Autonomous vehicles in China

A. National policies before the National Road Testing Guideline

Traffic matters are governed primarily by a national law, namely the PRC Road Traffic Safety Law, supplemented by a number of implementing rules, national guidelines and provincial or municipal regulations in China. To date, China has no comprehensive regulatory framework for AVs. While the National Road Testing Guideline has been published, it remains a subject of heated debate how self-driving cars should fit into the traditional transportation laws, product liability laws, etc.

Prior to the issuance of the National Road Testing Guideline, several policies and plans on this topic have been issued in 2017 by the State Council (the central government of China) and the primary industrial regulators, i.e. the National Development and Reform Committee (“NDRC”) and the Ministry of Industry and Information Technology (“MIIT”), evidencing the government’s determination to accelerate the development of intelligent vehicles at national level.

The State Council called for research on artificial intelligence and cultivation of an intelligent economy in a national plan in the middle of 2017 that encompasses development of self-driving technologies and intelligent vehicles.

Pursuant to that call, the NDRC and MIIT issued several action plans in the last quarter of 2017, including:

• the Three-Year Action Plan to Enhance the Core Competitiveness in Manufacturing Industry (2018-2020) issued by NDRC on November 27, 2017;
• the Implementation Plan for the Commercialization of Key Technologies for Intelligent Vehicles issued by NDRC on December 13, 2017;
• the Three-Year Action Plan for Bolstering the Development of the Next Generation Artificial Intelligence Industry (2018-2020), issued by MIIT on December 14, 2017;
• the Guidelines on Establishment of the National Standard System for Telematics Industry (Intelligent & Connected Vehicles) jointly issued by the MIIT and the Standardization Administration of China on December 29, 2017.

The NDRC includes intelligent vehicles as a key sector in its action plan and sets forth a number of key tasks for the commercialization of intelligent vehicle-related technologies. The NDRC also is committed to supporting and providing financial aid to qualified projects in this sector.

On the other hand, MIIT aims to establish a comprehensive system of national standards for AVs, such as terms and definitions relating to AVs, functional evaluation standards, information security standards and information perception standards. MIIT seeks to promulgate at least 30 key national standards by 2020, that will support AVs with driver assistance functions and low-level automated driving functions, and to develop a more comprehensive system with more than 100 national standards by 2025 geared to support high-level automated driving.

(i) The National Road Testing Guideline

On April 3, 2018, the MIIT, the Ministry of Public Security (the “MPS”) and the Ministry of Transportation (the “MOT”) jointly issued the Administrative Rules for Road Testing of Intelligent and Connected Vehicles (for Trial Implementation), i.e. the National Road Testing Guideline. The National Road Testing Guideline was promulgated to introduce a nationwide legal framework for testing AVs on public roads. It took effect on May 1, 2018 and aims to facilitate the development of automated driving technology through the wide deployment of public road tests.

Key points of the National Road Testing Guideline are set out as follows:

1. Definition of Intelligent and Connected Vehicle

   The National Road Testing Guideline defines the “intelligent and connected vehicle” as a new generation vehicle that is equipped with advanced car-borne sensors, controllers, actuators and other devices in combination with modern communication and network technologies, which can ultimately replace the operation by human drivers and achieve safe, efficient, comfortable and energy-saving driving. AVs should be capable of, among other things, intelligent information exchanging and sharing between the vehicle and humans, other vehicles, roads and cloud servers, perceiving complicated surrounding conditions, intelligent decision-making and collaborative control.

   The automation functions of AVs are divided into three different levels, namely conditional automation, high-level automation and full automation. Conditional automation is the driving mode where the system performs all driving tasks and the driver needs to intervene when requested by the system; high-level automation is the driving mode where the system performs all driving tasks and may request the driver to respond in certain circumstances but the driver may ignore such requests; and the full automation is the driving mode where the system performs all driving tasks that a human driver can perform under all road conditions without any intervention of the driver. These are generally understood to refer to L3, L4 and L5 under the definition of levels of automation as outlined by SAE International.

