news

Editor’s Note:Tencent Auto's editorial department calls the past decade of electrification the "stormy era" of China's auto industry. Now, standing at the historical node of 2024, which is known as the "first year of intelligent driving", we can't help but ask: What technical route will the major players in the industry adhere to? How will they build their own competitive barriers? Tencent Auto has specially launched a series of intelligent driving planning. Through interviews, actual tests, cross-comparisons, reviews and other methods, it strives to stand at the origin of history and further understand the huge changes that may occur in the auto industry in the next ten years, so as to provide readers and the industry with a more comprehensive content guide and leave some historical footnotes for the industry.

Tencent News "High Beam"

Contributing Author:Slight

edit:Stone Ding

Since more than two years ago, when car companies and intelligent driving solution providers turned to high-level driving assistance routes with greater implementation and monetization opportunities, autonomous driving, especially highly autonomous driving (L4) and fully autonomous driving (L5), have been labeled as "money-losing" and "pie in the sky".

Unexpectedly, halfway through 2024, the enthusiasm for autonomous driving has suddenly ignited again. There is more than one burning fuse.

Overseas, since the end of last year,TeslaAfter the release of FSD V12, a large number of Tesla owners uploaded the real performance of their vehicles in different scenarios on social platforms, and commented that Tesla already has capabilities similar to those of human drivers. Subsequently, Musk stated on different occasions that FSD V12.5 or 12.6 are planned to enter the domestic market before the end of the year.

then,Xpeng MotorsChairman and CEO He Xiaopeng went to the United States to test the FSD system, and said frankly that the overall performance of FSD is very sophisticated. Tesla, which has made a big bet on AI, will surpass and replace the current leader in autonomous driving Waymo in the short term and become one of the most important players in the autonomous driving industry.

But just when everyone was looking forward to seeing Tesla's long-publicized Robotaxi in August, the press conference was announced to be postponed, causing the stock price, which had been driven sharply by AI and FSD, to fall immediately, causing an uproar in the market.

In China, in May this year, Baidu's self-driving car, which confidently declared at Apollo Day that "self-driving technology is at least three to five years ahead of Tesla Robotaxi", ushered in a highlight moment against the backdrop of its old rival's default on delivery - the number of orders in Wuhan exploded. Subsequently, Baidu's Hong Kong stocks rose rapidly, not only forming a sharp contrast with Tesla, but also sparked a heated discussion in the industry about whether the era of self-driving has arrived. The various performances of the car on the streets of Wuhan have completely made self-driving cars popular.

Today, autonomous driving is gradually shedding the labels of "vague" and "pie in the sky" and is about to or has already appeared in people's daily lives, becoming a new role that cannot be ignored in transportation activities.

However, how to ensure that all this can be on the right track, to ensure that autonomous driving technology as a new quality of productivity has sufficient room for development and is not constrained by the existing environment, while ensuring that other traffic participants can pass safely and normally, and also to consider the general public's cognition and acceptance of new things.

All of this is still undecided, and it is by no means something that a specific autonomous driving company or companies can independently promote and complete. It is a huge and complex system facing the entire society, which needs to be supervised by the national government and promoted by the whole industry.

Autonomous driving is an industry that integrates multiple fields and has a complex legal framework.

As an industry that crosses multiple fields, the legal framework for autonomous vehicles is very complex. If divided into different fields, it can be roughly divided into product management, traffic management, infrastructure and information security.

Regulations such as motor vehicle registration, traffic law enforcement, and vehicle insurance, which are familiar to most drivers, belong to the traffic management category; network security, cloud platforms, and information data, which are emphasized by intelligent connected vehicles, are classified as information security content; as for road facilities, information communications, and high-precision maps that are strongly related to "vehicle-road-cloud", they are all classified in the infrastructure column.

These areas have their own focus and are intertwined and indispensable. They are closely woven into the legal framework for self-driving cars and provide them with protection.

Although there are many legal provisions, in the final analysis, they all serve the products. Road testing, standard management, and product access related to automotive products are top priorities.

To put it in a more down-to-earth way,If a car with autonomous driving function wants to be driven on the road normally, the two most important issues to be solved are access and driving.

The former means that car companies are qualified to produce self-driving cars, and self-driving vehicles can be announced for sale; the latter means that self-driving vehicles have the same road rights as other vehicles, and the body of the vehicle can be equipped with regular blue or green license plates and driven on public roads.

