Autonomous Driving Lands in Multiple Cities: How Difficult to Commercialize
China News Weekly reporter / Jiang Zhiyu
Published in the 1060th issue of "China News Weekly" magazine on September 12, 2022
Autonomous driving, which is regarded as one of the most difficult technologies to implement in the field of artificial intelligence, is receiving policy support from many Chinese cities.
Due to the limitations of the bottleneck of autonomous driving technology, the autonomous driving industry once "ebbed" after experiencing an early investment boom.
Industry insiders believe that it will take 5-10 years for autonomous taxis to land on a large scale, and commercialization is also slow.
Due to the long-term "testing" and "demonstration application" stage, autonomous vehicles lack the commercial operation qualifications in terms of policies, and enterprises have never "charged", making it difficult to enter the commercialization stage.
Beginning in 2021, many autonomous driving startups will try to commercialize unmanned taxis by looking for partners such as online car-hailing platforms, so as to promote the iterative upgrading of technology and large-scale promotion and application through commercialization.
However, since the beginning of this year, a series of policies have "opened up" to the commercialization of autonomous driving.
Among them, the “Guidelines for Transportation Safety Services for Autonomous Vehicles (Trial)” (Draft for Comment) issued by the Ministry of Transport on August 8, proposed to encourage the use of autonomous vehicles in specific scenarios on the premise of ensuring transportation safety. Engaged in taxi, bus passenger and general cargo transportation.
The industry generally believes that this is a positive statement from the transportation regulatory authorities on the commercialization of autonomous driving.
The commercialization of autonomous driving has also accelerated significantly in places.
On July 20, Beijing officially opened China's first pilot pilot for the commercialization of driverless travel services.
In August, Chongqing and Wuhan issued a pilot policy for fully unmanned commercialization of autonomous driving, allowing autonomous vehicles without a safety officer in the vehicle to carry out commercial services on social roads, and issued the first batch of unmanned demonstration operations in the country to Baidu qualifications.
All regions are competing for commercialization pilots and the degree of openness of policies. There are policy directions and economic benefits behind them.
Autonomous driving is not only a competition between enterprises, but also a contest between regions.
Multiple problems of large-scale landing
From Suzhou North High-speed Railway Station, passengers are likely to order a self-driving car through an online platform.
It may have a distinct appearance from other vehicles—a larger sensor in the shape of an umbrella on the roof, or it may look the same as a normal vehicle.
The "Yangtze River Delta Intelligent Driving Industry Demonstration Zone" in Xiangcheng District, Suzhou is located in the northern circle of Suzhou about 10 square kilometers. More than 500 autonomous vehicles are tested and applied on the road.
Passengers can call Robotaxi and Robobus through the online platform, and get on and off at the dedicated autonomous driving station. There is no charge yet.
Suzhou is an early city to realize the normal operation of self-driving buses. In Suzhou, passengers can take self-driving buses through the "Light Boat Travel" applet. There are currently 7 stations including Jiajing Tiancheng Garden and Yucheng Experimental Primary School. The bus line named "Qingzhou-Su Line 11" operates at 9:00 every day and ends at 16:30, and is positioned as a short-distance connection of 3 kilometers.
"Qingzhou Travel" was developed by Qingzhou Zhihang, an unmanned general solution company. Its co-founder and CEO Yu Qian told "China News Weekly" that Qingzhou Zhihang will land in Suzhou in 2020, from the initial dozens of employees to There are currently close to 400 people.
Qingzhou Zhihang’s self-driving bus for urban open roads, Longzhou ONE, has been operating in Suzhou High-speed Railway New City for two years, with a total of 8 lines opened.
Yu Qian said that it will take 5~10 years for Robotaxi to be commercialized on a large scale, and Robobus may only take 2~3 years to land due to the fixed operation route, so the application scenario is the first to fall on Robobus.
The difficulty of commercializing autonomous driving is mainly due to technical problems.
Yu Qian believes that there are still many technical breakthroughs in the widespread popularity of Robotaxi. Even Waymo can only achieve full unmanned driving in a limited area.