2. Testing Procedures and Requirements

   Before an AV can be tested on roads, a test permit (described in more detail below) must be obtained from the authority. The local counterparts of the MIIT, the MPS and the MOT at the provincial or municipal level are jointly responsible for administration of AV tests and issuance of test permits for AVs.

   The following requirements must be complied with in order to obtain a test permit from the authority:

   Requirements of the testing entity

   The testing entity shall be an independent legal entity registered in China that has necessary technical and financial capability to, among others, manufacture vehicles and their components, conduct related research and development activities, monitor the test vehicles remotely on a real-time basis, record, analyze and reproduce an incident involving the test vehicles, compensate the losses caused by the test vehicles.

   Before being permitted to test on public roads, it must complete certain tests as required by the authority in a closed test field. It shall take out traffic accident insurance with an insured amount of at least 5 million Yuan (approximately USD $750,000) or provide a letter of guarantee of the same amount for each test vehicle.

   Requirements of the test vehicle

   Test vehicles, including passenger vehicles and vehicles for commercial use but excluding low-speed vehicles and motorcycles, shall meet the following requirements.

   First, the test vehicle should not yet be registered with the authority but must satisfy all statutory inspection and testing requirements except for endurance requirement. If any statutory testing requirement is not satisfied due to the automation function, the entity applicant must prove that the safety of the vehicle has not been jeopardized.

   Second, the test vehicle shall be equipped with an autonomous driving system and have the function to switch between the autonomous driving mode and the manual driving mode safely, immediately and easily. The test driver shall be able to intervene and control the vehicle directly at any time under the autonomous driving mode.

   Third, the test vehicle shall have status recording and storage as well as online monitoring functions, which enables the real-time transmission of information relating to the driving mode, the location and the movement of the vehicle, and which can automatically record specified data during the period of at least 90 seconds prior to a traffic accident or malfunction of the test vehicle and store such data for at least 3 years.

   Fourth, the test vehicle must complete sufficient tests in a closed field, and its self-driving function must be tested and verified by a third-party testing institution recognized by the authorities.

   Requirements of the test driver

   The test driver shall have at least three years of unblemished driving experience with no record of drunk or drugged driving, no severe traffic violation record (e.g. speeding 50% over the speed limit or violation of traffic lights) for the recent one year, and no traffic accident record of causing death or serious bodily injury. It is also required that the test driver shall enter into an employment contract or a labor service contract with the testing entity. In addition, the test driver shall have a good technical understanding of the self-driving testing program and operation methods and have the capacity to deal with the emergency situations.

   The testing entity shall submit relevant materials to the authority evidencing that the above requirements are complied with, and the authority will decide whether to grant a test permit in respect of each test vehicle, which will be valid for no more than 18 months. After the testing entity receives the test permit from the authority, it shall apply for a plate for the test vehicle. If any information shown on the test permit such as the testing entity, the test vehicle or the test driver is changed, the testing entity shall re-apply for a test permit.

(ii) Local rules and regulations

Local transportation authorities in Beijing, Shanghai, Chongqing, Changsha, Guangdong, Fuzhou Pingtan, Changchu, Tianjin, Zhejiang, Jinan, Wuhan and Jiangsu have promulgated local rules to further regulate AVs in their own regions:

• the Beijing Administrative Rules on Acceleration and Promotion of Work relating to Road Testing of Autonomous Vehicles (for Trial Implementation) issued on December 15, 2017; 
• the Shanghai Administrative Measures on Road Testing of Intelligent and Connected Vehicles (for Trial Implementation) issued on February 27, 2018;
• the Chongqing Administrative Rules on Road Testing of Autonomous Vehicle (for Trial Implementation) issued on March 14, 2018.