Among the many laws and regulations currently, the policy documents that directly address these two issues mainly include the "Notice on the Pilot Work of Intelligent Connected Vehicle Access and Road Notification" issued at the national level.(Officially released jointly by the Ministry of Industry and Information Technology, the Ministry of Public Security, the Ministry of Housing and Urban-Rural Development, and the Ministry of Transport in November 2023), and regulations on the management of intelligent connected vehicles issued at the local level, a typical example of which is the "Regulations on the Management of Intelligent Connected Vehicles in Shenzhen Special Economic Zone"(Effective in August 2022), "Wuhan City Intelligent Connected Vehicle Development Promotion Regulations (Draft for Comments)"(Released in February 2024)、Beijing Autonomous Driving Vehicle Regulations (Draft for Comments)(Released in June 2024)wait.

How to legally produce and put self-driving cars on the road?

In the "Notice on Carrying out Pilot Work on Access and Road Notice for Intelligent Connected Vehicles" (hereinafter referred to as the "Notice"), several core issues are clarified:

1. What is the level of autonomous driving of the smart connected vehicles that are allowed to enter and go on the road?

According to the national standard "Automobile Driving Automation Classification", conditional autonomous driving and highly autonomous driving, which we usually call L3 and L4 autonomous driving vehicles, can participate in the pilot activities of access and road access. However, it is worth noting that the highest level of fully autonomous driving (L5) in the driving automation classification is not covered by the "Notice".

Why is L5 temporarily excluded? The reason is easy to understand.

By definition, conditional autonomous driving (L3) requires a driver in the driving seat who can take over in the event of vehicle disability or other emergency situations; highly autonomous driving (L4), although it is not mandatory to have a driver/safety officer in the car, can define the vehicle's driving area through electronic fences and other means. In other words, L3 and L4 autonomous driving vehicles are either supervised or restricted in movement, and are generally still within the "controllable" range. Using these two levels of vehicles as a breakthrough point to test the complete administrative process for the admission and road use of autonomous vehicles is most in line with the new policy's characteristics of small steps and fast progress and a controllable rhythm.

As for L5 autonomous driving cars, considering that current technological development has not yet reached this level, policy research on L5 can be finalized after L3/L4 policies have been accumulated and improved, and there is no need to rush.

2. Which companies are eligible to participate in the pilot activities?

In the "Notice", the assessment objects for the access and road use of autonomous vehicles are car companies, users and models, among which "users" are a newly added role.

The so-called "user" can be generally understood as the autonomous driving vehicle operating platform, such as the Baidu Carrot Run mentioned above, and the equally rapidly developing autonomous driving companies such as Pony.ai, AutoX, and Scion Intelligent.

There are two key points here.First of all, car companies have become one of the responsible parties for self-driving cars.

Traditionally, car companies, as manufacturers, should be responsible for the entry of vehicle models. However, almost all self-driving vehicles on the road are modified by self-driving car companies after purchasing the whole vehicle. If any problem occurs during road testing of self-driving cars, it is difficult to prove that the car companies are responsible because the vehicles have been modified.

But the Notice directly stated thatThe current situation where self-driving cars are only allowed to be modified later and put on the road for temporary license after suspension testing is not sustainable.

The Notice stipulates that car companies are responsible for the production and manufacturing of self-driving cars. After delivering them to users, the users are responsible for all situations of self-driving cars driving on public roads, including daily operation and maintenance and accident handling and compensation. Both car companies and users are responsible for self-driving cars and need to meet the requirements of the Notice for both car companies and users.

In other words, the kits related to the autonomous driving function will change from "after-sales modification" to "mass pre-installation". Car companies must endorse the production and manufacturing quality of autonomous vehicles and cannot evade their responsibilities by using "modification" as an excuse.

Secondly, self-driving cars can only be sold to users through TO B sales, and cannot be sold to TO C to enter the traditional passenger car market. In other words, individual consumers cannot purchase L3/L4 self-driving vehicles.

Autonomous driving vehicles are centrally managed by the user (operating platform). The biggest advantage is that they are easy to manage and have clear responsibilities and rights.

Take the popular car Luobo Kuaipao in Wuhan as an example. There is either a safety officer in the car ready to take over at any time, or a remote driver to provide one-to-many support for the autonomous driving system. Both the safety officer and the driver are professionals. They have received professional training and practical experience in how to take over the vehicle, how to adjust the vehicle posture to get out of trouble, how to predict vehicle driving abnormalities, and how to ensure the safety of the vehicle and passengers. Their abilities far exceed those of ordinary consumers.

Since the main purpose of the Notice is to allow self-driving cars to go through the administrative procedures for access and road use, and does not seek to allow vehicles to enter the passenger car market, it is undoubtedly a better choice to entrust the unified management of the vehicles to the user entity rather than to leave it to a variety of individuals.

For ordinary consumers, if they want to experience self-driving cars, becoming a passenger rather than a car owner is also a more suitable option.

3. If one party to a traffic accident is a self-driving car, how is liability determined?

As self-driving cars take to the road, an unavoidable issue is how to handle traffic accidents and determine liability.

