Lucid missing market segment begging to be exploited via Dream Drive

[BS8899]

I am not a LIDAR expert by any stretch of the imagination. From what I know, I question whether a passive (non-scanning) LIDAR can offer any useful extension to the cameras.

someone who is a lot smarter than I am can educate us.

Some older cars had "Ultrasound" sensors for ranging. Perhaps the (non-scanning) LIDAR on the Lucid can do "ranging". But that's a country-mile differences than 3D scanning, as in the Waymo!

 

[RM-S8]

I'm in the same boat as you with DreamDrive Pro. It irks me to no end that Lucid made all these promises, showed off marketing material that never came to fruition and won't say a god damn word about what's going on and if\when we can expect ANY features to be added to DreamDrive Pro.

These are not promises, they are rather future improvements that Lucid is working on .
Jokes aside, I hardly use DDP, so take my comment very lightly.

 

[BS8899]

These are not promises, they are rather future improvements that Lucid is working on .
Jokes aside, I hardly use DDP, so take my comment very lightly.

I don't think we should side-step the real issues. The currently enabled DDPro capabilities (lane-keep HA, 3D view, etc.) are roughly the same as what I had on my 2017 Lexus RX450. Implied or otherwise, people who have DDPro paid around $10,000 or so for it, correct? LIDAR was prominently mentioned/advertised as a differentiating capability.

Yet, nothing came of it. Worse yet, as mentioned by many, there is no roadmap, schedule, or even acknowledgment of what lies ahead. I don't think that builds customer commitment!

Even a forward-looking-only LIDAR (as opposed to a scanning (as in a Waymo) or phased-array LIDAR) if properly implemented, can be a differentiator. But, where is the roadmap?!

 

[Batteryman]

Waymo is truly amazing. I live in Tempe, where interacting with these vehicles happens every day, and I have come to the point of actually seeking them out when I have a choice. Get behind one at a traffic light. When the light turns green, it moves right away. What? Not finishing that text? On the road, it travels dead center in the lane. What? Not looking elsewhere or reading an email and drifting all over the place? It's fast, dead nuts accurate, signals turns, doesn't run lights, and is reliable as hell. Now, throw it a curve, like a non functioning left turn lane sensor, and it will sit there ad nauseum, but I can live with that. It won't shoot me in a fit of rage either.

I think the world will be a better place, and accident/insurance rates should drop dramatically, when many drivers have access to at least some of this technology. I hope that, rather than having so many auto makers reinvent the wheel, something like Waymo technology will be easily accessible (and adaptable) for use by companies like Lucid.

Oops, I just remembered that Waymo is owned by Google, like Android Auto. Never mind.

Oh wow. The light changed about ten minutes ago. I better move. Thank God everybody in Arizona is afraid to even honk at laggards like me!

I always wonder how Tesla can achieve what Waymo does with just cameras. Anyway I look at it, it just won’t work to the degree of safety that should be acceptable.

 

[Bobby]

I don't think we should side-step the real issues. The currently enabled DDPro capabilities (lane-keep HA, 3D view, etc.) are roughly the same as what I had on my 2017 Lexus RX450. Implied or otherwise, people who have DDPro paid around $10,000 or so for it, correct? LIDAR was prominently mentioned/advertised as a differentiating capability.

Yet, nothing came of it. Worse yet, as mentioned by many, there is no roadmap, schedule, or even acknowledgment of what lies ahead. I don't think that builds customer commitment!

Even a forward-looking-only LIDAR (as opposed to a scanning (as in a Waymo) or phased-array LIDAR) if properly implemented, can be a differentiator. But, where is the roadmap?!

We have gone back-and-forth on this for a very long time. I agree with you, we got less than we paid for. I have a feeling we will see some big upgrades this year.

 

[segbrk]

I am not a LIDAR expert by any stretch of the imagination. From what I know, I question whether a passive (non-scanning) LIDAR can offer any useful extension to the cameras.

someone who is a lot smarter than I am can educate us.

Why do you think it's passive?

Lucid's LiDAR supplier

I was finally able to dig up who is supplying Lucid LiDAR modules. It is Robosense which is using a 905nm technology. This appears to be the most common for ADAS. Robosense is China based and also supplying Xpeng, Zeeker and GAC. The lasers are EEL (Edge Emitting Lasers) made by OSRAM and the...

lucidowners.com

 

[segbrk]

Some older cars had "Ultrasound" sensors for ranging. Perhaps the (non-scanning) LIDAR on the Lucid can do "ranging". But that's a country-mile differences than 3D scanning, as in the Waymo!