Waymo is the self-driving company with the longest test miles in the world.
According to the 2021 self-driving road test report released by the California Department of Transportation (DMV), Waymo, which ranks first in test mileage, ran 2.32 million miles throughout the year.
In the annual road test data released by the California DMV, MPI (Miles Per Intervention, the average number of miles traveled between two manual takeovers) is considered to be one of the key indicators to measure the level of autonomous driving technology.
In 2021, Waymo’s safety officers have taken over 292 times, and the MPI is 7,965 miles, which means that it needs to take over once every 7,956 miles. According to the MPI comparison alone, the reliability of Waymo’s self-driving system is about 1/45 of that of human drivers.
In 2019, Waymo's MPI was 13,200 miles, and the number of takeovers did not drop in the past two years.
"The marginal benefits of traditional data collection methods are no longer significant. This is an objective technical issue," said Di Di, head of the strategy of China Zhixing, an autonomous driving technology and travel service provider. But the long-tail problem of 90 to 100 is difficult to break through, and an increase of 1% will require a tenfold or a hundredfold price.”
Technical difficulties will limit commercialization.
"The main benefit point of shared autonomous driving in the future is to replace part of the income of human drivers." Didi said.
According to the 2021 annual report of Didi Chuxing, an online car-hailing platform, drivers’ remuneration and subsidies already account for about 80% of the net fare.
Replacing human safety officers with autonomous driving systems is the only way to commercialize, but the autonomous driving industry has been slow to "remove" the safety officers from the car.
Didi believes that in addition to technical issues, economic benefits are also a problem for autonomous driving companies, "L4-level autonomous driving requires equipment such as lidar, millimeter-wave cameras, and large computing power chip units. Most of the manufacturing and modification costs are more than 500,000 yuan."
Founded in 2018, China Intellectual Property Corporation announced in 2019 that it will develop autonomous driving through a technical route of vehicle-road coordination.
After an early investment boom, the autonomous driving industry has calmed down due to obstacles in technology and commercialization.
In order to break the bottleneck of autonomous driving technology and commercialize it as soon as possible, two different technical routes have been developed in the field of autonomous driving.
In addition to the single-vehicle intelligent routes adopted by Waymo and Qingzhou Zhihang, there is also the vehicle-road coordinated route adopted by China Zhixing, which emphasizes the empowerment of smart roads to vehicles.
In Suzhou, the vehicle-road synergy program "Light Vehicles and Familiar Roads" jointly developed by China Zhixing and Tianyi Transportation has been implemented.
E-Surfing Transportation is an intelligent transportation operation and service provider jointly funded by China Telecom, CITIC Bank and Suzhou State-owned Assets Platform. It will be established in 2021.
"Light cars and familiar roads" refers to smart cars with L2 lightweight configuration, which can achieve L4 autonomous driving by utilizing the capabilities of smart roads.
In this plan, CIIC is responsible for the overall solution of vehicle-side autonomous driving, and Tianyi Transportation is responsible for roadside construction.
Under the traditional bicycle route, the perception completely relies on the sensors of the autonomous vehicle itself, but in the "light vehicle and familiar road" scheme, even if all the sensors on the vehicle end are covered and only rely on the perception of the road side, the car can still obtain the road side through the on-board 5G communication module. Intelligent information to make driving decisions.
"As a 'plug-in', smart roads can make up for the lack of bicycles in terms of technical safety." Didi said, "In terms of economic benefits, some sensing and computing devices are placed on the road side, and the vehicle side is relatively lightweight. Retaining lightweight sensors in the middle of the road makes the vehicle and road intelligence redundant, and it can also reduce the overall cost and increase efficiency of intelligent transportation.”
In the core business district near the Suzhou North Station of the High-speed Rail New City, Tianyi Transportation has deployed a "holographic road" with 12 intersections and a length of 6 kilometers. Relying on nearly 200 different types of sensors, it is digitally coordinated with the back-end vehicles and roads. The twin platforms are connected, which can track and record the status of all traffic participants in the road in real time in real time, and transmit the road conditions to the matching autonomous vehicles to realize "vehicle-road coordination".
