A Tesla is more like the coyote than the roadrunner

https://cdn.steemitimages.com/DQmNhJAhwqRf9sVmav2krfetBhtd6W7KTG55vsy5mpEdPPu/imagen.png
Souce

Do you remember those scenes from the Roadrunner cartoons, in which a road painted on a wall tricks the Coyote into crashing head-on, now he replaces the character with a Tesla equipped with one of the most famous autopilot systems in the world, that was exactly what happened and an impressive test carried out by YouTuber and former NASA engineer Mark Rover.

https://www.youtube.com/watch?v=IQJL3htsDyQ

Could it be that Elon Musk was wrong when criticizing LIDAR technology? In the video, Mark Rover tested a White Tesla with the autopilot system that uses only cameras against another vehicle equipped with Lidar sensors, known for using lasers to identify objects with extreme precision. One of the most striking tests was with a false wall painted to look exactly like a road, the sky and the surrounding landscape, something that would hardly happen in real life, but which perfectly illustrates the limits of vision-based systems.

https://cdn.steemitimages.com/DQmZh54pSWoKYyvntuPBeaigUj5uTdriSwHESsUfFhHLCf2/imagen.png
Souce

The result was dramatic and caught everyone's attention. The vehicle with Lidar was not fooled for a moment by the paint on the wall. This is because lidar works by emitting pulses of laser light and precisely measuring the distance of objects. It does not matter what has been painted on the wall. The system immediately identifies the barrier as something physical and brakes without hesitation. However, the Tesla system, which is based exclusively on cameras and computer vision, was completely fooled by the visual illusion, crossing the false wall without even touch the brakes.

Elon Musk, who is against lidar in his Teslas, has stated on multiple occasions that he considers lidar an expensive, unnecessary and Obsolete, Musk believes that the future of autonomous driving depends exclusively on computer vision based on cameras supported by artificial intelligence and neural networks. According to him, human vision is sufficient to drive, so replicating that system through cameras would be the most efficient way to achieve true level five autonomy in which the car would not need human intervention in any situation.

The false wall test revealed the great weakness of camera-based systems, they depend on visual interpretations and can be easily fooled by optical illusions or challenging environmental situations, such as dense fog or heavy rain, now based on this test do you really believe that Musk is right or will it be that the future of autonomous cars depends precisely on the combination of different sensors to guarantee maximum safety.

The images without reference were created with AI Le immagini senza riferimento sono state create con l'intelligenza artificiale

Thank you for visiting my blog. If you like posts about #science, #planet, #politics, #rights #crypto, #traveling and discovering secrets and beauties of the #universe, feel free to Follow me as these are the topics I write about the most. Have a wonderful day and stay on this great platform :) :)

! The truth will set us free and science is the one that is closest to the truth!

security

Experts