These local rules contain similar but more detailed requirements in respect of the testing entity, the test vehicle and the test driver to the National Road Testing Guideline, with local nuances. For instance:

• The Fuzhou Pingtan rule allows an independent legal person registered in Taiwan to conduct road testing.
• Under the Changsha rule, a testing entity shall take out safety production liability insurance.
• The Chongqing rule prohibits the test driver from working for more than two consecutive hours or working for more than 6 hours per day, and a dozen local rules require the test driver to have a minimum of 50 hours’ experience in operating autonomous driving system.

Applicants for road testing permits must comply with both the National Road Testing Guideline and the relevant local rules.

(iii) NDRC draft strategy

On January 5, 2018, the NDRC issued the Strategy for Innovation and Development of Intelligent Vehicles (Draft) (“Draft Strategy”) for public comments, which marks a further step of the government towards its goal of promoting AVs.

The Draft Strategy envisages that by 2020, a systematic framework for China will be in place for technology innovation, industrial ecosystem, infrastructure network, regulations and standards, product regulation and information security. The Draft Strategy aims to massively develop AVs in China and sets an ambitious goal that by 2020, all new vehicles are expected to have partial or full autonomous driving capabilities and 90% of LTE-V2x, and by 2025, China hopes to have almost 100% of new vehicles as AVs, and a full regulatory regime and industry specifications are expected to be fully established and by 2035; China will become an AV superpower.

The Draft Strategy recognizes the following tasks for the development of intelligent vehicles in China:

• promoting an independent and controllable technology innovation system for intelligent vehicles;
• creating an inter-sector and integrated industrial ecosystem for intelligent vehicles;
• setting up an advanced and complete road infrastructure system for intelligent vehicles;
• formulating further regulations and standards for intelligent vehicles;
• building up a scientific and normative product regulation system for intelligent vehicles;
• building up a comprehensive and efficient information security system for intelligent vehicles.

B. Data protection and data privacy

AVs contain various sensors that are designed to collect massive data of the vehicle’s operation and user’s preference as well as its surroundings. The sensors generally are cameras, radar, thermal imaging devices and “LIDAR,” and will collect data such as statistics, photos and videos. With the development of AVs, the concerns of data privacy and unreasonable disclosure of personal information go high.

The map for autonomous driving is a new type of electronic navigation map (“ENM”), and data collection, editing, processing and production of automatic driving maps can only be undertaken by an entity holding an ENM license issued by the restructured Ministry of Natural Resources of the PRC (“MNR”). Foreign investors are prohibited from making ENM. Where an ENM license holder cooperates with automakers in developing and testing maps for autonomous driving, the relevant surveying and mapping activities shall be conducted by the ENM license holder alone. Without the approval of the provincial branch of MNR, mapping data generated from autonomous driving technology testing or road testing (including adding contents, elements or precision to traditional ENM) shall not be provided to or shared with foreign entities or individuals or foreign-invested enterprises incorporated in China (including wholly foreign-owned enterprises and joint ventures).

China’s Cybersecurity Law, which became effective as of June 1, 2017, and a series of underlying rules, regulations, guidelines and industry standards, have imposed new regulatory requirements in terms of data privacy and data protection. These new legal requirements will have significant implications for industry players in the AV industry in relation to the collection, use, processing and cross-border transfer of data. In particular:

• “Personal data” is defined as all kinds of information recorded in an electronic or other form, which can be used, independently or in combination with other information, to identify a natural person’s personal identity, including but not limited to the natural person’s name, date of birth, ID number, biology-identified personal information, address and telephone number. Furthermore, “biometric data,” including genes, fingerprints, voiceprints, palm prints, pinna, iris and facial recognition features, would fall within the category of “sensitive personal data.”
• The collection and use of personal data must follow the principles of legality, proportionality and necessity. The data controllers shall expressly indicate the purposes, means and scope of data collection and use, and obtain prior explicit consent (rather than implied consent) from relevant sensitive personal data subjects.
• The data controllers are recommended to pseudonymize any personal data they receive.
• For the sharing and transfer of sensitive personal data, the data controllers are required to give notice to the data subjects (as to the purpose as well as the type, identity and data security capability of the data recipients), obtain the data subjects’ explicit consent, and adopt security measures (e.g. encryption).
• Any personal data and critical data collected and generated during the business operation of critical information infrastructure (“CII”) operators within China shall be stored in China and not transferred abroad, unless any data export is made on the ground of business necessity and has passed a security assessment by competent authority, the network operators themselves or delegated third parties (as the case may be). CII covers a wide range of sectors including public communication and information services, energy, transportation, water conservancy, finance, public services and e-government, as well as “other infrastructure, that, in the event of damage thereto, loss of function thereof or leak of data therefrom, could seriously jeopardize national security, national economy or public interest of China.”

