First Jerry rig everything, now mark rober 🧐

[IIIEnforcerIII]

I had to re-read your comment multiple times, and not sure if I got your point but I suspect you're thinking of it as a "this vehicle performs better than that vehicle", but that wasn't the point of the video. The point was a demonstration of the kind of tricks LIDARs can do, which in parallel highlighted some of the limitations of Tesla's safety features that at this time only rely on the cameras.

People keep bringing up FSD in this thread, but the only thing FSD could have done there was to kick out forcing the driver to take over controls. Cameras were simply unable to see into the wall of fog or into the wall of water, it doesn't matter what software those camera feeds were pumped into, AP or FSD.

I agree. I think most people are being way too sensitive. I didn't notice the brand of car with Lidar. I didn't care about the product Luminar, etc. Any thinking person should be asking them same questions, is Vision enough to do X thing. I live in a very snowy place. I wish liar was included because it is better in some scenarios. I wonder all the time if my cameras are seeing much, or how they react with the glare of the sun, rain, snow, fog. I simply thought it was interesting to see the difference. I'm only on the trial FSD and won't ever buy it, so it doesn't matter to me really. But it did answer my question about which each technology can do. Was it a perfect comparison? Probably not. But I'm simply not losing any sleep over it.

Sponsored

 

[freyguy]

Seen another post on x that adds more fuel to this mess. apparently the phone was edited post production. the phone in the lidar vehicle that was mounted was an apple phone as you can clearly see by the reflection. but in the shot it has a G brandling but its not the correct orientation. I didn't wanna believe it but sadly I went back gave mark sober another view and here yet again another flop of mark hiding using an apple product and trying to hide it. Is there no integrity?

 

[shaneaus]

TLR - Lidar and Radar signals have diminishing results the worse the weather. The processing time needed to interpret the second set of data (radar/lidar) and combine it with the vision data takes too long. By the time the data can be used the data is too old and no longer relevant.

---
The issue with Lidar and with radar are with particles in the air.

Lidar = light emissions being reflected. When there is inclement weather there are particles (rain/moistures drops, snow, dust in the event of a dust storm). All of these particles reflect and/or diffuse the light creating a vast amount of "noise." The heavier the rain or snow the less effective Lidar functions.

This same principle applies to radar. This is why the manufacturers of speed detection radar units don't allow law enforcement to use the radar in raining/snowing conditions. The Officers COULD use the radar to measure speed -but, it takes experience to know that the range becomes much reduced and usage at a specific location/conditions to understand what signal is being returned. So, the manufacturers just state in their training material that they don't approve the devices for use in inclement weather.

Radar and lidar devices all work on the same principles (a signal is emitted and the return is interpreted). If one is driving in heavy rain or snow radar/lidar won't be much (if any) more effective than vision.

In Tesla's case the software interpretation of that signal data from radar or lidar takes time. Then, that interpretation must be compared to the interpretation of the video signal. And, that comparison takes time. This is what Tesla was referring to when they talked about latency.

By the time the lidar signal is received, interpreted, compared to the vision, and the software can use the data... Even though it is only milliseconds - It is too late to use the combined data. The software is already trying to calculate the next set of data. Over a shirt period of time the lidar processing just gets farther behind or is discarded as no longer useful and the process starts over again with the same result. By the time the data can be used it is too old to be relevant.

This is why Tesla is staying with vision. One data set that can be used for everything (signage, lines in the roadway, detecting vehicles, obstacles,etc) and that data doesn't need to be combined with anything before usage.

 

[cybercricket]

TLR - Lidar and Radar signals have diminishing results the worse the weather. The processing time needed to interpret the second set of data (radar/lidar) and combine it with the vision data takes too long. By the time the data can be used the data is too old and no longer relevant.

---
The issue with Lidar and with radar are with particles in the air.

Lidar = light emissions being reflected. When there is inclement weather there are particles (rain/moistures drops, snow, dust in the event of a dust storm). All of these particles reflect and/or diffuse the light creating a vast amount of "noise." The heavier the rain or snow the less effective Lidar functions.

This same principle applies to radar. This is why the manufacturers of speed detection radar units don't allow law enforcement to use the radar in raining/snowing conditions. The Officers COULD use the radar to measure speed -but, it takes experience to know that the range becomes much reduced and usage at a specific location/conditions to understand what signal is being returned. So, the manufacturers just state in their training material that they don't approve the devices for use in inclement weather.

Radar and lidar devices all work on the same principles (a signal is emitted and the return is interpreted). If one is driving in heavy rain or snow radar/lidar won't be much (if any) more effective than vision.

In Tesla's case the software interpretation of that signal data from radar or lidar takes time. Then, that interpretation must be compared to the interpretation of the video signal. And, that comparison takes time. This is what Tesla was referring to when they talked about latency.

By the time the lidar signal is received, interpreted, compared to the vision, and the software can use the data... Even though it is only milliseconds - It is too late to use the combined data. The software is already trying to calculate the next set of data. Over a shirt period of time the lidar processing just gets farther behind or is discarded as no longer useful and the process starts over again with the same result. By the time the data can be used it is too old to be relevant.

This is why Tesla is staying with vision. One data set that can be used for everything (signage, lines in the roadway, detecting vehicles, obstacles,etc) and that data doesn't need to be combined with anything before usage.

You're making it significantly more complicated than it needs to be. LIDAR/RADAR can be used as a sidecar for AEB input, it doesn't need to be integrated into any of the autonomy features.