However, "Light Cars and Familiar Roads" have not yet disclosed the road test mileage and takeover situation.
Some people in the industry believe that the main body of responsibility in vehicle-road coordination is not clear, the decision-making contradiction between the vehicle end and the road end is difficult to resolve, and it will take a long time to verify the true collaboration between the vehicle and the road.
Some people in the field of vehicle-road coordination pointed out that under the current technical conditions, the validity of the data at the vehicle end and the road end is mainly judged by confidence, and then it is decided whether to "listen to the car" or "listen to the road".
However, this system still requires massive collection and verification of big data to improve the accuracy of system decision-making.
This also requires the continuous running-in of vehicle-road collaboration and autonomous driving companies, and a large amount of test data verification to improve the credibility of decision-making.
Why are the government and enterprises so "enthusiastic"?
Although there are still technical difficulties in the large-scale implementation of autonomous driving distances, autonomous driving companies are already seeking commercialization pilots.
"Allowing self-driving passenger cars to carry out operating charges is the beginning of commercialization." Didi said, "Autonomous driving must run in a certain area, from any point to point, normal operation, and business model can be allowed to land, Only then can we in turn promote the iterative upgrading of technology and large-scale promotion and application.”
Since commercial operations require licenses, and autonomous driving companies do not have relevant qualifications, from 2021, autonomous driving companies will “save the country from the curve” by cooperating with third-party travel platforms.
In August this year, Pony.ai, an autonomous driving company, announced that Pony.ai Robotaxi will be launched in Cao Cao’s Beijing area.
Qingzhou Zhihang, China Zhixing and other companies are also planning to jointly operate Robotaxi in Suzhou with the car-hailing platform.
However, partnering with a ride-hailing platform has its downsides.
An industry source who did not want to be named said that this means that outside the automatic driving section, the need to restore the manual driving of online car-hailing drivers is actually a hybrid operation model, which is not completely commercialized for autonomous driving companies. , the boundary of responsibility is not clear, and the user experience is not good.
Although the bottleneck of autonomous driving technology still exists, the progressive policies introduced from the national to the local level also mean support for the autonomous driving industry.
On August 8, the Ministry of Transport issued the "Guidelines for Transportation Safety Services for Autonomous Vehicles (Trial)" (Draft for Comment), which proposed to encourage the use of autonomous vehicles in specific scenarios to engage in taxis, Bus passenger transport and general cargo transport.
Cheng Xiuyuan, vice president of Qingzhou Zhihang, told China News Weekly that the "Guide" encourages autonomous driving companies to conduct commercial pilots with a prudent attitude, which is conducive to promoting the closed-loop commercialization of the autonomous driving industry.
Xie Liangbing, dean of the Standard Ranking City Research Institute, previously stated that the previous competition among major cities in China revolved around striving for a national-level intelligent networked vehicle test demonstration area, the promulgation of autonomous driving-related policies, the issuance of test licenses, and the construction of test roads. , and now, all regions have completely entered the competition stage for the commercialization of autonomous driving.
Allowing the commercial operation of local autonomous vehicles is seen as a signal that local governments support the development of autonomous driving.
On July 20, Beijing officially opened the first pilot commercialization of the “driverless” travel service in China.
Immediately afterwards, Chongqing and Wuhan, unwilling to lag behind, announced the commercialization of "fully unmanned" autonomous driving pilots.
In addition, Shenzhen, Pingtan, Xiamen, Changzhou, Zhengzhou and other places have also repeatedly reported new progress in the commercialization of autonomous driving.
Beijing established a high-level autonomous driving demonstration zone in September 2020. Kong Lei, deputy director of the Beijing Economic and Technological Development Zone Management Committee, serves as the director of the demonstration zone work office.