C. Intellectual property

The increasing research and development in the fields of biometrics and autonomous driving has inevitably led to the birth of a lot of novel, inventive and practically applicable technologies. These technologies may be applied for and protected as patents in China under the Chinese Patent Law. The Chinese Patent Law was only first promulgated in 1984, but the Chinese patent landscape has developed rapidly over the past 35 years to support the evolution of new technologies in China.

Under the Patent Law of China, an invention, which is the subject of a patent application, must not be an existing technology, i.e. a technology known or disclosed to the public either locally or abroad. As compared with the technology existing before the date of application, an invention must have prominent substantive features and represent a notable progress, and a utility model must have substantive features and represent progress. The term of protection is twenty years for an invention patent and ten years for a utility model, counted from the date of application.

In 2017, the Guidelines for Patent Examination were amended and the requirements for claims for software-related inventions were relaxed. Accordingly, it now seems to be the case that a computer-readable medium having instructions for performing a technical method may seek patent protection. Many software innovations incorporated into AVs in China may be eligible for patent protection.

Autonomous driving

The China Patent Protection Association published the “In-depth Analysis Report on Patents for Artificial Intelligence Technology” in November 2018. The top five applicants for patents in the field of autonomous driving in China were named as being Baidu, Ford, Toyota, Dajiang (also known as DJI) and Beihang University. Baidu is an Internet company, which invested heavily in research and development on autonomous driving projects in recent years. Ford and Toyota are existing well-known vehicle manufacturers. DJI is a manufacturer of unmanned aerial vehicles (drones) for aerial photography and videography. Beihang is a Chinese University specialized in scientific research in this field. It can be seen that the industry players in the patents for autonomous driving are not restricted to the experienced vehicle manufacturers, but also manufacturers of hi-tech autonomous driving devices, Internet companies which are very good at analyzing data, as well as academic or research institutes.

Since Baidu is an extensive filer for Chinese patent applications in the field of autonomous driving, it is worth studying what patent applications Baidu has filed for over the past few years. The Chinese Patent Registry published an analysis of Baidu’s patent applications relating to autonomous driving on its official website in March 2018. According to the analysis, Baidu filed for its first autonomous driving-related patent in 2014, but since then the number of patent applications has increased rapidly. Baidu has filed patent applications in various technologies, with the main focus on environmental awareness (e.g. laser radar and image acquisition devices which help to detect obstacles and lane lines), operational control (e.g. switching and control of driving mode), and trial and evaluation. The report commentator was of the opinion that Baidu was relatively weak in relation to the technology of planning and decision-making and also data fusion using data gathered from multiple sensors.

Biometrics

According to a recent report published by Frost & Sullivan, biometric technology in China may be classified into six categories, namely face recognition, fingerprint recognition, vein recognition, gait recognition, iris recognition and voiceprint recognition. According to the Sullivan data, the market size for biometric identification in China has expanded steadily, and the market size in China in 2017 increased by 38.4% compared to 2016.

Face recognition technology has become more widespread in China and related technologies such as computer and optical imaging have also developed rapidly. According to the statistics published in October 2018 by the China Business Research Institute, the market size for Chinese face recognition reached RMB 2.91 billion in 2017. From 2014 to 2017, the number of patent applications for face recognition technology in China continued to increase, with an average annual growth rate of 36%. In 2017, the number of face recognition patents in China reached 2,698.

