This site is reader-supported. When you click through links on our site, we may be compensated.
Tesla’s focus on cameras and sensors has been evident since the beginning of its self-driving journey. With the company having several options to utilize in its quest for autonomy, Elon Musk has always maintained that Tesla would utilize cameras and radar, not LiDAR, for its self-driving projects. Musk once called LiDAR “a fool’s errand” and refused ever to consider the system for Tesla.
Instead, Tesla has used eight external body cameras, twelve ultrasonic sensors, and a front-facing radar to improve the accuracy of its self-driving cars. To make the cars even more predictable and safe, Tesla also uses a Neural Network to gather data and information that learns other drivers’ behavior. The 20 billion miles of real-world data Tesla has gathered over the years contributes to the Neural Net’s development and makes the entire Autopilot and FSD system more accurate as more miles are driven.
With all of these sensors, cameras, and radars, most would think that Tesla’s FSD suite, which is already showing vast improvements since the FSD Beta rollout in late 2020, just needs more miles and time to improve. But this isn’t the case, because Tesla wants to install a more refined radar sensor in its cars to improve accuracy even further.
Originally obtained and reported by Electrek, the FCC document doesn’t reveal full details on the device because of a confidentiality agreement. This will keep the finer points of the system until July 2021.
The document says that the new radar will operate in a 60 GHz band.
“The equipment under test (EUT) was a Vehicle Millimeter-wave Radar Sensor operating in 60 GHz band (60-64 GHz).”
Ultimately, using a new, more refined, and more accurate sensor will help Tesla develop its self-driving suite even further. Last year, Musk revealed during the Q3 2020 Earnings Call that the company planned to move to a 4-dimensional training program, moving away from its “~2.5D” system that it currently uses. 4D is “essentially video” and would help improve the system’s accuracy.
“So what we’ve been doing, thus far, has really just been like 2D — mostly 2D, and like I said, well correlated in time. So just hard to convey just how much better a fully 4D system would work — does work. It’s capable of things that if you just look — looking at things as individual pictures as opposed to video — basically, like you could go from like individual pictures to surround video, so it’s fundamental. So the car will seem to have just like a giant improvement.”
The 4D system would be capable of handling different traffic scenarios in a more sophisticated and accurate fashion because video is easier to gain information from than pictures. It will likely work in combination with Tesla’s Dojo supercomputer, which is set to be released late this year.