If the self-driving car is innocently involved in a traffic accident, then the party at fault should bear the responsibility. What if the self-driving car actively causes the illegal behavior or traffic accident?

As a guiding document at the national level, the Notice also provides the judgment logic of the core guidance opinions. The original text itself is difficult to understand, but in simple terms, it is to judge according to the different states of the vehicle.

• If the vehicle does not have its automatic driving function activated when an accident occurs, the driver of the vehicle will be held responsible for negligent driving according to current traffic regulations;

• If the vehicle is in the state of automatic driving activation when an accident occurs, the user shall bear the responsibility, and the compensation shall be borne jointly by the insurance company and the user;

• If it is clear that the root cause of the accident is the lack or failure of the autonomous driving function, then the user can seek compensation from the car manufacturer.

Although there are few practical details, the above-mentioned judgment criteria basically cover possible situations. Further refinement will depend on the public security, transportation and other departments to continuously accumulate experience and make attempts during the pilot activities.

After the "Notice" was officially released at the end of last year, only half a year later, 9 automobile manufacturers had jointly used the entity, becoming the first batch of joint applicants to implement the "Notice" and entering the list of consortiums for the pilot program of intelligent connected vehicle access and road use.

More importantly, local governments have begun to issue L3 autonomous driving road test licenses, involving car companies includingBMW，Benz，SAIC Zhiji、Chang'an、GAC、NIO, Seres,Extreme、ZeekrBrands such as domestic brands, international brands and new forces are blossoming in all aspects.

• July 2023,BYDObtained a conditional autonomous driving test license for expressways in Shenzhen, becoming the first automaker to obtain a conditional autonomous driving test license in China;

• In November 2023, Seres will useQJM9(equipped with ADS 2.0) obtained 11 test licenses (5 Chongqing licenses, 6 Shenzhen licenses). SERES is the only automaker that has obtained L3 autonomous driving test licenses in two places at the same time;

• December 2023,Polar FoxBecame one of the first companies approved for Beijing's conditional autonomous driving (Level 3) highway road test license;

• In December 2023, Mercedes-Benz became the first batch of Beijing's conditional autonomous driving (L3) highway test licenses; BMW and Zhiji officially announced that they had officially obtained the L3 level autonomous driving test licenses for expressways in Shanghai;

• In March 2024, Zeekr's products obtained the Shanghai Expressway L3 autonomous driving test license;

However, it should be pointed out that the release of L3 autonomous driving road test licenses does not mean that these companies and models have gone through the administrative procedures for access and road use. This is only a prerequisite and preparatory work for the pilot work. These road test vehicles need to run enough time and mileage before entering the next step of testing and safety assessment. Only after passing the assessment can they finally submit the pilot application for access and road use.

Wuhan, Shanghai, Guangzhou, Shenzhen, and Beijing, these rapidly developing cities with strong comprehensive technological strengths, have all formulated management standards that are more suitable for local areas, more detailed, and easier to implement under the framework of the "Notice on the Pilot Work of Access and Road Notification for Intelligent Connected Vehicles", combining their respective urban traffic management and operation characteristics.

A typical example is that the road areas open for autonomous driving road testing and trial runs in various cities have different characteristics.

Taking Shanghai as an example, in addition to the Anting National Intelligent Connected Vehicle Demonstration Zone in Jiading District, which is home to many automakers, the demonstration application of L3 autonomous taxis is mainly concentrated in Jinqiao and Lingang New Area. These are remote suburban areas that are a certain distance from the city center, with wide roads, clear road signs, stable traffic flow, and basically no mixed traffic of people and vehicles.

The Guangzhou area is more down-to-earth. Areas such as Nansha and Huadu have steadily carried out mixed-traffic pilot projects. The driving range of autonomous driving vehicles covers typical commercial and living areas such as office buildings, star-rated hotels, scenic spots, schools, shopping centers, and communities.

Wuhan has gone a step further. The operating range of Baidu Carrot Run has expanded from a small area outside the southeast fourth ring road to most of the city, crossing the Yangtze River Bridge and running into the streets and core urban areas of Wuhan.

From remote suburbs with vast land and sparse population, to expressways and trunk roads with good road conditions, to urban cores with mixed traffic of people and vehicles, bustling and risky. The open areas have different characteristics and road grades, and the corresponding management requirements are naturally different.

Under the general principle that "the same requirements in different cities can be recognized remotely", if car companies or autonomous driving technology companies want to expand into more cities, they only need to meet the "special" requirements of different cities.

For enterprises, this operating mode can avoid repeating the same test content during the "opening of the city" process and save resources, while also making the greatest investment in overcoming "city-specific requirements" and quickly improving the autonomous driving capabilities to cope with different traffic environments.