Iris recognition technology has also developed rapidly in China in recent years. According to the China Intellectual Property Right Net, Chinese applicants started to file for patents relating to iris recognition technology worldwide around the same time as applicants from South Korea, the United States and Japan. The number of worldwide patent filings by Chinese applicants was quite low in the early years, but the number of worldwide patent filings by Chinese applicants increased drastically after 2006 and has since occupied the largest portion. Among the top ten patent applicants in the global market (by volume), six of the applicants came from China. As of November 2016, it was revealed that 93% of the applicants filing patents relating to iris recognition technology in China were domiciled in China. These Chinese applicants are mainly research institutes and enterprises in Beijing, including China Institute of Automation, Chinese Academy of Sciences, China University of Science and Technology, Beijing Tiancheng Shengye Technology Co., Ltd and Beijing Zhongke Hongba Technology Co., Ltd, etc. This record of patent applications shows that shows that Chinese corporations are keen to develop biometrics technology and are aware of the benefits of seeking patent protection not only in the domestic market but also globally.

Integration of biometrics to autonomous driving

Sullivan suggested that there is a trend of integrating different biometric technologies rather than relying on a single biometric technology in order to improve security. For example, Chinese manufacturers have integrated palm and palm veins, fingerprints and faces into the same identification. The use of multiple biometrics technology can achieve better recognition performance and reliability than a single biometrics technology, increase the difficulty of forging human biometric features and improve product safety. Nevertheless, composite biometrics technology is not a simple additive integration between biometrics, and it is necessary to develop new algorithms based on the characteristics of different biometrics to achieve geometrical improvements in computational efficiency and accuracy. Similarly, it is not surprising to see how biometric technology will be integrated to the field of autonomous driving for the convenience of the public.

There appears to be little overlapping to date between the key players in the fields of biometrics and autonomous driving. It is anticipated that different companies will cooperate and obtain the right to use other biometric or autonomous driving technologies through licensing or acquiring other technology companies. There are provisions restricting the import and export of technologies in China. A Chinese citizen or legal entity which grants a license or assigns its patent to a foreign company will be subject to the Regulations of Administration of Import and Export of Technology in China. For example, the import and/or export of technologies may be restricted in order to protect the public interest or enable acceleration 
of development of certain industries locally. It remains to be seen how different entities will cooperate in order to integrate these technologies.

D. Challenges to insurance industry

General auto insurance regime in China will apply to AVs, but no doubt the widespread adoption of AVs will have a great impact on the automobile insurance industry. For example, insurance costs are expected to shift from the individual car owners to the automobile manufacturers gradually because the automakers will likely be held accountable for accidents occurred during the self-driving mode. Insurance premiums will drop considerably, since accidents will decrease as human drivers will make fewer mistakes with the assistance of the automated system. Commercialization of artificial intelligence and big data technologies and mass production of AVs in the near future will have far-reaching consequences for insurance businesses in China. Currently, however, given the early stage of the development and testing of AVs, it has been provided in the National Road Testing Guideline that an AV testing entity must take out traffic accident insurance with an insured amount of at least 5 million Yuan (approximately USD $750,000) or provide a letter of guarantee of the same amount for each test vehicle.

E. Conclusion

Recent developments in the sector are welcomed by the industry and clearly show China’s determination and commitment to bolster the AV sector. The national and local road testing guidelines and rules represent a firm step towards an upgraded and intelligent automobile industry. It is expected that more regulations and national standards will be promulgated shortly. Testing permits for AVs have been granted to large Chinese and international automotive and tech companies, such as Alibaba, Baidu, Tencent, Shanghai Automotive, BMW, etc., in multiple cities in China in 2018, and we expect more testing permits will be issued. Interaction between new technologies and traditional laws may present both opportunities and challenges for the industry players and they should keep a close eye on future developments.