Urban automatic assisted navigation driving accelerates the implementation of smart driving, the first year of intelligent driving begins

Since the beginning of this year, a number of new car-making forces have been actively deploying relevant industrial chains and actively promoting the full implementation of city-level intelligent assisted driving.

Recently, Li Auto officially announced that urban NOA (Navigation Assisted Driving) will be launched for internal testing in Beijing and Shanghai in the near future, and the commuting NOA function will be launched within half a year.

Coincidentally, Xpeng Motors also announced not long ago that the urban NGP (intelligent navigation assisted driving) function has been officially opened in Beijing. It is reported that before this, Xpeng Motors’ NGP functions have been gradually opened in Shanghai, Shenzhen, Guangzhou and other places. At the same time, Cyrus and Avita also announced plans to launch city-level intelligent assisted driving functions.

It is generally believed in the industry that applying intelligent assisted driving functions from highway sections to urban highways is a key step for high-end intelligent assisted driving to move toward driverless driving .

Since the beginning of this year, a number of new car-making forces have been actively deploying relevant industrial chains and actively promoting the full implementation of city-level intelligent assisted driving. Thalys insiders said that due to policy catalysis and further maturity of technology, 2023 can be said to be the first year of intelligent driving. With the large-scale release of urban NOA in the second half of this year, the smart driving market is expected to be further catalyzed, and the penetration rate of advanced autonomous driving is expected to increase rapidly.

Smart driving welcomes the first year of development

According to my country's "Automotive Driving Automation Grading" standards, autonomous driving technology is divided into six levels. Navigation-assisted driving such as NOA is between L2 and L3 autonomous driving. This function can be implemented based on the navigation route set by the user. Intelligent navigation assists driving from point A to point B. At present, different car companies have slightly different names for intelligent navigation-assisted driving. Some call them NOA, while others call them NGP, NOP, and NCA. No matter how it is named, the basic functions and internal logic of its technical implementation are consistent.

In recent years, navigation-assisted driving represented by NOA has become more and more popular, especially new car-making forces represented by Tesla , Xpeng Motors, etc., which basically regard high-end intelligent driving assistance functions as the main features of their products. One of the selling points.

It is worth mentioning that before 2023, the intelligent navigation-assisted driving developed by new car-making forces will be mainly used in high-speed scenarios, and can realize active lane changes, automatic up and down ramps, etc. Public data shows that in 2022, China's sales of new intelligent connected passenger cars equipped with assisted autonomous driving systems will reach 7 million units, a year-on-year increase of 45.6%, and the proportion of new energy vehicles equipped with assisted autonomous driving systems will reach 48%. Among them, the number of high-speed NOA equipment reached 139,000.

By 2023, various car companies have unanimously expanded the use scenarios of NOA from highway sections to urban sections. In April this year, Huawei's high-end intelligent driving system AD S 2.0 was officially released. The system can achieve a continuous driving experience for highways, urban areas, and parking scenarios. The new system was first installed on the high-end intelligent driving version of the Wenjie M5, and will be launched on Avita 11, the new HI version of the Afar S and other models.

On June 15, Xpeng Motors officially opened urban NGP in Beijing, which is currently mainly applicable to various ring lines and major expressways in Beijing. Liu Dehao, an algorithm expert at Xpeng Motors, said that autonomous driving is currently an inflection point from the first half to the second half. At this stage, mass-produced urban assisted driving capabilities and full-stack closed-loop capabilities are the "ticket" for companies to enter the second half of competition. .

In addition to the above-mentioned car companies, Lideal, Weipai, Zhiji, NIO, Baidu, Pony.ai, and Haomo.com are also accelerating the layout of urban NOA.

Multiple factors promote the implementation of NOA

In 2022, with the collapse of the self-driving star startup Argo  AI , the entire industry fell into a downturn. Although it is backed by two major automobile giants, Ford and Volkswagen, the problem of commercialization still plagues this start-up. At the same time, the collapse of Argo AI also reflects the current status of the industry: that is, the higher the level of autonomous driving technology, the harder it is to achieve commercial breakthroughs.

