This week, the Ministry of Industry and Information Technology disclosed that it will launch pilot projects for the access and on-road traffic of intelligent networked vehicles, organize city-level "vehicle-road-cloud integration" demonstration applications, support conditional autonomous driving (L3 level), and Commercial application of higher-level autonomous driving functions.
Just in November last year, the Ministry of Industry and Information Technology publicly solicited opinions on the "Notice on Carrying out the Pilot Program for the Access and Passage of Intelligent Connected Vehicles (Draft for Comment)", clearly stating that: qualified road motor vehicle manufacturers and those capable of mass production Carry out access pilots for intelligent networked car products equipped with automatic driving functions that meet the conditions.
According to the description of the application documents for this pilot program, first of all, it is only for the B-side pilot users with independent legal person qualifications and certain ability requirements; at the same time, the consortium (pilot car manufacturers, pilot users) is required to provide products with automatic driving functions Promotion plan for the next 3 years.
This means that car companies are beginning to aim at the real and practical positive research and development path to realize the transformation of the traditional autonomous driving industry chain. In the past, autonomous driving companies usually participated in the selection and customization of software and hardware, and then purchased new cars (partially cooperated with car companies) for modification and debugging.
And just last month, the relevant person in charge of the Shanghai Economic and Information Commission stated that it is further deepening its cooperation with Tesla to promote the layout of functional sectors including autonomous driving and robots in Shanghai.
In addition, at the beginning of this year, Mercedes-Benz officially stated that it will start the test of L3 conditional autonomous driving in China, and the expert team of China's local R&D department will deeply participate in the system optimization of DRIVE PILOT for the operation of China's road conditions.
In terms of self-owned brands in the Chinese market, several brand models including Xiaopeng, Ideal, Avita, Zhiji, and Huawei Wenjie are planning to upgrade from high-speed NOA to urban NOA. This is also seen as a critical step from L2 to L3.
At the same time, in terms of reducing the cost of high-end intelligent driving, most car companies have begun to seek how to avoid relying on high-precision maps in NOA scenarios (for car companies or users, it is also a big expense), while BEV+Transformer The blessing can solve the problem to a certain extent.
According to the monitoring data of Gaogong Intelligent Automobile Industry Research Institute, from January to April 2023, 239,600 passenger cars with NOA optional configuration will be delivered in the Chinese market, a year-on-year increase of 87.63%. Among them, the proportion of models below 300,000 has exceeded 25%. .
At the same time, the penetration rate of new energy vehicles continues to rise, which also lays the foundation for the subsequent scale-up of NOA vehicles.
According to the monitoring data of Gaogong Intelligent Vehicle Research Institute, in the first quarter of 2023, 1.2296 million new energy passenger vehicles will be delivered in the Chinese market (excluding imports and exports), a year-on-year increase of 23.01%, accounting for 30.01% of the overall new car deliveries, an increase of about 9% year-on-year. percent.
Among them, in terms of intelligence, in the first quarter, 639,100 new energy passenger vehicles with standard L2 (including L2+) assisted driving were delivered, a year-on-year increase of 64.21%, which was significantly higher than the overall market growth rate (19.35%). The loading rate of front-mounted vehicles exceeded 50%, reaching 51.98%.
Among them, from January to April, the allocation rate of new energy vehicle NOA (including optional equipment, including hardware standard configuration + software paid subscription) has exceeded 20%, reaching 20.16%. This means that from basic L2 to high-speed NOA, it has become the main configuration in the new car market. From high-speed NOA to urban NOA, it has become a key link for new energy car companies to continue to improve the competitiveness of their models.
Last year, Wu Xinzhou, vice president of Xiaopeng Automobile's autonomous driving, publicly stated that the company's intelligent assisted driving route planning is: in 2022, it will mainly focus on the laying of single-scenario assisted driving; Moving toward fully autonomous and unmanned driving.
The ideal car that has passed the threshold of scale, based on BEV perception and Transformer model, realizes end-to-end training of urban NOA navigation assisted driving (without relying on high-precision maps) will start to land this year; realize expressways and closed roads in 2025 The next level of L4 automatic driving.
In the view of this company, the realization of L4 autonomous driving requires the application of a large amount of data and the gradual accumulation and iteration. The market penetration rate of L2 ADAS has reached the critical point of crossing the gap, which has become an important foundation for major car companies and suppliers to sprint to high-level autonomous driving (NOA, L3, L4).
Just this week, Li Auto released a new function called "commuting NOA" for the first time (somewhat similar to low-speed memory parking). Users can set their own routes and learn NPN features from their own cars. NOA can be used on urban commuting routes. Function. Among them, the relatively simple route can basically be activated within 1 week, and the more complicated route is expected to complete the training in 2-3 weeks.