 

[gooshjkc]

Even though I own a CyberTruck I can be unbiased here. These YouTubers on the other hand are not. They do it for one reason and that’s all - to get clicks. From there make money and get more sponsors. Trust me, if there were no $$$ behind these things, first they wouldn’t do it and most likely can careless. CyberTrucks are not flawless. Far from it. However, on two of these examples, Tesla has done something remarkable. Lidar might be an advance tech, but I don’t think it can replace the human eye. On which Tesla Vision is getting closer and closer to doing. Lidar has a lot of potential and I can see it being adapted to a lot of things, but for the road not so much. Now, coming to tow. Yes aluminum is not stronger than steel, but Tesla gave the specs of its strength and Jerry proved it plus some. I would say that’s a win not a loss. Now, I would tell anyone if they wish to buy a truck to tow. Any EV truck is NOT the way to go. Sorry, ICE trucks still have the leg up on this category. Maybe short distances like 50 miles and anything under 10K pounds go for it, but anything above those metrics - nope.

 

[BannedByTMC]

What is test load? Some hitch standard in testing?

I am just saying it broke at ~10x of Tesla max tongue weight.

Exactly and the cybertruck exceeded this by almost 10x.

or he didn’t understand what he was testing or how he was testing it.

Y'all need to read and understand the following:

It is designed to tow an 11k trailer with 1,100 pound tongue weight which means the receiver is designed to withstand much more than 1,100 pounds. In other words, if it broke at 3,300 pounds, that would be a fail, not a "great! it handled 3x what it needs to".

Static load rating is not the same as maximum force it's designed to handle. Loads seen during towing will be potentially much higher than 1,100lbs. Yes Jerry took it beyond that but it did not do "10X" anything

 

[shaneaus]

You're making it significantly more complicated than it needs to be. LIDAR/RADAR can be used as a sidecar for AEB input, it doesn't need to be integrated into any of the autonomy features.

Well, I'm just repeating what Tesla/Elon has already stated. And, at least for me, if the Lidar/Radar isn't being used with the self driving software then there would be absolutely no point in having it. It would just be a complete waste of money. Without self driving I'd prefer to save money and request the vehicle without radar/lidar sensors.

 

[CTInProcess]

I unsubscribed soon as I made my way through this hit piece.

It's funny how these are coming with more frequency after DOGE started dusting off old logbooks in high places.

Same here. I’m not sure why people are so unhappy that Elon is getting rid of all investigations into Tesla and companies. People should be happy as this will allow more Beta Testing without red tape ….

Everything is perfect, I’m glad they are nitpicking phones, maybe they should also start looking at shoes and time. He’s a fraud!!!

 

[cybercricket]

Well, I'm just repeating what Tesla/Elon has already stated. And, at least for me, if the Lidar/Radar isn't being used with the self driving software then there would be absolutely no point in having it. It would just be a complete waste of money. Without self driving I'd prefer to save money and request the vehicle without radar/lidar sensors.

Let's also throw out AEB since it doesn't really work in the conditions where it's needed the most. The only complication is that starting in 2029 it's mandatory.

 

[shaneaus]

Let's also throw out AEB since it doesn't really work in the conditions where it's needed the most. The only complication is that starting in 2029 it's mandatory.

Vision can work just as well for aeb as by the time the other systems "see" a vehicle you can probably see it with your own eyeballs.

Regardless, it sounds like you and I won't agree. And, I'd prefer not to hijack this thread seeing as what we are discussing isn't related to the OP's topic.

I, mostly, was just responding to a previous comment in which someone didn't understand why Tesla did away with Lidar/Radar. But, I'm on mobile and forgot to "reply" to the other person.

 

[cybercricket]

Vision can work just as well for aeb as by the time the other systems "see" a vehicle you can probably see it with your own eyeballs.

That's what the demo in the video demonstrated - a LIDAR can punch much farther into the fog / rain than cameras everything else being equal. Tesla AEB failed in those tests.

Regardless, it sounds like you and I won't agree. And, I'd prefer not to hijack this thread seeing as what we are discussing isn't related to the OP's topic.

It's absolutely related to the topic.

 

[mongo]

That's what the demo in the video demonstrated - a LIDAR can punch much farther into the fog / rain than cameras everything else being equal. Tesla AEB failed in those tests.

Wait are we talking about AEB or AP/FSD? If the latter, how did Mark get AP to drive down the center line?

 

[cybercricket]

Wait are we talking about AEB or AP/FSD? If the latter, how did Mark get AP to drive down the center line?

There was a combination of AEB/AP across all tests, but from the product standpoint AEB is the lowest common denominator and the vehicle should have engaged the brakes regardless of AP use. He did also use AP visualization to demonstrate how cameras couldn't see into the water. Not sure if I've answered your question, but again I'm highlighting that the context isn't autonomy, but rather "ranging" using vision only.

 

[dalton108]

I never heard of Mark Roper before this, but he may be getting some instant karma.

 

[nevetsyad]

Vision alone can only go so far. Aka in ideal conditions. The reason why us humans slow down or don’t drive in certain conditions, ie blizzards or sand storms.

I can’t wrap my brain around why a manufacturer would choose only one sensing technology instead of a network of different types for different weather/road conditions.

Fact: I’m not an engineer.

Are you suggesting not slowing down in wet, icy, vision restricted conditions? As an FYI, lasers can't see through snow, and worse, droplets from real fog, snow or rain make the laser sensors crap out, VW, BYD, Mercedes, etc. default back to cameras and radar (if radar is still working) in those conditions.

Sponsored