a shady thing to do with no explanation.

I paid the 10k, but I never expected hands free driving soon since it’s a new company. But nice to know you have the tech when anything gets added. Why would you think Lucid will miraculously have Tesla like FSD so soon? It took Tesla 20 years!

 

[idiot900]

I have DD Premium, and my Lucid Motors account shows that I can buy DD Pro for an additional $12k...(my 1/2023 build had all the hardware included)

Perhaps one day I'd consider paying the difference between what I paid for DD Premium and what Lucid actually charges currently for DD Pro, but $12k for lane keeping is absurd in today's market.

 

[borski]

Personally, I would not enjoy the car as much without DD Pro. I agree that you are paying a heavy premium for “future ready”, but not having HA or the 360 cameras would be a problem for me. I just wouldn’t get it expecting Lucid to add more DD Pro features because they are absolutely silent on what, if anything is coming. It’s been a long time since any meaningful features have been added and at this point I don’t expect anything.

Oh ye of little faith

 

[HC_79]

I paid the 10k, but I never expected hands free driving soon since it’s a new company. But nice to know you have the tech when anything gets added. Why would you think Lucid will miraculously have Tesla like FSD so soon? It took Tesla 20 years!

I never expected Lucid to have FSD 3 years in. What I did expect is for them to add things like auto lane changing and auto park/summon etc. like promised.

Instead, they’ve chosen to bury their head under a rock and say nothing. You just need to look at the Tesla path to know this “future ready” hardware will outdate very quickly so the longer Lucid leaves it the more inclined I am to believe nothing will be added.

 

[hmp10]

. . . this “future ready” hardware will outdate very quickly so the longer Lucid leaves it the more inclined I am to believe nothing will be added.

Except for Forward Looking Infrared (FLIR) I don't know of any new sensing technology that is on the automotive horizon. My guess is that all of Lucid's shortcomings with ADAS right now lie in the software domain.

I understand the frustration people are having with Lucid's lack of progress in this area; it is the primary reason my brother right now is refusing to consider a switch from Tesla to Lucid for his next car trade, so I hear about it a good bit. However, there have been news reports that Lucid is working with a South Korean firm (I forget its name) to switch to its technology and maybe away from HERE, so I don't the effort is yet DOA.

 

[Adnillien]

there have been news reports that Lucid is working with a South Korean firm (I forget its name) to switch to its technology and maybe away from HERE

The Korean group is HL group (HL Klemove) that they are rumored to be working on ADAS sensors with. HL is one of the bigger 24GHZ radar sensor suppliers and are actively trying to move into the 77GHz radar sensors which are used for all of the Air radar modules. With respect to Here maps, I doubt that Lucid will move away from Here maps.

 

[HC_79]

Except for Forward Looking Infrared (FLIR) I don't know of any new sensing technology that is on the automotive horizon. My guess is that all of Lucid's shortcomings with ADAS right now lie in the software domain.

I understand the frustration people are having with Lucid's lack of progress in this area; it is the primary reason my brother right now is refusing to consider a switch from Tesla to Lucid for his next car trade, so I hear about it a good bit. However, there have been news reports that Lucid is working with a South Korean firm (I forget its name) to switch to its technology and maybe away from HERE, so I don't the effort is yet DOA.

The problem is that hardware outdates very quickly, it doesn't become obsolete but it goes through upgrade cycles. Is the LiDAR sensor fitted in the car today still relevant to the LiDAR sensors that are available today, are the cameras still relevant to what cameras are available today, etc. Lucid dragging its feet on the software portion of DDPro will only make this gap worse.

The Korean group is HL group (HL Klemove) that they are rumored to be working on ADAS sensors with. HL is one of the bigger 24GHZ radar sensor suppliers and are actively trying to move into the 77GHz radar sensors which are used for all of the Air radar modules. With respect to Here maps, I doubt that Lucid will move away from Here maps.

I agree about HERE. Last I read about why automakers keep choosing HERE had nothing to do with user facing data but more what the car needs for ADAS\Autonomous driving. HERE maps apparently has better HD mapping for this need over Google etc. so while everyone is kicking and screaming to have Google Maps its not beneficial to what the car needs behind the scenes. Seems like a trade off and most chose the behind the scenes need as more important. Maybe LA area is mapped out better but I don't find HERE to be that bad. Every time its done a weird route its done it for a valid reason (traffic, accident, etc.) and when i've crossed checked its weird routing with Apple Maps and Google they're often saying to take the same route to a point where I now don't question it.

 

[Adnillien]

The problem is that hardware outdates very quickly, it doesn't become obsolete but it goes through upgrade cycles. Is the LiDAR sensor fitted in the car today still relevant to the LiDAR sensors that are available today, are the cameras still relevant to what cameras are available today, etc. Lucid dragging its feet on the software portion of DDPro will only make this gap worse.

The Airs LiDAR resolution and field of view is still very much state of the art. 150m range, 120 degrees horizontal field of view and 25 degrees vertical with 0.05 degrees of angular resolution. I think the bigger issue for becoming obsolete is the NVIDIA compute power. Lucid is using a newer sensor fusion strategy is referred to early fusion compared to the legacy automakers. In early fusion, all of the sensor information is combined before target identification. The advantage is that all sensor information, rather than filtered targets, is available for the perception algorithm. The disadvantage is that it requires a very high power central processor for the fusion. Since it is a newer strategy, the perception algorithms are not as well developed and hence the delay in rolling out new Dream Drive features. It is also the reason that the NVIDIA processing power is key to making Dream Drive Improvements.

Legacy automakers tend to use late fusion where the target recognition is done separately for each sensor (camera, radar, LiDAR) and the resulting targets are overlayed together for the final perception algorithm. This strategy distributes the processing power across many processors and reduces the workload on the central processor. It also eliminates and simplifies a lot of sensor data before being passed to the central processor for combining.

 

[Drendino]

The Airs LiDAR resolution and field of view is still very much state of the art. 150m range, 120 degrees horizontal field of view and 25 degrees vertical with 0.05 degrees of angular resolution. I think the bigger issue for becoming obsolete is the NVIDIA compute power. Lucid is using a newer sensor fusion strategy is referred to early fusion compared to the legacy automakers. In early fusion, all of the sensor information is combined before target identification. The advantage is that all sensor information, rather than filtered targets, is available for the perception algorithm. The disadvantage is that it requires a very high power central processor for the fusion. Since it is a newer strategy, the perception algorithms are not as well developed and hence the delay in rolling out new Dream Drive features. It is also the reason that the NVIDIA processing power is key to making Dream Drive Improvements.

Legacy automakers tend to use late fusion where the target recognition is done separately for each sensor (camera, radar, LiDAR) and the resulting targets are overlayed together for the final perception algorithm. This strategy distributes the processing power across many processors and reduces the workload on the central processor. It also eliminates and simplifies a lot of sensor data before being passed to the central processor for combining.

Is it your opinion that our cars do not have an adequate process server? If so, how can Lucid remedy that on existing models? I assume Gravity will have much more capacity.

 

[Adnillien]

Is it your opinion that our cars do not have an adequate process server? If so, how can Lucid remedy that on existing models? I assume Gravity will have much more capacity.

I have no opinion on the adequacy of the Air's NVIDIA central processor because we do not have enough information about what is needed or the current capability. It is only my opinion that the central compute power will become obsolete before the sensors. We do know that Gravity will have significantly better central compute power than the Air.