The Corner Cases Data Challenge for Full Self-Driving (FSD)

May 31, 2024Updated 9:55 am

Tesla CEO Elon Musk introduced Robotaxi in China during his recent visit, as China Daily reported on May 8. While Tesla's Robotaxi demonstration test has received support, its Full Self-Driving (FSD) system has yet to be fully approved for use in China.

In this article, we discuss the benefits of Tesla's FSD technology, the role of data in FSD, the key considerations, such as the corner case and regulatory challenges of landing in China, and Chinese local Robotaxi.

The Game-Changing Potential of Tesla's FSD Technology

Tesla's Full Self-Driving (FSD) technology uses advanced machine learning from vast amounts of data to enable autonomous driving with features such as autopilot, traffic-aware cruise control, and advanced navigation.

In contrast to conventional systems, FSD opts for a comprehensive neural network over a combination of sensors and radars. This approach enhances adaptability, cuts costs, and improves performance in real-world situations[1].

At present, FSD is notably advanced in the United States.

Photo by Jorgen Hendriksen on Unsplash

Data: The Fuel Behind FSD's Success

According to Tesla's Impact Report, vehicles with Autopilot engaged have a better safety record, with only 0.18 accidents per 1 million miles[2].

Tesla's performance is made possible by collecting and analyzing large amounts of vehicle data to optimize its algorithms, and its vast fleet of Autopilot-enabled vehicles provides a significant data advantage.

Elon Musk predicts 6 billion miles of test mileage for autonomous driving[3]. Tesla has completed 1 billion miles so far[4].

Musk plans to launch Robotaxi for commercial use on August 8 and aims to collect more road test mileage using the FSD feature.

Key Considerations For The Chinese Version of FSD

There is no denying the enormous potential of the Chinese market. Not only is China one of the largest automotive markets in the world, but it is also a major development site for electric vehicles (EVs) and autonomous driving technology.

Tesla reportedly has more than 1.7 million vehicles in China[5], providing a wealth of driving data for FSD technology to further optimize and adapt the Autopilot system.

In addition, Chinese consumers are more receptive to new technologies, providing a fertile testing ground for Tesla's innovative technology[6].

Despite the potential of FSD, there are significant barriers to its widespread adoption in China.

First, China's complex road conditions and different traffic rules require Tesla to further adapt and optimize its FSD system.

Last year, a Chinese Robotaxi confused food stalls and pedestrians for trucks, underscoring the challenges in autonomous driving[7]. These incidents emphasize the significance of collecting diverse data on rare or complex road scenarios, known as corner cases, to boost the precision of self-driving technology.

Corner cases can involve a variety of road conditions, from bustling intersections during peak hours to roads under severe weather conditions such as heavy rain, snow, and fog. They also encompass special scenarios like sidewalks, backtracking, and rollovers. Traditional methods often fall short, so real-world test drives, varying weather conditions, and different traffic scenarios are essential.

In this case, understanding corner cases is crucial for the safety of self-driving vehicles in China. Collecting extensive, diverse, and high-quality data on long-tail scenarios is essential for improving autonomous driving algorithms.

As a premier provider of AI data services, maadaa.ai has released the report “maadaa.ai White Paper | Corner Cases in Auto-driving — The Data Challenges”. maadaa.ai's global network helps to address these data challenges for safer autonomous driving.

It can significantly aid in developing safer and more reliable autonomous driving technology[8].

image by maadaa.ai Auto-Driving Dataset

Next are China's unique regulatory challenges. Data security and privacy are also important considerations in China. Tesla must ensure that its systems are fully compliant with Chinese regulatory requirements.

In addition, Tesla has to address privacy and security concerns to gain public trust.

Despite these hurdles, Tesla is making progress.

Most notably, Tesla is collaborating with Baidu on Advanced Assisted Driving maps for the Chinese version of FSD, aiming to solve the map data issue for Autopilot in China and gain regulatory approval.

The Rise of Robotaxi in China

China's challenging road conditions and dense urban traffic make operating robotaxis more difficult than in other countries. That's one of the key reasons that Chinese Robotaxi companies in China are advancing faster than Tesla.

maadaa.ai has also covered the case study on Chinese Robotaxi to focus on Robotaxi in China, the review, the respective technologies, and the important factors for Robotaxi[9].

In fact, companies like Pony.ai, WeRide, and Baidu have already initiated driverless passenger transportation programs.

Baidu's Apollo Go, China's first mass-market Robotaxi platform, has launched manned testing and operational services in cities like Beijing, Guangzhou, Shenzhen, etc.

Pony.ai is the first domestic self-driving company to obtain a taxi operating license in 2022. Pony.ai has obtained a “front-row unmanned” testing permit in Beijing Yizhuang and mixed-traffic pilot operation qualifications in Guangzhou city.

Founded in 2017, WeRide has over 600 autonomous vehicles and has conducted research, development, testing, and operation in over 26 cities worldwide.

In addition, several cities have set up demonstration zones to promote Robotaxi to verify the safety of the technology and the feasibility of commercial operation.

As of November 2023, Beijing's high-level self-driving demonstration zone has reported a cumulative self-driving mileage of 17 million kilometers.

Shanghai has opened a total of 1,003 self-driving test roads with a total mileage of more than 2,000 kilometers as of March this year.

The latest progress comes from Hangzhou, where the “The Hangzhou Intelligent Connected Vehicles Testing and Application Promotion Regulation” took effect on May 1, with Hangzhou's main center serving as an area for testing and applying intelligent connected vehicles[10].

China's self-driving test area is expanding and moving from testing to commercial operation, indicating the maturation of self-driving technology, policies, and regulations in China. This will also pave the way for Tesla to conduct driverless taxi testing on Chinese roads.

In addition, the emergence of Tesla's FSD technology is expected to further stimulate the growth of the Robotaxi industry in China.

Photo by Decry.Yae on Unsplash

The Path Forward: Collaboration and Innovation

Tesla's Fully Self-Driving (FSD) technology could significantly influence Chinese autonomous driving companies, inciting both fierce competition and potential technological partnerships.

Given its proven success in the market, Tesla's FSD could propel Chinese companies to expedite their technological advancements. Tesla's business model could also offer them new insights and opportunities for collaboration.

However, realizing this potential will require overcoming significant challenges, including regulatory hurdles, data collection and management, and ensuring the safety and reliability of the technology.

The success of Tesla's FSD technology hinges on its reliance on real-world data and collaboration with local partners. The rapidly changing regulatory landscape in the Chinese market and the dynamic Robotaxi industry will afford Tesla valuable opportunities to further hone its FSD capabilities.

Sources：

Any further information, please contact us.