According to industry insiders, compared with the "part-time" status of the autonomous driving working groups in some cities, the Beijing Self-Driving Office is an entity with a "high level".
At present, the commercialization pilot in Beijing is in the stage of "unmanned driving". Baidu and Pony.ai have become the first batch of licensed companies. They will invest 30 unmanned vehicles in the core area of the Economic and Technological Development Zone within 60 square kilometers. Normalized charging services.
The competition for the first city of the fully unmanned commercialization pilot is also very fierce.
"Chongqing Daily" released a report on August 8, saying that it learned from Baidu that China's first full-vehicle unmanned demonstration operation qualification was issued in Yongchuan.
Phoenix.com Hubei issued a document on the same day that Wuhan government departments issued Baidu the first batch of unmanned demonstration operation qualifications in the country.
According to the latest "Guidelines for Transportation Safety Services for Autonomous Vehicles (Trial)" (Draft for Comments) issued by the Ministry of Transport, fully autonomous vehicles engaged in transportation operations should be equipped with remote drivers or safety officers.
At present, Yongchuan's Carrot Run has allowed "no one in the car", but it is also equipped with remote security personnel.
Although many places are scrambling to introduce policies to encourage the commercialization of autonomous driving pilots, there is still a distance from the real "commercialization".
Cheng Shidong, director of the Urban Transportation Research Office of the Comprehensive Transportation Research Institute of the National Development and Reform Commission, believes that the current commercialization of autonomous driving is only a "symbolic fee", and a large-scale mature business model has not yet been formed, and it is still in the testing stage.
Cheng Shidong suggested that there is no need for various regions to compete for commercialization pilots in terms of policies.
"Autonomous driving companies are not yet profitable, and the impact and drive on economic benefits is not obvious. Companies do not need to go to so many cities for road testing, tropical, temperate, mountainous, plain, and some cities that meet specific scenarios are enough, mainly for testing. The state of unmanned driving in different scenarios.”
Didi also believes that it should be "penetrated in a single city".
"CITIC's grand strategy is to penetrate through a single city and replicate in multiple cities. There must be breakthroughs at a single point. Piloting dozens of kilometers in a certain city is not a real commercialization." Didi said, "There is a The term is called ODD (operational design domain, operational area), and it must be fully verified in this area and have commercialization capabilities before going out.”
However, in promoting the commercialization of autonomous driving, local governments and enterprises have hit it off, and they are looking at a bigger market cake in the future.
In order to promote the development of autonomous driving, the Suzhou Municipal Government established the state-owned platform Pioneer (Suzhou) Digital Industry Investment Co., Ltd. Its chairman Wang Jiali believes that intelligent networked vehicles involve multi-disciplinary, cross-border, subdivided industrial chains, and strong industrial drive. .
Some people in the industry believe that in the context of the saturation of traditional infrastructure projects, the smart road may become the outlet for new infrastructure investment.
In May 2021, the Ministry of Housing and Urban-Rural Development and the Ministry of Industry and Information Technology jointly announced the pilot program of "Double Intelligence" (the coordinated development of smart city infrastructure and intelligent connected vehicles), which is considered to be a new round of "city building movement".
Under the tuyere, the construction of autonomous driving test roads in various places has already entered a white-hot stage.
Wang Jiali revealed that on the road side, Suzhou City has built a 63.4-kilometer autonomous driving open area, and is also building a 98.6-kilometer-long third-phase test road. After completion, there will be about 160 kilometers of autonomous driving test routes.
In June 2021, Chongqing announced that 60 new roads in Yongchuan District and Chongqing High-tech Zone with a total of 116.76 kilometers will be opened for autonomous driving tests.
In August this year, the Shenzhen Municipal Bureau of Transportation announced that it will add 56.68 kilometers of test roads. So far, the accumulated test road mileage for intelligent connected vehicles has been about 201.37 kilometers.
In September, the Beijing High-level Autonomous Driving Demonstration Zone Office announced that the construction area would be expanded from 60 square kilometers in the core area of the Economic and Technological Development Zone to 500 square kilometers in the city.