Although the commercialization path is tortuous and long, the industry does not seem to have stopped exploring high-level autonomous driving technology. Liu Dehao analyzed that there are currently two paths for autonomous driving in the industry. One is bottom-up. Through sufficient mass production and operation of low-level autonomous driving, a large amount of data is accumulated to achieve data closed loop and technology iteration, and then gradually move towards more advanced technologies. High-level capability evolution; the second is a top-down approach, directly targeting the L4 or even L5 level and attacking high. The commercialization approach of this approach is mainly focused on specific operating scenarios.

"Currently, most OEMs adopt a bottom-up approach. We need to implement powerful functions with a low-cost, highly robust hardware architecture. We need to adapt to China's localized road conditions and driving habits and deliver high availability to consumers. , high-scenario and easy-to-use assisted driving products ." Liu Dehao believes that the ability to mass-produce urban assisted driving is the only way for car companies to achieve higher-level autonomous driving.

Thalys insiders said that there are multiple reasons why car companies are focusing on deploying urban NOA this year. From an enterprise perspective, in the competition of new energy vehicles, intelligent strength is the key factor in winning the second half. NOA is an intelligent driving function that tends to the L3 level. It is the superposition of low-level intelligent driving functions such as AC , LCC, ALC, etc. Complex urban NOA is the only way to achieve fully autonomous driving. The latter is also the way for car companies to create product differentiation. important areas of advantage.

At the same time, policy-level support is also a key factor driving the vigorous development of this market. Since the beginning of this year, in order to promote the commercial application of higher-level autonomous driving functions, the national level has introduced a series of policy support and guidance. Following Beijing, Shanghai, Guangzhou, Wuhan and other cities, Hefei also launched driverless taxis in June this year, while Shenzhen also launched L4 driverless commercial pilot operations at the same time.

Xin Guobin, Vice Minister of the Ministry of Industry and Information Technology, recently stated that pilots for the access and on-road access of intelligent connected vehicles and city-level "vehicle-road-cloud integration" demonstration applications will be launched in the near future to support the commercialization of L3 and higher autonomous driving functions. application.

Thalys insiders said that the national level and local governments have timely introduced a series of policies and plans to inject a boost into the development of the autonomous driving industry and catalyze L3 level autonomous driving to enter the pilot stage on the road.

In addition, the development of algorithms , data and chip technology has promoted iteration and industrial chain upgrading, and is also a key factor in promoting the large-scale mass production and commercialization of city-level NOA. Industry insiders judge that due to policy catalysis, further maturity of technology and lower prices, 2023 will be the first year of NOA. With the large-scale release of urban NOA in the second half of this year, the smart driving market is expected to be further catalyzed, and the penetration rate of advanced autonomous driving will increase. Expected to improve quickly.

Moving away from high-definition maps has become a trend

It is worth noting that with the accelerated implementation of urban NOA, different companies have also differentiated in their choices of technical routes. In the high-speed NOA implementation stage, with the exception of Tesla, various car companies still rely heavily on high-precision maps. Compared with ordinary navigation maps, the positioning accuracy of high-precision maps can reach centimeter level and can provide information such as road shapes, road markings, traffic signs and obstacles.

However, due to the high cost of high-precision maps, low frequency of maintenance and updates, and the need to improve coverage, there are bottlenecks in the application of high-precision maps in urban NOA.

As early as 2019, Tesla CEO Elon Musk publicly stated that over-reliance on high-precision maps would make the autonomous driving system extremely fragile and make it more difficult to popularize it.

Not long ago, Yu Chengdong, Huawei's Managing Director, CEO of Terminal BG, and CEO of Smart Car Solutions BU, also said that it is too difficult for high-precision maps to cover the whole country. China's roads are changing almost all the time. If you rely on high-precision maps, there is no way to popularize them.

Because of this, in the process of developing urban NOA, more and more car companies have adopted the model of “removing high-precision maps”. Recently, Li Auto demonstrated the company's newly launched internal beta version of Urban NOA to many media. It is reported that the ideal urban NOA achieves the four major functions of not relying on high-precision maps, identifying all things, planning and decision-making, and continuous evolution through large model algorithms.