The benefits are obvious to local governments. They do not need to read the test reports of pilot enterprises from beginning to end, nor do they need to worry about any minor flaws in the management requirements. When encountering special situations, they can refer to the management requirements and handling methods of sister cities, work together, and move forward based on each other's experience. This is the optimal solution in terms of efficiency and cost.

The only special case that has some content that breaks through the "Notice" is the "Regulations on the Management of Intelligent Connected Vehicles in Shenzhen Special Economic Zone".

Shenzhen's policies are ahead of the curve, but technical verification remains to be solved

Compared with the above-mentioned "Notice", the "Regulations on the Management of Intelligent Connected Vehicles in Shenzhen Special Economic Zone" (hereinafter referred to as the "Shenzhen Regulations"), which was officially released in June 2022, has made breakthroughs in both applicable objects and scope of use.

Why can a local policy break through the national policy framework? The reason is that in 2019, the central government granted Shenzhen the right to pilot in emerging fields such as artificial intelligence and driverless cars, and encouraged and supported Shenzhen to use its special zone legislative power, which gave birth to the cutting-edge "Shenzhen Management Regulations".

First, the scope of application of the Shenzhen Regulations has been expanded from conditional autonomous driving (L3) and highly autonomous driving (L4) to fully autonomous driving (L5), covering all levels of autonomous driving.

Secondly, the Shenzhen Regulations clearly state that as long as smart connected vehicles are included in the Shenzhen smart connected vehicle product catalog and obtain relevant access,You can sell, can be driven on the road, and can also engage in business transportation services with the permission of the transportation department. The key here is "saleable", which means that the smart connected cars in Shenzhen can be directly sold to consumers in the passenger car market. This is fundamentally different from the restriction of the "Notice" that only provides TO B transportation services.

The applicability of L5 fully autonomous driving and its targeting of the passenger car market are major strides forward for the Shenzhen Regulations, but in other areas, they are relatively conservative and contractionary.

Taking the fully autonomous driving L5 model as an example, although the Shenzhen Regulations allow L5 vehicles to not install manual driving devices such as steering wheels, brakes and accelerators, and not be equipped with drivers, they restrict the driving areas and road sections of the vehicles. In the formal definition of autonomous driving, the L5-level autonomous driving system should complete dynamic driving tasks in all road environments. The so-called "all roads" means that there should be no restrictions on driving areas. Therefore, the Shenzhen Management Regulations are essentially a contraction of "fully autonomous driving."

For highly automated driving L4, the "Shenzhen Regulations" also have a certain degree of "definition downgrading", requiring that there must be someone in the driving seat of an L4 vehicle, and no unmanned or remote driving is allowed. This can basically be regarded as the same level of handling as L3 level automated driving.

In addition, considering that Shenzhen’s smart connected vehicles can be sold to ordinary consumers, the corresponding determination of traffic accident responsibility will also be different from the determination of rights and responsibilities in the "Notice".

The main principle is "whoever benefits shall be held responsible", with whether there is a driver in the car as the primary criterion for judgment: as long as there is a driver in the driving seat, the responsibility and compensation of the smart connected car shall be borne by the driver first; if it is a fully autonomous driving L5 vehicle, there are only passengers in the car but no driver, then the vehicle owner or manager shall bear the relevant responsibility and compensation.

This principle for determining accident responsibility is very simple and clear, which makes it easy for traffic management departments to quickly locate the person responsible for the accident, avoid disputes among drivers, car owners, car companies and managers, and provide help and compensation to accident victims as soon as possible.

It can be seen that the vehicle's autonomous driving level is not the main basis for judgment. Whether the vehicle has design defects and functional failures, and whether the autonomous driving function was activated at the time of the accident, will be judged based on the information from the on-board equipment, roadside equipment and regulatory platform. The driver or owner can seek compensation from the manufacturer and seller based on the final conclusion of the accident analysis.

Unlike the "Notice" which saw some car companies take the lead in taking the lead half a year after its release, it has been two years since the "Shenzhen Regulations" were released. As of now, no car company has officially announced that one of its cars can be included in the smart connected vehicle product catalog in Shenzhen, and publicly sold with L3 or even higher levels of autonomous driving functions.

As autonomous driving technology becomes increasingly sophisticated, it is equally important to handle the issues of social integration and improving social acceptance. To a large extent, these leading cities are leveraging their advantages in policy environment, business environment and urban construction to undertake the task of exploring replicable experiences.

Since the beginning of this year, many car companies have been trying to apply for pilot projects in the "Notice" and "Shenzhen Regulations" at the same time. The two complement each other and can fully prepare and support the eve of large-scale application of the autonomous driving industry.(The author of this article is an intelligent driving engineer at a certain OEM)