Biz / Auto

Tesla's FSD enters a new phase as self-driving competition intensifies

Shen Mengdan
Tesla's latest software update, "FSD (Supervised)," boosted its Full-Self Driving system to the next level in terms of safety and autopilot technology.
Shen Mengdan
Tesla's FSD enters a new phase as self-driving competition intensifies
Ti Gong

Tesla's Full-Self Driving system was upgraded to a new level with its most recent software update called "FSD (Supervised)," indicating the company's newest achievement in safety and autopilot technology.

On March 31, Tesla released a revised version of FSD V12 in the United States. For the first time, the word "Beta" is removed and replaced with "Supervised," indicating the end of the FSD Beta testing program.

Tesla's FSD is at the heart of the company's EV, allowing it to work fully autonomously.

FSD, or full self-driving, refers to the car's ability to drive in most cities without human involvement, to deliver a safer driving environment.

Tesla expressly introduced the notion of "FSD (Supervised) v12" in this update, which renews the city driving function with an end-to-end neural network trained on millions of video clips and replaces more than 300,000 lines of C++ code.

This shows that Tesla is increasingly moving away from traditional programming in the development of its autopilot technology and toward more powerful machine learning methodologies.

The FSD (Supervised) can perform lane change operations, choose forks in the road that the navigation route indicates, avoid other cars and obstructions, and make left and right turns, according to Tesla's update notes.

Nonetheless, it underlines the significance of the driver remaining aware and prepared to take over the driving when using the FSD (Supervised) component of self-driving.

Tesla's FSD enters a new phase as self-driving competition intensifies
Ti Gong

Tesla released a new version of FSD V12 in the US on March 31, removing the word "Beta" and replacing it with "Supervised."

Currently, the two main competitors in the autonomous driving industry are Huawei and Tesla, with their latest versions of Huawei ADS 2.0 and Tesla FSD 12.

Huawei ADS 2.0 makes use of Huawei's ground-breaking GOD-occupied grid network algorithm as well as the BEV sensor network.

The GOD network is significant because it no longer needs to recognize and name all impediments in the traffic environment but rather simply needs to plan which areas the vehicle can go through. The GOD network may theoretically "recognize" and avoid a barrier that is not on the BEV's whitelist of sensor networks, significantly increasing safety.

Tesla FSD's Occupancy Network is based on the same concept, but the difference is that Tesla FSD is a pure visual fusion solution, whereas Huawei's GOD network is a multi-sensor fusion of LIDAR, millimeter-wave radar, and cameras, which is superior in recognizing obstacles in inclement weather and low-light conditions.

Aside from Huawei, Xpeng's automatic driving system leads the new domestic car-making forces.

Its latest intelligent driving system, XNGP, adds the function of full intelligent assisted driving under urban road conditions (urban NGP), as well as an enhanced version of high-speed NGP and VPA memory parking based on the first-generation XPilot system, making it China's only system capable of full intelligent assisted driving under urban road conditions.

The conclusion may be drawn that Tesla's FSD differed significantly from the intelligent driving programs of domestic automobile manufacturers.

Tesla uses a pure vision program, relying solely on cameras to acquire information, which necessitates advanced algorithms, AI model training, and data processing capabilities.

Because its FSD program is based on data (mostly from road tests in the United States), it may cause issues when applied to China, where road conditions may be far more complex.

Domestic vehicle businesses, on the other hand, prefer the fusion perception program due to the complexity of domestic road conditions, which necessitates the use of many technologies to acquire information.

Huawei, Xpeng, Chang'an, and other domestic automobile firms have a competitive advantage in the local market because of their intelligent driving technologies that are well-suited to domestic roads.

Last November, Tesla made an official announcement about the promotion of the FSD landing plan in China, and the Tesla China owner's manual revised the FSD beta's function description.

When Tesla arrived in Shanghai in 2019, the "catfish effect," which drives members of a group to perform better in the face of competition, energized China's EV industry.

FSD is likely to enter China as a "catfish" in the realm of intelligent driving, intensifying industry competition.

Tesla's FSD enters a new phase as self-driving competition intensifies
Ti Gong

Avita 11 by Chan'an, Huawei, and CATL is fitted with the latest ADS 2.0.

Special Reports