LiDAR - NASA confirms cameras better than LiDAR - Musk is correct

[qdeathstar]

Let's look at the past and present:

Without LIDAR, no one could reach the finish line for the first DARPA challenge in 2004.

By the 2007 DARPA Urban Challenge (Youtube), there were about 36 participants but only 6 were able to finish the race and they all had LIDAR.


1) Carnegie Mellon University, Tartan

2) Stanford, Junior

3) Virginia Tech, VictorTango

4) MIT, Talos:

Finished the race among the 6: University of Pennsylvania, Little Ben

Finished the race among the 6: Team Cornell’s Skynet

View attachment 825396

Waymo has been able to use LIDAR to first let the blind man ride in its Prius with no human driver in 2012:

Waymo has been letting its driverless cars drive the public (not just NDA riders) around since 2020.

Companies that don't have LIDAR can't do that today.

With the advance of Tesla Vision, it still collides with stationary objects, and AI Addict were expelled from FSD beta as an example.

In my opinion: Collision Avoidance Technology: LIDAR has been proven since the 2007 DARPA Urban Challenge. In the meantime, Tesla Vision has been proven to be unreliable for Collision Avoidance Technology in the past and present because it still requires a human driver to blame.

What needs now in the LIDAR-equipped system is intelligence.

I disagree. It doesn’t prove anything. If you equip a car with millions of dollars worth of sensors it will be able to drive itself. But that car isn’t practical to sell to the masses. So it’s not a practical answer to level 5 self driving.

Technology no one can afford is useless.

 

[Daniel in SD]

I disagree. It doesn’t prove anything. If you equip a car with millions of dollars worth of sensors it will be able to drive itself. But that car isn’t practical to sell to the masses. So it’s not a practical answer to level 5 self driving.

Technology no one can afford is useless.

The cost of all that 2007 sensor technology is absolutely low enough to sell to the masses in 2022. The actual issue is that those vehicles cannot drive themselves at anywhere close to human performance.

 

[diplomat33]

I disagree. It doesn’t prove anything. If you equip a car with millions of dollars worth of sensors it will be able to drive itself. But that car isn’t practical to sell to the masses. So it’s not a practical answer to level 5 self driving.

Technology no one can afford is useless.

That's hyperbole. AVs today can do L4 and don't have millions of dollars worth of sensors. Mobileye is planning to do L4 consumer cars with just 1 lidar, 12 cameras and 6 radar, definitely not "millions of dollars worth of sensors". In fact, Mobileye is planning to sell L4 on consumer cars for less than $6,000. That would definitely be affordable.

 

[Tam]

...If you equip a car with millions of dollars worth of sensors it will be able to drive itself...

That's the thing: In the old days, prior to LIDAR, no matter how many millions of dollars researchers could spend, they could not have a reliable Collision Avoidance Technology.

15 years ago, or by the 2007 DARPA Challenge, the LIDAR's price went down to below $100,000 each, and that helped 6 cars to the finish line when it was not possible before.

Humans have 2 eyes and can drive fine. Tesla has more: 8 cameras but the difference is: It is lacking a capable brain. It's possible to have Tesla Vision to avoid obstacles but it needs a better brain or computer and source codes which we don't have today. To imitate a human brain would take millions or billions of dollars and many many years. We'll be lucky if it's in our lifetime.

Because we don't have billions of dollars each to get a computer that has human brain capacity, we have the cheap Tesla HW3, and soon HW4. But they are by no way will be as capable as the human brain so it will continue to collide with obstacles.

Thus, Tesla Vision is not cheap or here. It's like the old days when there was no LIDAR. Why? Again, it's not just Vision but it's also the "brain". A computer that can competently process like a driver's brain is not here.

That's why LIDAR is now cheaper. Its computer does not need to be a driver's brain. It is just a machine that can do repetitive tasks well with no thinking. We only tell it the repetitive tasks to perform: It can report the thickness of the paint on the lane marker and you just need to write a program that is not dangerous. It can report the size of a sandbag on the road and you can write a program to avoid the sandbag. Again, there's no need for a powerful human driver's brain: you only need a machine that can do repetitive tasks. That means you don't need a highly capable computer like a driver's brain.

