The end-to-end architecture has gradually become the main theme of this year's competition in the field of intelligent driving. In order to compete for a leading advantage in the market, automotive companies have begun to increase their research and development efforts.
According to a later report by automotive media, the R&D department of NIO Intelligent Driving has recently completed a structural adjustment. Previously, the company's intelligent driving research and development part was focused on perception, regulation, and integration. After adjustment, the perception and control teams were merged into a large model team, and the integration team was reorganized into a delivery team. The merged large model team is led by Peng Chao, the team leader of Yuan Weilai Perception.
The restructured autonomous driving R&D department is still managed by Ren Shaoqing, Vice President of NIO Intelligent Driving R&D. After this adjustment, Ren Shaoqing conveyed to the team that he wants to abandon the technology route of "perception decision-making regulation" that has been used in the industry for many years. This means that NIO will more clearly explore using end-to-end large models to achieve advanced intelligent driving.
NIO Motors did not respond to the request for comment from Interface News.
According to public information, Ren Shaoqing participated in the establishment of the autonomous driving company Momenta and later joined NIO Motors in 2020. Peng Chao used to work as a senior visual algorithm engineer at Momenta and was the former employee of Ren Shaoqing at the company.
From relying on a large number of rules for manual programming to the era of deep neural networks, end-to-end intelligent driving models have integrated perception, positioning, navigation, prediction, and planning steps compared to before, reducing response time and bringing a more anthropomorphic driving experience. The processing methods for emergency braking and overtaking scenarios have been optimized.
Tesla released the FSD v12 version at the beginning of the year, once again sparking a wave of discussion on the future direction of intelligent driving in China. Zhou Guang, CEO of autonomous driving startup Yuanrong Qixing, once said that Tesla FSD's entry into China is accelerating competition. In contrast, the advantage of domestic companies is that the road conditions in China are more complex than those in the United States, and the data collection information for Chinese roads of the same kilometer is richer.
In addition to NIO, domestic companies such as Huawei, Xiaopeng, and Ideal have also proposed end-to-end intelligent driving solutions.
Huawei is expected to fully switch to the end-to-end architecture of ADS 3.0 in mass production in August this year. Ideal plans to launch a self-developed high-end intelligent driving solution based on the end-to-end large model by the end of this year or early next year. Xiaopeng previously announced the implementation of end-to-end network boarding, but there is still a gap compared to a model that has completed perception and control.
Traditional car companies are also striving to catch up, and Great Wall plans to achieve end-to-end intelligent driving mass production on three models this year, in cooperation with supplier Yuanrong Qixing.
In terms of the number of intelligent driving teams alone, compared to its competitors Zhonghua with over 7000 people, BYD with nearly 4000 people, and Xiaopeng with about 3000 people, NIO's personnel configuration is relatively streamlined. The company's plan is to launch an end-to-end active security feature in the first half of this year, and it is reported that the latest version including end-to-end AEB will be launched soon.
An industry analyst told Interface News that as the competition in the automotive market focuses on intelligent driving, new energy brands may hope to build technological barriers through self-development, and relatively weaker automotive companies may seek supplier cooperation.
Sources from intelligent driving solution suppliers point out that suppliers who have sensed the end-to-end trend in the industry have already started vertical layout two years ago. Whether NIO and other host manufacturers can catch up with the progress and form product advantages is still unknown, and the competition is still fierce.
However, the aforementioned individuals also pointed out that intelligent driving has not yet become a core factor affecting consumer car purchasing decisions. The 2024 McKinsey China Automotive Consumer Insight Report shows that under the influence of price wars, Chinese consumers have seen an increase in acceptance of intelligent driving, but a significant decline in willingness to pay. The key to breaking through is to launch intelligent driving with a better driving experience and more widespread usage scenarios, moving from "usable" to "user-friendly".