According to people in the field of vehicle-road collaboration, smart roads depend on the government’s investment in 5G new infrastructure. After the pilot is successful, companies will promote the project and then participate in government bidding to get the project.
Some self-driving companies have already jumped on the bandwagon.
The autonomous driving full-stack technology and operation service provider Mushroom Chelian signed a strategic cooperation with Sichuan Tianfu New District on July 28, with a total project value of 3 billion yuan.
Previously, the Mushroom Car Alliance had also implemented hundreds of millions of autonomous driving projects in Dali, Yunnan, Hengyang, Hunan, Tongzhou, Beijing and other places.
Baidu Apollo also won the bid for the “new infrastructure” project of smart transportation in Huangpu, Guangzhou in 2020. The project value is close to 460 million yuan, and this year, it won the new urban infrastructure construction of Guangzhou Huangpu “Smart +” Checheng Network with 503 million yuan. project.
The demonstration area should be placed and more "managed"
As autonomous driving accelerates from road testing to commercial operation, how to support system management measures while developing rapidly is also a problem that needs to be solved in autonomous driving demonstration areas in cities.
In 2019, Suzhou issued the first autonomous driving road test license.
Due to the integration and mutual recognition of the Yangtze River Delta region, they are all tested at the Wuxi Jiaotong Institute, and China Intellectual Property Corporation also obtained relevant licenses in 2019.
Didi said that the license has 17 tests, each of which must be tested 30 times in a row, and the road test license can only be obtained after consecutive passes.
Wang Jiali, chairman of Pioneer, pointed out that at present, the authority for issuing road test licenses in Jiangsu Province has been delegated to various prefecture-level cities, and in the future, third-party testing institutions can also conduct certification tests in Suzhou.
Due to the limitation of the city's energy level, Suzhou has not yet issued a commercialization pilot policy, and local autonomous driving enterprises are in the "demonstration application" stage, that is, the free trial operation stage for non-specific public.
Wang Jiali said that Suzhou has a set of access standards for "demonstration applications", and autonomous vehicles need to pass a total of 2,000 kilometers of simulation test virtual scenarios before they can operate with people.
In addition, Suzhou is applying for a high-quality automobile development pilot zone in Jiangsu Province, and the pilot zone may make breakthroughs in regulations and policies in the future.
In the technological process of autonomous driving, "removing the safety officer" is a particularly critical sign.
Wang Jiali believes that Suzhou considers piloting unmanned operation of low-speed operation vehicles such as sanitation vehicles and logistics vehicles. Because their operation time is usually in the early morning and at night, they are not motor vehicles, and the safety risk is small.
For Robotaxi and Robobus, unmanned operations will be more prudent due to the current technical bottlenecks.
The Beijing Demonstration Zone has set a clear stage for the development of Robotaxi, and has successively established four stages: manned testing, highway testing, unmanned testing and commercialization pilot.
In the process of unmanned testing, Beijing clearly divides the autonomous driving test into three stages according to the presence or absence of safety officers.
According to the "Beijing High-level Autonomous Driving Demonstration Zone Development Report", the first stage is for the safety officer to move from the main driver to the passenger seat, that is, "the main driver is unmanned", and the safety officer only passes the installed braking device when necessary. Take over to ensure the safe operation of the vehicle.
In the second stage, the safety officer will be moved from the co-pilot to the rear.
The third stage will completely cancel the in-vehicle safety personnel and enter the real unmanned test stage.
Another problem facing self-driving companies is that standards are not uniform across cities.
Taking road test licenses as an example, the road test licenses issued by some cities are not mutually recognized, and enterprises have to retake the test.
In addition, licenses are issued for vehicles rather than enterprises. According to industry insiders, it takes 2 million yuan to obtain an autonomous driving license in Beijing, which is a large cost for enterprises.
The cost of obtaining a Wuxi license is 500,000 yuan, and you can test 5 vehicles at a time.
Since self-driving companies are not yet profitable, they mainly rely on investment and government subsidies.
