Will high-speed L3 intelligent driving really arrive next year?

Since cars entered the era of intelligence, discussions about autonomous driving have been endless. Since the second half of 2023, leading companies have announced that they will initially have full-scenario autonomous driving capabilities.

In June this year, the Ministry of Industry and Information Technology and four other departments identified nine consortia to carry out the pilot program for the access and road use of intelligent connected vehicles, marking a new stage in China's promotion of intelligent connected vehicles. Compared with the preliminary notice in November last year, the new guidelines explain the stages and significance of the pilot program in more detail. It is reported that the first batch of units to enter the pilot program for the access and road use of intelligent connected vehicles include Changan , BYD , GAC, SAIC, BAIC Blue Valley, FAW , SAIC Hongyan, Yutong Bus and NIO.

It is worth noting that manufacturers such as Jiyue, Zeekr, Mercedes-Benz and BMW obtained L3 test licenses in 2023, but did not appear on the pilot list for 2024. Among them, Jiyue announced in March this year that its point-to-point navigation assistance has covered more than 400,000 kilometers of roads across the country, and according to the plan, it will achieve navigation capabilities covering 100% of the country's highways and elevated roads in 2024.

In addition, Xpeng Motors ' G9 SUV obtained the L4-level Guangzhou test license as early as 2022, and subsequently obtained the qualification for autonomous driving road testing on Chinese roads. The company said on social media on June 4 that it expects that its models equipped with XNGP will have the ability to perform NOA nationwide in the third quarter of this year.

The latest data shows that by the end of April 2024, my country has opened more than 29,000 kilometers of intelligent connected vehicle test roads, issued more than 6,800 test demonstration licenses, and the total road test mileage has exceeded 88 million kilometers. The Ministry of Industry and Information Technology said that at present, some products have met certain conditions for mass production and application.

So how far are we from L3? The industry believes that promoting pilot projects is just the beginning, and improving the laws and regulations for L3 and above autonomous driving models is the key to large-scale commercialization. However, the challenges faced by large-scale commercialization are not just the imperfections of laws and regulations.

Cui Dongshu, secretary general of the China Passenger Car Association, said that the biggest difficulty in promoting L3 and above assisted and autonomous driving is undoubtedly the limited level of existing technology. Existing intelligent assisted driving systems still have defects in identifying the environment, establishing the ability to recognize the environment at the level of experienced drivers, and information processing capabilities.

At the same time, China's road conditions are complex, especially in cities, which has a great impact on China's autonomous driving and has a significant long-tail effect. If the relevant needs are to be met, the existing intelligent driving system still has difficulties in decision-making and prediction in complex environments. Therefore, the car needs to have a significant improvement in the coordination of intelligent algorithms and related hardware systems.

Although L3 autonomous driving is regarded as a watershed in the legal sense of autonomous driving, it is still not completely independent of humans from a technical point of view. It can provide services and work completely autonomously and still relies on humans for monitoring and taking over in times of crisis. Of course, this also shows that since L3 is relatively less technically difficult, it should be easier to implement in the short term.

Previously, Yu Chengdong also said in an interview that the true arrival of L3 intelligent driving depends on national regulations, but with Huawei's capabilities now, it may be able to be done soon. That is, closed sections such as highways and expressways can be realized quickly, but L3 intelligent driving in urban areas may take some time to polish.

In addition, there are reports that Shanghai is promoting 10 Tesla vehicles to carry out full-automatic driving (FSD) pilot projects. Cui Dongshu said that Tesla FSD's entry into China will undoubtedly have a promoting effect. This promotion is different from the promotion of hardware. Behind the promotion of hardware is the upgrading of suppliers and supply chains, which drives the entire industrial chain. At the same time, as long as the hardware can be purchased, everyone can have it as standard. Behind the software is the company's exclusive intellectual property rights, which is also the core asset of Hyundai Intelligent Driving Company. The introduction of related software into the market will not directly lead to an overall leap in the level of domestic intelligent driving.

Zhu Xichan, a professor at the School of Automotive Engineering of Tongji University, said that there is no specific standard to clearly define L3 and L4 autonomous driving technologies. my country's mass-produced models are generally at the L2 assisted driving stage, and some brands will promote their products at the L2+ level. According to data from the Ministry of Industry and Information Technology, in 2023, 3.6842 million new energy passenger vehicles in my country will be equipped with L2 assisted driving functions as standard, with a penetration rate of 50.56%.

Fang Yinliang, global director and partner of McKinsey, also said that the development of autonomous driving still faces challenges including various factors such as cost. The key to when L3 or higher level autonomous driving can be truly implemented lies in whether it can truly meet consumers' expectations and solve consumers' needs in various scenarios in the future.

The obvious difference from L2 level to L3 level or even higher-level autonomous driving is that the application scope of the designed operating area of ​​autonomous driving will be greatly improved. Under certain designed operating conditions, the system will be able to perform all dynamic driving tasks; this also means that more stringent requirements will be placed on the overall safety and reliability of the intelligent driving system, requiring multi-dimensional upgrades in takeover mechanisms, redundant safety, human-computer interaction, product operation safety, network security, data security, and data recording.

Cinda Securities believes that the gradual implementation of L3 autonomous driving tests is expected to promote further iterations of high-level intelligent driving. Automakers have greatly increased their attention to high-level intelligent driving, and intelligent driving is also expected to become an important consideration for car owners when buying a car.

Many car companies have also begun to deploy L3 intelligent driving solutions. On June 15, Jin Yuzhi, CEO of Huawei's Intelligent Automotive Solutions BU, publicly stated that he hopes to open up high-speed L3 intelligent driving next year. He said that China currently attaches great importance to L3 intelligent driving.

At the same time, he also discussed in depth the many highlights and future plans of Huawei Qiankun Intelligent Driving. First of all, he emphasized that Huawei Qiankun Intelligent Driving's pure vision solution has surpassed Tesla's FSD, demonstrating Huawei's deep strength in the field of intelligent driving technology.

Jin Yuzhi revealed that next year, up to 2 million cars will be equipped with Huawei Qiankun Intelligent Driving System, which is comparable to the deployment level of Tesla FSD, fully demonstrating Huawei's strong momentum in the intelligent driving market. In terms of autonomous driving level, he expects Huawei Qiankun Intelligent Driving to reach L3 level next year, and in the next three to five years, it is expected to reach L4 or even higher levels, further promoting the development of intelligent driving technology.

Li Auto Chairman and CEO Li Xiang pointed out at the 2024 China Automotive Chongqing Forum that Li Auto plans to push the national mapless NOA (navigation assisted driving) function to all users in the third quarter of this year. At the same time, the company will also push the end-to-end + VLM autonomous driving system trained based on 3 million clips to test users. It is expected that as early as the end of this year and at the latest in early next year, Li Auto will launch a more complete autonomous driving system trained with more than 10 million clips to provide users with a supervised L3 autonomous driving experience. With the improvement of computing power and the continuous optimization of models, unsupervised L4 autonomous driving will be realized in the next three years.

I believe that with the continuous advancement of science and technology and the promotion of policies, the widespread application of high-speed L3 intelligent driving will be just around the corner.