Lucid also has ultrasound sensors. Those are the parking sensors that give you high precision at low range.

 

[segbrk]

Why do you think it's passive?

Lucid's LiDAR supplier

I was finally able to dig up who is supplying Lucid LiDAR modules. It is Robosense which is using a 905nm technology. This appears to be the most common for ADAS. Robosense is China based and also supplying Xpeng, Zeeker and GAC. The lasers are EEL (Edge Emitting Lasers) made by OSRAM and the...

lucidowners.com

Adding some more clarification. I'm not a LIDAR expert by any means either, but maybe have slightly more exposure to it.

"Passive" lidar is not a thing people commonly talk about. It means measuring sunlight and other natural light sources as they reflect off of objects, rather than the "active" type actually projecting a light source from the sensor rig and measuring the time before its reflection is received. Some things do use the passive type, like satellites, but generally when people say LIDAR they mean the type with lasers in it. Lucid's is the type with lasers.

Simple LIDAR, like the type you might put on a hobby robot or something, has a single beam and single sensor. It only gives you ranging, like you said. Not that different from ultrasonic, it's just that light works better than sound for some applications. Lucid's LIDAR is not simple.

Waymo has a big 360º LIDAR hat on top of its head. That's a whole load of lasers and sensors pointing in all directions. That's obviously great for building a complete picture of the surroundings.

Lucid's is in between. It's not just one direction, it's a 120º wide forward facing view. That won't give it a complete picture of surroundings, but it's enough to map out in much higher 3D detail than cameras all the stuff you're about to drive into. It also has a much higher range and speed than ultrasonics. I suspect the LIDAR is what we can already thank for the Air's relatively smooth tracking of the vehicle in front for ACC/HA. I think it will also be an advantage for automatic lane changing eventually, and for tracking non-car objects in front of you (pedestrians, bikes, etc.) where non-stereoscopic cameras might be easily fooled by e.g. a person wearing clothes that blend into the background behind them. Not saying Tesla's system isn't more advanced than my home security camera "AI," but... my cameras will tell me they see a small car instead of a regular sized person if I wear a shirt with a car on it. That type of mistake is pretty much impossible to fully correct for without some kind of 3D vision.

 

[Adnillien]

Adding some more clarification. I'm not a LIDAR expert by any means either, but maybe have slightly more exposure to it.

"Passive" lidar is not a thing people commonly talk about. It means measuring sunlight and other natural light sources as they reflect off of objects, rather than the "active" type actually projecting a light source from the sensor rig and measuring the time before its reflection is received. Some things do use the passive type, like satellites, but generally when people say LIDAR they mean the type with lasers in it. Lucid's is the type with lasers.

Simple LIDAR, like the type you might put on a hobby robot or something, has a single beam and single sensor. It only gives you ranging, like you said. Not that different from ultrasonic, it's just that light works better than sound for some applications. Lucid's LIDAR is not simple.

Waymo has a big 360º LIDAR hat on top of its head. That's a whole load of lasers and sensors pointing in all directions. That's obviously great for building a complete picture of the surroundings.

Lucid's is in between. It's not just one direction, it's a 120º wide forward facing view. That won't give it a complete picture of surroundings, but it's enough to map out in much higher 3D detail than cameras all the stuff you're about to drive into. It also has a much higher range and speed than ultrasonics. I suspect the LIDAR is what we can already thank for the Air's relatively smooth tracking of the vehicle in front for ACC/HA. I think it will also be an advantage for automatic lane changing eventually, and for tracking non-car objects in front of you (pedestrians, bikes, etc.) where non-stereoscopic cameras might be easily fooled by e.g. a person wearing clothes that blend into the background behind them. Not saying Tesla's system isn't more advanced than my home security camera "AI," but... my cameras will tell me they see a small car instead of a regular sized person if I wear a shirt with a car on it. That type of mistake is pretty much impossible to fully correct for without some kind of 3D vision.

Here are a few more details on the LiDAR in our Lucids.

• 120 degrees horizontal
• 25 degrees vertical
• Angular resolution .05 degrees
• At 100 meters, .05 degrees of resolution is about 8.7 cm

The resolution of LiDAR is the key advantage over Radar which even the best imaging Radar has 60 degree field of view and 15 degrees vertical with a resolution of 1 degree horizontal and 2 degrees vertical. At 100 meters 1 degree of resolution is 1.75 m.

