laitimes

L3 level automatic driving has reached the domestic boarding node? A number of industry experts threw forward-looking views

Car stuff (public number: chedongxi)

The author | Wooden rice

Edit | Juice

On April 8, the 2022 Automotive Evaluation International Summit and C-NCAP Annual Conference was held in Tianjin, focusing on the development trend of intelligent and connected vehicles and how to improve the domestic autonomous driving industry.

At this meeting, THEC Released the annual Blue Book of China Automobile Evaluation, summarizing the evaluation results of C-NCAP (China New Car Evaluation Regulations) and CCRT (China Automobile Consumer Research and Evaluation) from 2018 to 2021, and interpreting and analyzing them.

In the next stage, in addition to C-NCAP and CCRT, the China Automobile Center is also promoting the verification and implementation of a special regulation system for autonomous driving evaluation for the actual needs of Chinese users.

In addition, a number of academicians, represented by Li Keqiang, academician of the Chinese Academy of Engineering and professor of the School of Vehicles and Transportation of Tsinghua University, also gave special speeches on the topic of intelligent networked vehicles and automatic driving, and jointly discussed the current development trend and challenges of intelligent and connected vehicles in China.

First, safety is the key Automatic driving urgently needs a supporting evaluation system

At the meeting, Academician Li Keqiang delivered a speech entitled "Key Technologies of Intelligent and Connected Vehicles for The Next Generation of AI-Driven China".

In recent years, Academician Li Keqiang has led a scientific research team to continuously explore how to comprehensively improve the driving safety, energy consumption economy, riding comfort, and traffic patency in the application process of vehicle autonomous driving by improving driving intelligence.

He believes that the "cloud-enabled automatic driving system" that integrates "real-time dynamic data, basic service content, and collaborative computing environment", that is, "the intelligent networked car of the Chinese solution", has become the consensus of the industry, which has unique characteristics in system definition, key technologies and industrial ecology. At the same time, the next generation of AI-driven Chinese solution intelligent networked vehicles have achieved phased results in three typical key technologies, such as complex environment perception, brain-like intelligent decision-making, and intelligent networked car-cloud brain integration system.

Academician Li Keqiang stressed that the intelligent networked vehicles of the Chinese solution need to have a system architecture of "vehicle-road-cloud integration" and the technical characteristics of "hierarchical decoupling and cross-domain sharing", and meet the three conditions of "meeting China's infrastructure standards, meeting China's networking operation standards, and meeting China's new system architecture automotive product standards".

L3 level automatic driving has reached the domestic boarding node? A number of industry experts threw forward-looking views

▲Professor and academician Li Keqiang of Tsinghua University

In addition, Academician Li Keqiang also pointed out that "the new generation of artificial intelligence technology should be truly applied in intelligent networked vehicles and form a large intelligent travel system, and there are a series of technical problems that need to be broken." "Autonomous driving is an advanced stage in the development of intelligent and connected vehicles, and safety is a major core issue in the process of realizing autonomous driving in the future.

On the one hand, in the current technical background, the intelligence of autonomous driving is still very insufficient, and it is impossible to make accurate judgments on complex situations. Academician Li Keqiang said that the current mainland intelligent networked vehicles on the traffic lights of the environmental perception accuracy rate has reached 95%, pedestrian perception accuracy of 90% (about 80% in the dynamic environment), vehicle perception accuracy rate of 98%, although compared with the past has made great progress, but still not reached absolute accuracy, its decision-making intelligence is far lower than the average human average, especially in harsh environments, the perception performance of vehicles by rain, snow, fog, haze, glare, sand and dust and other weather impact significantly, there is a situation of manual takeover, It also needs to be continuously improved through technical means.

On the other hand, intelligent networked cars are a new generation of electronic information terminals, which will also face attacks from sensors, operating systems, internet of vehicles and other ways, and its information security challenges are becoming increasingly severe.