According to industry sources, some companies mainly rely on government purchases of projects to maintain operating projects that have not yet been commercialized.
Taking Suzhou as an example, the Bureau of Industry and Information Technology has a number of subsidies for road test support and scenario support. 20% of the actual road test service expenses will be subsidized.
In cities that have already allowed commercialization pilots, Beijing has clear access requirements.
Cheng Xiuyuan, vice president of Qingzhou Zhihang, told China News Weekly that the commercial operation of self-driving minibuses in Beijing needs to meet the 4,000-kilometer no-load test and the 6,000-kilometer simulated load test. ”, including the requirements for other indicators such as takeover rate, in order to obtain the commercial operation qualification for non-specific public charges.
The Beijing Demonstration Zone has specific criteria for the admission mechanism, such as mileage, number of orders, vehicles, number of takeovers, and test mileage for new vehicles.
For example, in terms of the number of orders received, within the scope of implementation, the number of orders received in the pilot phase of commercialization with the main driver is required to be no less than 20,000 times, and the number of orders received in the demonstration application stage with a passenger in the driver and no one behind the steering wheel is required to be no less than 1,000 times.
According to data disclosed by Beijing, the total autonomous driving mileage in the demonstration area is currently about 3.08 million kilometers.
However, Beijing did not disclose the takeover situation, only the proportion of autonomous driving status was announced, and the overall proportion in 2021 will remain between 70% and 80%, indicating that the proportion of manual takeovers is constantly decreasing.
In the U.S., the self-driving industry has a higher level of data disclosure.
According to the regulations of the California DMV, autonomous vehicles need to report the takeover situation to the DMV every year, and detail the location of the takeover, whether it is safe, the weather conditions at that time, and the road traffic conditions, etc., in order to obtain a new year's road test license.
DMV will also issue a report every year, revealing the test mileage and the number of takeovers of each company.
In China, no data such as the takeover rate has been released.
Cheng Shidong believes that the government or third-party agencies should access the enterprise's autonomous driving system to obtain real data such as road test mileage and takeovers.
"Whether the autonomous driving technology is qualified or not is not the same logic as the human driver's license test. It is impossible to pass several tests to determine whether the driving technology has passed the test. The data of the autonomous driving vehicle should be tracked in real time, and the technology should be judged through the real road test mileage and takeover data. mature."
The issue of data ownership may become a contradiction in the field of intelligent networked vehicles in the future.
On the one hand, in the construction of vehicle-road collaboration, smart roads will collect a large amount of information and data of urban traffic scenarios, which may conflict with personal privacy rights; on the other hand, how to supervise enterprise data is also a difficult problem.
Self-driving companies have no incentive to publish data.
Su Qing, the former head of Huawei's autonomous driving product department, once said that the real takeover data is a commercial secret for any company, and it is meaningless to compare a specific type of external data.
Some people in the autonomous driving industry say that some private data is encrypted and requires permission to transmit.
"Just like a car driven by a human driver, the traffic police will not know where the car usually goes in real time." After an accident, the autonomous vehicle will report relevant information to the supervision platform.
Yang Diange, a professor at the School of Vehicles and Transportation at Tsinghua University, recently issued an article saying that data management of smart cars should take into account both safety and development, and "should be coarse and not fine."
He believes that the technical route of smart cars is still in the exploratory stage, and the relevant data security management regulations should not be too detailed, and enterprises should be given a certain degree of freedom on the premise of keeping the basic security bottom line.
At present, basic constraints have been formed on the cross-border and data storage of smart car data, and qualification management is also adopted for the collection and storage of geographic information data. Appropriately leave room for enterprises to explore, and then form some detailed regulations after the technology is mature, otherwise the phenomenon of "one pipe and one death" will easily occur.
People close to the regulatory level revealed to China News Weekly that although companies report testing to regulators every six months, not all data will be reported to regulators because indicators such as the takeover rate involve corporate secrets.