The difference in resolution is what allows the ADAS system to separate targets. I agree that DDPro is not yet taking advantage of the capability but I think and I hope by the end of the year that we start to see it.

FOV 60 degrees Resolution 1 degrees

FOV 15 degrees Resolution 2 degrees

 

[Buffalo Bob]

Wow! Thanks all for the great information. I for one certainly have a lot better understanding of the technical capabilities built into our DDP equipped cars. We might not have all the spinning gizmos bolted all over those Waymo vehicles, but we are way ahead of the pack from a hardware standpoint, with the software to make it all worthwhile hopefully not too far behind.

 

[Buffalo Bob]

I always wonder how Tesla can achieve what Waymo does with just cameras. Anyway I look at it, it just won’t work to the degree of safety that should be acceptable.

I have no direct experience with Tesla FSD, but from what I've read here, they are still pretty far from achieving what Waymo does. However, given what little technology they have to work with, FSD seems to be a pretty remarkable achievement.

 

[BS8899]

Adding some more clarification. I'm not a LIDAR expert by any means either, but maybe have slightly more exposure to it.

"Passive" lidar is not a thing people commonly talk about. It means measuring sunlight and other natural light sources as they reflect off of objects, rather than the "active" type actually projecting a light source from the sensor rig and measuring the time before its reflection is received. Some things do use the passive type, like satellites, but generally when people say LIDAR they mean the type with lasers in it. Lucid's is the type with lasers.

Simple LIDAR, like the type you might put on a hobby robot or something, has a single beam and single sensor. It only gives you ranging, like you said. Not that different from ultrasonic, it's just that light works better than sound for some applications. Lucid's LIDAR is not simple.

Waymo has a big 360º LIDAR hat on top of its head. That's a whole load of lasers and sensors pointing in all directions. That's obviously great for building a complete picture of the surroundings.

Lucid's is in between. It's not just one direction, it's a 120º wide forward facing view. That won't give it a complete picture of surroundings, but it's enough to map out in much higher 3D detail than cameras all the stuff you're about to drive into. It also has a much higher range and speed than ultrasonics. I suspect the LIDAR is what we can already thank for the Air's relatively smooth tracking of the vehicle in front for ACC/HA. I think it will also be an advantage for automatic lane changing eventually, and for tracking non-car objects in front of you (pedestrians, bikes, etc.) where non-stereoscopic cameras might be easily fooled by e.g. a person wearing clothes that blend into the background behind them. Not saying Tesla's system isn't more advanced than my home security camera "AI," but... my cameras will tell me they see a small car instead of a regular sized person if I wear a shirt with a car on it. That type of mistake is pretty much impossible to fully correct for without some kind of 3D vision.

@segbrk's description above is, of course, correct. I was trying to draw a contrast between the Lucid LIDAR and the 3D-scanning LIDAR on the Waymo and I inadvertently used the term "passive" (referring to the non-3D scanning Lucid LIDAR) which, as @segbrk said, refers to a different thing. I should have said, "non-3D surround-view scanning" when comparing the Lucid LIDAR to the Waymo implementation. Unlike Waymo, which is driverless, and uses 3D LIDAR surround view scanning to detect/measure 3D surroundings, Lucid's LIDAR is a "looking forward" sensor. It is never intended to be "driverless" to free the driver to go to sleep or to play video games. Since there are plenty of discussions going back to 2022 on this forum, I won't repeat them.

As @Bobby pointed out, the delivery of ADAS LIDAR-enabled features has been a disappointment. The hope is, that major ADAS capabilities will be forthcoming. I hope that's true. But I worry software is not exactly Lucid's strong suit! Just as important, Lucid continues to tout DDPro as a high-value feature and charges accordingly. But there is little progress and no communicated roadmap and timeline. Disappointment and frustration are simply "expectations not met".

 

[Buffalo Bob]

My assumption has always been that the technology packed into (or onto) those powerhouse Waymo-type vehicles was never intended to be the norm, but rather just the means for continually building a more and more (dare I say) robust database of information that can someday be effectively applied to vehicles with more practical installations. That's not to say that the technology built into a DDPro Lucid won't be eclipsed over time, but when the time does come to transition all of that learning to everyday vehicles, it's hard to imagine that today's DDPro Lucid won't at least meet the minimum threshold, if not more likely set the bar as the best in class. Does anyone know what plans companies like Google have for bringing all of this knowledge into the mainstream? Perhaps there are already alliances built that I don't know about.

