[Service Development Team Byungin Kim]
Tesla recently decided to release radar from Model 3 and Model Y sold in North America. Since LiDAR has already been excluded, we have declared that autonomous driving will be implemented only with the camera and AI technology on the car.
Up to now, both sensors of lidar and radar have been recognized as indispensable for autonomous driving, and most autonomous vehicles have made efforts to realize safe autonomous driving by properly using radar, lidar, and camera together.
Below is a brief description of the radar and lidar.
Radar
Radar is an electromagnetic wave-based sensor. electromagnetic waveIt is possible to measure the distance to the object and measure the direction of the object by analyzing the electromagnetic wave that hits the object, reflects and returns, and is less affected by the surrounding environment.
LiDAR
LiDAR is a system that measures the location coordinates and distance of a reflector by measuring the time it takes to shoot a laser pulse and return it to be reflected. Because it is measured by shooting light directly, it has little effect on darkness and can accurately recognize surrounding shapes.
Tesla Vision
At AI Day a few days ago, there was a structural explanation of Tesla Vision powered by 8 cameras and neural network processing without a radar sensor.
The first picture shows eight cameras mounted on the vehicle taking 360-degree photos and converting them into 3D vector space on a monitor installed next to the driver's seat.
A person in charge of autonomous driving explained that Tesla's AI works just as Tesla's neural network technology recognizes and judges the environment and situation of a snow road. Currently, Tesla has developed a neural network system with increased precision while increasing the vector space, which is a weakness of the existing FSD, and simplifying all tasks such as camera calibration and caching. “The neural network has proven its technological prowess by showing that it can effectively draw a map in real time while driving,” said Carpati, confidently saying that it shows a big difference from General Motors’ driving assistance systems 'Super Cruise' and Waymo. These competitors explained that unlike Tesla, they were sticking with a pre-mapping strategy.
As Tesla announced at AI day, the results of actual driving are expected in the future whether the camera alone will show performance that exceeds the existing autonomous driving.