"The government will supervise the data of test vehicles, but it is only a minimum requirement for safety precautions, and will not evaluate the technology itself." The above-mentioned person believes that "the industry needs to have an inclusive attitude towards new industries and technologies. growing up very fast."
The definition of safety responsibility needs to be improved
With the development of the autonomous driving industry, relevant laws and regulations also need to be improved.
According to data disclosed by Beijing, a total of 18 traffic accidents involving road tests of autonomous vehicles occurred in the demonstration area.
According to the driving mode, there were 6 accidents in manual driving mode and 12 accidents in automatic driving mode (including 3 accidents with responsibility and 9 accidents without responsibility).
On average, there is one accident every 171,000 kilometers.
There are high insurance requirements both at home and abroad for applicants for autonomous driving road test licenses.
The California Department of Motor Vehicles requires businesses to take up $5 million in insurance that proves they can be held liable for bodily injury, death or property damage caused by self-driving vehicles.
In China, the "Management Specification for Road Testing of Intelligent and Connected Vehicles (Trial)" requires autonomous driving companies to obtain a certificate of compulsory traffic accident liability insurance, and a certificate of traffic accident liability of not less than 5 million yuan per vehicle or not less than 5 million RMB Yuan for the accident compensation letter of self-driving road test.
However, the current laws and regulations for dealing with traffic accidents and violations are designed according to the traditional thinking that the vehicle has a driver. In autonomous vehicles, human beings no longer control the vehicle as a full-power operator, and the corresponding responsibility division is also more problematic. complex.
The "Shenzhen Special Economic Zone Regulations on the Administration of Intelligent Connected Vehicles" (hereinafter referred to as the "Regulations"), which came into effect on August 1, is China's first ICV regulation.
According to industry insiders, since self-driving vehicles have not formed a large scale, self-driving models have not been regulated as new traffic participants. Previously, traffic control regulations have been missing, and "traffic control departments are passively coping."
Cheng Shidong believes that Shenzhen, with the legislative power of the special economic zone, has issued the first regulations clarifying the handling of road violations involving autonomous driving, filling the gap in the lack of responsible subjects for autonomous driving accidents.
Zeng Ganyu, a staff member of the Wisdom Department of the Shenzhen Transportation Bureau, told China News Weekly that the "Regulations" were led by the Shenzhen Municipal People's Congress to set up a special class with the Municipal Transportation Bureau, the Municipal Public Security Traffic Police Bureau, and the Municipal Bureau of Industry and Information Technology. Legislation completed within time.
Zeng Ganyu said that the special class organized dozens of on-site symposiums, went to many places to inspect the development of enterprises, and produced a supporting research report of nearly 900,000 words by sorting out 43 existing articles in China and legislative norms in Europe and the United States.
Who is responsible in the event of an accident with an autonomous vehicle?
The "Regulations" make provisions around two types of situations: "with a driver" and "without a driver".
For the responsibility belonging to the intelligent networked vehicle, and there is a driver, the driver shall take the responsibility first; for the driver without a driver, based on the principle of "whoever benefits, who bears the responsibility", the vehicle owner and manager shall come first. Liability.
If it is subsequently discovered that the accident is caused by the defects of the car itself, the driver, vehicle owner, and vehicle manager can claim compensation from the vehicle manufacturer and seller in accordance with the regulations, and generally follow the rule of "pay first, then recover".
That is to say, for autonomous vehicles, as long as there is a safety officer on the vehicle, the safety officer will be responsible first.
Cheng Shidong believes that it is reasonable to delineate the responsible subject as the safety officer, who can still operate the vehicle in an emergency.
However, some practitioners who did not want to be named said that the determination method based on safety officers positions automatic driving as the role of auxiliary driving.
Afterwards, if other cities introduce corresponding regulations, they can further refine the identification of accident responsibility, and determine whether the human or the system is driving when the accident occurs, so as to determine whether it is the responsibility of the safety officer or the system.
"China News Weekly" Issue 33, 2022
Statement: The publication of "China News Weekly" manuscripts is authorized in writing