"Any human takeover means a potential autonomous driving accident." Academician Li Keqiang stressed many times in his speech that safety is the first demand of autonomous vehicles. To break through the shortcomings of self-driving cars in terms of system safety, a matching evaluation system should be established to identify hidden dangers as soon as possible.

At present, through years of exploration and research, based on the experience accumulated by C-NCAP and CCRT professional evaluation of China Automobile Evaluation, and with the support of various enterprises and experts in the industry, China Automobile Center has made a stage breakthrough in automatic driving evaluation technology, and is promoting the verification and implementation of automatic driving evaluation procedures for the actual needs of Chinese users.

Second, last year, smart cars accounted for more than 45% of China has reached the L3 autonomous driving node

Che Dongxi learned at the meeting that according to relevant professional statistics, 711 new cars will be listed in the Chinese market in 2021, of which 328 have intelligent driving functions, accounting for more than 45%. With the increasing variety of intelligent products, China's autonomous driving evaluation solutions that meet the needs of users at higher levels have also attracted much attention.

At present, the automatic driving function level of the vast majority of mass-produced models at home and abroad is focused on the L2 level, with assisted driving and pilot assisted driving functions as the mainstay, but the driver is still the driving body and bears the main responsibility. Recently, a small number of models equipped with L3 autonomous driving functions have been certified in Europe and Japan.

Under the active guidance of favorable policies, China's autonomous driving industry has also shown a vigorous upward trend. At the meeting, a number of guests analyzed the development of autonomous vehicles from different perspectives and put forward forward-looking views.

Yuan Jianhua, chief researcher of the Institute of Traffic Management Science of the Ministry of Public Security, introduced the existing policies and regulations on autonomous driving and intelligent automotive in mainland China. He stressed that the mainland attaches great importance to the development of autonomous driving technology, especially the safety of technical applications. Among them, the state adopts technical regulations such as standards and norms to create the most basic guarantee for consumer use and safety management. Industry evaluation standards verify product performance in more dimensions, promote enterprises to continuously improve technology and quality, and provide an important basis for the formulation and optimization of relevant national standards, systems and regulations in the future.

Zhu Xichan, a professor at Tongji University's School of Automotive, said that the grading standard for automatic driving should be "moderate." In the standard evaluation process, on the basis of the pursuit of "zero deaths", vehicles that can be safer than a skilled and cautious driver, but have a small probability of residual accident risk, should be recognized and certified. If the relevant access standards are too much pursuit of 100% "0 accidents", it may affect the riding experience and traffic conditions of smart cars, and limit the development and application of autonomous driving technology.

L3 level automatic driving has reached the domestic boarding node? A number of industry experts threw forward-looking views

▲Conference theme

Wang Qingfeng, test director of WEILAI Auto's autonomous driving system, proposed, "High-level automatic driving evaluation should be considered from the three dimensions of safety, comfort and ease of use, focusing on the statistical test of the combined scenario, focusing more on the performance data performance within the equivalent mileage, attaching importance to the 'end-to-end' full-scenario test that is more oriented to the user's usage habits, and focusing on the user's driving experience on the basis of focusing on safety." ”

"At the critical point when L3-level autonomous driving products are about to enter the market, in order to facilitate consumers to clearly understand the difference between autonomous driving and assisted driving in the scope of capabilities and the boundaries of use, guide consumers to correctly, rationally and safely select cars for use, and introduce and implement autonomous vehicle evaluation procedures," qin Kongjian, chief expert of intelligent network evaluation of China Automobile Center, said when introducing the user-oriented automatic driving evaluation system.

Conclusion: Intelligent networking technology continues to develop

Intelligent networked vehicles have now become a big hit in the development of the entire automotive industry, and the recent proposal of the double carbon target has also profoundly affected the industrial ecology of the entire industry.

At the International Summit on Automobile Evaluation and the Annual Meeting of C-NCAP, enterprises and experts and scholars in various industries have shown their attention to the ecology of intelligent and connected vehicles, and with the introduction of domestic supporting policies, the mainland's intelligent networking and automatic driving industry will also accelerate the pace of development.

Read on