 

[gavram]

My hopes for Dream Drive have changed over time. Originally I thought I wanted more autonomy, not FSD but maybe auto lane change, lane centering on city streets, etc. but the cameras make it so easy to change lanes that I don't think I'd use it much and spending time in a Tesla has made me realize that the more the car is making its own decisions, the more consistently nervous I am that it's going to make a mistake.

At this point, my only real want is to be able to use lane centering on a highway with hands on the wheel but without getting the nag alert to torque the wheel. If Lucid had used a good touch sensitive, capacitive, wheel I'd probably be good with current set up but my hope would be better use of the in cabin camera or something to eliminate the torque requirement.

 

[hmp10]

My hopes for Dream Drive have changed over time. Originally I thought I wanted more autonomy, not FSD but maybe auto lane change, lane centering on city streets, etc. but the cameras make it so easy to change lanes that I don't think I'd use it much and spending time in a Tesla has made me realize that the more the car is making its own decisions, the more consistently nervous I am that it's going to make a mistake.

At this point, my only real want is to be able to use lane centering on a highway with hands on the wheel but without getting the nag alert to torque the wheel. If Lucid had used a good touch sensitive, capacitive, wheel I'd probably be good with current set up but my hope would be better use of the in cabin camera or something to eliminate the torque requirement.

Early on I read that Lucid's facial recognition camera would be used to insure driver attention when using ADAS instead of having to keep a hand on the wheel. I don't know why that never materialized, because the facial recognition camera in my car is quite sensitive (and aggressive) if my eyes so much as shift momentarily to read what music selection is playing -- even when wearing sunglasses.

 

[BobFM]

I'd love to at least get the functionality that I have in my wife's Suburban which has Super Cruise. Hands-free driving with auto-lane change from TX to FL is pretty fantastic over the 12+ hour trip. Even trips to Dallas and Houston would be much less painful. For shorter trips than that though, I prefer to drive this wonderful machine...

Agree with @Bobby that I also have a feeling there'll be more updates to the system. Just a feeling though...

 

[Sport]

Would like to thank all who kindly replied to my post. Have learned a lot including how as a senior to better drive the often construction-impaired TX Interstates. 4 additional points:

1) agree we have to give Lucid some slack. They are producing great debut cars as a startup and need time to deliver all their promises.

2) This said I will be scrutinising the new Volvo EX 90. It may be able to deliver safety features I lamented in my post are still not available as promised on Air GT. Pity company no longer Swedish owned. Would appreciate learning any views on how Gravity and EX 90 compare as details and test drives emerge.

3) Read recently (apologies source not retained) that in tuning the Sapphire, Lucid found ways to nicely enhance handling of all their cars. So they added new OTA tuning to Pure and said they will (eventually) do this for Air GT. Another case that we early adopters get 2nd class attention.

4) If I eventually sell my Air GT any suggestions how best to do this? Know am looking at hefty write off. Yet again a big disappointment. We naively hoped car would hold its value like a similarly priced Range Rover.

 

[segbrk]

3) Read recently (apologies source not retained) that in tuning the Sapphire, Lucid found ways to nicely enhance handling of all their cars. So they added new OTA tuning to Pure and said they will (eventually) do this for Air GT. Another case that we early adopters get 2nd class attention.

Pure actually inherited some newer components from the Sapphire. It’s likely not that Pure is getting attention first, but that it was just a straight copy of tuning for those components developed for the Sapphire. Doesn’t mean we’re left in the dust, it just means we have different components. Stuff like this is always going to happen with newer model years. Maybe similar changes will be made to our handling when they figure out how to apply it to our older components. Maybe the improvements don’t apply to us at all, and we were already better off with more mature tuning.

 

[geosynch]

....but would settle for a Homelink button I could press with the car in reverse.

Hear, hear! This is such an obvious eff up that really should be simple to fix. Makes me crazy ever time I back out of the driveway.

 

[Bus Driver]

Phantom braking, it also was driving straight for a train. You can’t achieve Robotaxis with just cameras. Good good example is a child on the side of the road. If there’s an adult holding the hand, you will not slow down, but if the child is playing with the ball or alone, you will. What do you do in bad weather with just cameras? Without redundancy, it’s unsafe and will make life threatening mistakes no matter how Musk spins it.

That’s not what EEEEHLOHN(!) thinks.