 

[qdeathstar]

That's hyperbole. AVs today can do L4 and don't have millions of dollars worth of sensors. Mobileye is planning to do L4 consumer cars with just 1 lidar, 12 cameras and 6 radar, definitely not "millions of dollars worth of sensors". In fact, Mobileye is planning to sell L4 on consumer cars for less than $6,000. That would definitely be affordable.

As I said, plans are plans. Once level 5 is a solved problem, we can make bold statements.

 

[qdeathstar]

That's the thing: In the old days, prior to LIDAR, no matter how many millions of dollars researchers could spend, they could not have a reliable Collision Avoidance Technology.

15 years ago, or by the 2007 DARPA Challenge, the LIDAR's price went down to below $100,000 each, and that helped 6 cars to the finish line when it was not possible before.

Humans have 2 eyes and can drive fine. Tesla has more: 8 cameras but the difference is: It is lacking a capable brain. It's possible to have Tesla Vision to avoid obstacles but it needs a better brain or computer and source codes which we don't have today. To imitate a human brain would take millions or billions of dollars and many many years. We'll be lucky if it's in our lifetime.

Because we don't have billions of dollars each to get a computer that has human brain capacity, we have the cheap Tesla HW3, and soon HW4. But they are by no way will be as capable as the human brain so it will continue to collide with obstacles.

Thus, Tesla Vision is not cheap or here. It's like the old days when there was no LIDAR. Why? Again, it's not just Vision but it's also the "brain". A computer that can competently process like a driver's brain is not here.

That's why LIDAR is now cheaper. Its computer does not need to be a driver's brain. It is just a machine that can do repetitive tasks well with no thinking. We only tell it the repetitive tasks to perform: It can report the thickness of the paint on the lane marker and you just need to write a program that is not dangerous. It can report the size of a sandbag on the road and you can write a program to avoid the sandbag. Again, there's no need for a powerful human driver's brain: you only need a machine that can do repetitive tasks. That means you don't need a highly capable computer like a driver's brain.

We don’t have any car capable of autonomous driving on any road. Waymo does it on set routes on certain roads.

 

[qdeathstar]

The cost of all that 2007 sensor technology is absolutely low enough to sell to the masses in 2022. The actual issue is that those vehicles cannot drive themselves at anywhere close to human performance.

That wasn’t my point. My point is you could have a lot of sensors but if the price point is too high who cares.

 

[qdeathstar]

Also, @!igs im not sure if you got the memo or not but now disagree adds to my rep just like a like does, so thank you for the increase in my reaction/post ratio brah…

 

[Tam]

We don’t have any car capable of autonomous driving on any road. Waymo does it on set routes on certain roads.

That's why it's categorized from the low 0 to the highest 5.

L3 is a step toward improvement when there's an accident while the L3 is in operation, the driver is not liable. Usually at low-speed traffic jams on highways.

Mercedes-Benz Takes Legal Responsibility for Its Level 3 Technology

While the announcement on liability acceptance communicates Mercedes-Benz’s confidence in the future of SAE Level 3 technology, the industry as a whole needs to do much more to overcome early consumer reluctance and concerns about next-level vehicle automation.

www.wardsauto.com

Mercedes opens sales of Level 3 self-driving system on S-Class, EQS

Mercedes will offer the conditional self-driving system, the first to be approved for European public roads, as an option starting at $5,300.

europe.autonews.com

L4 is the same responsibility as above: While the car is in L4, a blind man rider can't be blamed for not having a driver's license. It is currently geofence, so that's why it is not L5.

L4 Waymo and Cruise are taking rides and getting paid for it with no drivers in their cars. The downside is it is L4: Geofenced and not everywhere.

For consumer products, an L2 with reliable Collision Avoidance Technology is a happy medium between L2 and L5: It's a dumb machine that can repetitively avoid obstacles but it lacks intelligence and that's where the driver's brain is for: Not to repetitively doing the collision avoidance tasks but to be reserved for intelligence decision: The car might get stuck and doesn't know what to do next, the human driver can then manually make the decision to bail the machine out.

That's where the GM Ultra Cruise, an L2, with LIDAR will fit in approximately in 2023:

 

[Daniel in SD]

That wasn’t my point. My point is you could have a lot of sensors but if the price point is too high who cares.

The price point of the sensors isn’t too high though. I think you vastly underestimate the value of a self-driving car.
People seem to think theses sensors are fundamentally expensive just like people used to think batteries were always going to be too expensive. Volume and technology advances have made them cheaper over time.