However, regarding the activation of L3 and L4 policies, some industry insiders pointed out that based on cost and regulatory considerations, "automatic driving access certification and potential liability costs are still relatively high in the short term, and the scale remains to be seen."
Previously, the "Draft for Solicitation of Comments" proposed that pilot automobile manufacturers should assume the main responsibility for the quality and production consistency of ICV products, continue to fulfill the requirements for software upgrade management and filing commitments, and ensure that the production and sales of ICV products Responsible for quality and safety.
In comparison, L2+/L2++ still belongs to the field of driver assistance at the regulatory level, but on the basis of traditional L2 ADAS functions, the function ODD of some scenarios is enhanced. For example, in some countries and regions where policies allow, conditional "hands-free" can be achieved.
At present, typical applications in the pre-installation market include high-speed pilot assisted driving, which provides a system experience close to automatic driving on the basis of high-precision maps and other perception redundancy. Speed limit/curvature adjusts vehicle speed. In urban scenes, functions such as traffic light recognition and road sign recognition have been added.
From an industry perspective, L2+/L2++ is easier to implement than L3, and the system cost is lower. On the one hand, the system retains the responsibility of the driver, but can give the driver a near-autonomous driving experience under certain conditions. On the other hand, it is easier for car companies to promote large-scale boarding without taking responsibility.
At the same time, another big attraction of L2+/L2++ is that it is an important stepping stone on the road to true autonomous driving. In particular, with the help of Tesla's "shadow mode", car companies and smart driving system suppliers can gradually solve difficult edge scene problems through large-scale data collection and model training.
According to Gaogong Intelligent Vehicle Research Institute, in the Chinese market, starting from 2023, NOA scale (from high-speed to urban areas) will be the protagonist; starting from 2024, a small number of car companies will try L3/L4 level Policies of different cities); from 2025, L3/L4 will enter the cycle of small-scale mass production.
Gaogong Intelligent Automobile Research Institute predicts that with the launch and delivery of more NOA-equipped new vehicles in the second half of the year, especially the addition of first-line brand mass-produced models, the number of NOA pre-installed standard equipment this year may sprint to the 500,000 mark. At the same time, considering the scale of optional (including software OTA payment) models, it is estimated that the potential NOA carrying capacity for the whole year will exceed 1.5 million vehicles.
At the same time, the city-level "vehicle-road-cloud integration" demonstration application will activate the 5G and V2X markets and start ahead of schedule. Accelerating the technical application of 5G vehicle-road coordination and promoting the coordinated development of electrification and intelligent networking have also become the main target directions of recent policies.
In July 2020, 3GPP announced that the Rel-16 version was frozen, completing the 5G V2X standard based on the 5G framework for the first time. The biggest difference from the previous version is that Rel-16 makes full use of 5G's large bandwidth, low latency, and high reliability features to improve the perception capabilities between vehicles, and then focus on supporting autonomous driving technology.
Previously, LTE V2X was more about digitizing infrastructure and providing basic security services, including digitizing the system of traffic lights and then broadcasting them as wireless signals; while the Rel-16 version of 5G V2X technology is more about in-vehicle Collaborative perception, path planning, and even emergency braking between vehicle and road section perception.
According to the monitoring data of Gaogong Intelligent Vehicle Research Institute, in 2022, 170,100 standard C-V2X passenger cars in the Chinese market (excluding imports and exports) will be delivered, a year-on-year increase of 268.18%, and the pre-installation rate is 0.85%. In the quarter, 46,100 C-V2X vehicles were delivered as standard, a year-on-year increase of 151.91%, and continued to maintain a high growth rate.
Gaogong Intelligent Vehicle Research Institute predicts that with the gradual improvement of domestic roadside infrastructure, the 5G deployment rate will exceed 25% in 2025, which will have the opportunity to drive the LTE/5G-V2X pre-installation rate to exceed the 10% mark in the same period .
In addition, during this year's Barcelona Communication Exhibition, Huawei revealed that 2023 will be a critical year for 5.5G. Standards, spectrum, industry chain, ecology, applications and other fields are progressing faster than expected. The 5.5G standard will be delivered in 3GPP releases 18, 19 and 20, of which release 18 will be frozen in the first half of 2024, and will further enhance perception and collaborative communication capabilities.
At the same time, 5G, V2X, as an upgrade path for vehicle communication, provides additional support functions (such as high-precision positioning, vehicle-road coordination), to a certain extent, it provides the possibility of hardware reuse for the development of related functions of car companies. It is also a way to reduce costs.