Driving through the “Three Levels”

  Demonstration operation is in the ascendant. When will the large-scale commercial application of intelligent networked vehicles be realized?
From L3 to L4: the hardest jump

  In Wuhan Economic and Technological Development Zone (hereinafter referred to as “Wuhan Economic and Technological Development Zone”), many self-driving taxis in trial operation are wearing a huge “mushroom head”, which integrates many sensors. They drive along fixed lines, making it obvious at a glance that this is a test car.
  Contrary to people’s imagination, driverless cars are not “no one”. A reporter from Fudian Finance and Economics saw that these vehicles are generally equipped with a safety officer in the driving position. Safety officers do not actively intervene while the vehicle is in motion, and in case of emergency, they will take over the vehicle.
  Li Kai, CEO of Dongfeng Yuexiang Technology Co., Ltd., told reporters that at present, whether it is self-driving buses, taxis, or unmanned retail vehicles and logistics vehicles in office areas and scenic spots, the speed is generally about 30 kilometers per hour. The reasons for maintaining slow operation are mainly due to immature technical conditions, followed by more secure and reliable management. In case of emergencies, relevant units can also provide solutions in time. Li Kai said that both speed and safety are important indicators of test operation, and they are also important factors for whether it can be put into use on a large scale in the future. “From the current situation, there are still many improvements and improvements in technology.”
  According to the “Data Tracking Report on China’s Autonomous Vehicle Market” released by IDC, in the first quarter of 2022, the penetration rate of L2-level autonomous driving in the passenger car market reached 23.2%, and the entire market was from L2 (partially automated driving) to L3 ( Conditional autonomous driving) + level development stage.
  According to industry insiders, moving from the L3 level to the L4 level (highly autonomous driving) is the most important step in realizing unmanned driving. This is the most critical leap and the most difficult leap, just like crossing a big hurdle.

On August 8, in Wuhan, Baidu Apollo’s autonomous driving travel service platform “Tuipao” started the commercial operation of autonomous driving without a safety officer in the car.

  ”From the L3 level to the L4 level, there is at least a 50-fold increase in the amount of computing and data, which requires us to innovate in the fields of large computing power chips and artificial intelligence core algorithms.” According to Lu Chihua, a professor at the School of Automotive Engineering, Wuhan University of Technology, my country It is an advantage to have a large amount of data and a large number of application scenarios for ICVs. However, at present, the top algorithm technology and large computing power chips are controlled by foreign companies. We mainly rely on the introduction. In addition to being expensive, we often face “stuck necks”.
Large-scale commercialization calls for cost reduction

  Wuhan Guangyu Mingsheng Intelligent Technology Co., Ltd. (formerly Wuhan Guangting Technology Co., Ltd., hereinafter referred to as “Guangyu Intelligent”) is one of the earliest companies in China to carry out research and development in the field of intelligent connected vehicles.
  The cost burden of driverless vehicles comes from the hardware first. “Smarter cars require more sensitive systems, which have higher requirements for chip computing power and sensor performance.” Su Xiaocong, general manager of Guangyu Intelligence, introduced that the research and development and investment of these security “weights” require a lot of money. At present, the cost of the self-driving taxis (Robotaxi) in demonstration operations in many places at home and abroad generally exceeds 1 million yuan, and some even reach 2 million yuan.
  Taking Baidu Apollo’s fifth-generation unmanned vehicle as an example, the total cost of the model is 480,000 yuan, and the monthly cost is about 8,000 yuan based on five years of stable operation.
  However, in recent years, the cost of driverless cars has also gradually decreased. “The complete vehicle cost of some models has gradually dropped from 2 million yuan to 400,000 yuan.” Chen Li, general manager of Wuhan Smart Ecological Technology Investment Co., Ltd., introduced that taking lidar as an example, the previous 64 lines (meaning in the vertical direction) Lidars with multiple transmitters and receivers, the more lines, the larger the amount of data processed and the higher the efficiency, all need more than 2,000 US dollars (equivalent to about 14,000 yuan), and now there are 64 lines, 96 lines, 128 lines. The line of laser radar can be produced independently in my country, and the average price of a laser radar is only 6,000 yuan.
  In July this year, Baidu launched the Apollo RT6, an unmanned vehicle that can remove the steering wheel. The cost of a single vehicle is only 250,000 yuan. According to the 5-year operation, the monthly cost is only 4,167 yuan, which is nearly 50% lower than the cost of the previous generation of unmanned vehicles. .
  In addition to hardware, with the improvement of the related supply chain and manufacturing process, the material cost and production cost of the vehicle will continue to decrease.
How to manage driverless vehicle accidents

  As autonomous driving moves toward commercial use, management regulations must keep up. If the driverless system fails in operation and causes a traffic accident, the principle of disposal is very important, but because the handling and driving rules are very different from those of traditional cars, the existing laws and regulations cannot be applied.
  For example, Su Xiaocong said, “For example, what standards and conditions should an intelligent networked car meet before it goes on the road? What licenses need to be obtained? If a traffic accident occurs due to unmanned driving, for vehicle manufacturers, software or chip suppliers, etc., the How to determine and divide responsibilities?”
  It is reported that Shenzhen has been the first to test the waters in terms of laws and regulations. On August 1 this year, the “Shenzhen Special Economic Zone Management Regulations on Intelligent Networked Vehicles” came into effect. This is the first domestic regulation on the management of intelligent networked vehicles. It clearly requires that intelligent networked vehicles must be registered by the traffic management department of the public security organ before they can go on the road. drive. The division of responsibility for traffic accidents has also made clearer regulations.
  In June this year, the Wuhan Municipal Bureau of Economy and Information Technology, the Wuhan Municipal Public Security Bureau, and the Wuhan Municipal Transportation Bureau issued the “Implementation Rules for the Management of Road Testing and Demonstration Application of Intelligent Connected Vehicles in Wuhan (Trial)”, allowing the main body of demonstration applications to carry out in-vehicle Remote testing, demonstration, and commercial operations without safety officers provide detailed guidance and support for companies developing autonomous driving businesses. However, this detailed rule is only aimed at the field of testing and demonstration applications, and does not provide for large-scale commercial matters such as the registration of ICVs and their official road operation.
  The “14th Five-Year” Development Plan for the Automobile Industry of Hubei Province proposes to accelerate the application of 5G, Internet of Things, and Beidou systems in intelligent networked vehicles, and to take new energy vehicles as the first carrier of intelligent networked technology to promote cross-border enterprises. Collaborate to improve technical standards and legal norms. By 2025, L2 (partially autonomous driving) and L3 (conditional autonomous driving) ICVs will account for more than 50% of annual car sales, and L4 (highly autonomous) ICVs will begin to enter the market.
  Obviously, the industrial planning objectives and steps are clear, and the supporting and perfecting of relevant management regulations should be followed in due course.