Words? |? Liumao
On September 14, this afternoon, another grand event in the field of intelligent networked automobiles this year-2020 World Intelligent Networked Automobile Congress officially opened in Shanghai Automobile Convention and Exhibition Center. In the same period, the site also held a "Zhilian Tianxia"? Leading the future "theme forum activities.
Self-driving cars exhibited at the event site
At the beginning of the conference, including Sino-German Intelligent Networked Automobile Promotion and Application Center and Shanghai Intelligent Networked Automobile Data Center, * * * six intelligent networked automobile-related projects were unveiled, started or signed at the event site.
Li Jun, Chairman of China Automotive Engineering Society, Professor Tsinghua University, Academician of China Academy of Engineering, Zu Sijie, Vice President and Chief Engineer of Shanghai Automotive Group Co., Ltd., Meng Xing, Chief Operating Officer of Didi Autopilot Company, Yu Kai, founder and CEO of Horizon, and other industry experts and practitioners came to the stage to share their thoughts on the development path of intelligent networked vehicles and their concrete achievements in intelligent networked vehicles and self-driving vehicles.
Li Jun believes that the development of self-driving cars needs to break through the technical route of intelligent bicycles. There are five "soul problems" in bicycle intelligence. In order to solve the limitation of bicycle intelligence, Li Jun put forward eight suggestions. Zu Sijie said that SAIC will officially put into production L3-class 5G smart cars by the end of the year, and plans to complete the transportation task of 20,000 standard containers this year.
The following is the dry goods content of this conference.
First, the construction of intelligent networked vehicles in Shanghai has accelerated.
With the deep integration of informationization, big data, artificial intelligence and automobiles, automobiles are evolving from traditional means of transportation to smarter new travel carriers. From a strategic perspective, intelligent networked vehicles have been regarded by more and more countries as an important part of seizing the opportunities for future development.
China has unique market and policy advantages in building intelligent networked automobiles and developing intelligent networked automobile industry.
At the opening ceremony of the conference this afternoon, six other projects related to intelligent networked vehicles completed the unveiling, launching and signing ceremony.
2065438+In July 2008, China and Germany signed the Statement of Joint Intention on Cooperation in the Field of Automatic Networked Driving. Subsequently, the Sino-German Association of Automobile Manufacturers signed a memorandum of cooperation with China and initiated the establishment of the Sino-German Intelligent Networked Automobile Promotion and Application Center.
In August, 2020, the Jiading District Government of Shanghai reached a cooperation intention with China Automobile Industry Association, confirming that the Sino-German center project will officially land in Jiading, Shanghai.
This afternoon, the unveiling ceremony of the Sino-German Intelligent Networked Automobile Promotion and Application Center was officially held at the 2020 World Intelligent Networked Automobile Conference, which finally brought this cooperation to the ground.
In addition to the Sino-German Intelligent Networked Automobile Promotion and Application Center, the opening ceremony of the conference also held the signing ceremony of the strategic cooperation of the National Intelligent Networked Automobile Innovation Center, the launching ceremony of the cooperation of the national-level Yangtze River Delta regional automobile networking experimental zone, the release of the results of the 5G intelligent transportation demonstration zone and the official launching ceremony of the second phase of the project.
National Yangtze River Delta Regional Vehicle Networking Experimental Zone Cooperation * * * Launched
The main application scenarios of Shanghai Intelligent Networked Automobile Data Center and Shanghai Intelligent Networked Automobile were unveiled and released at the opening ceremony.
It is understood that next, Shanghai will promote Jiading's global opening (***463 square kilometers), and build Jiading's self-driving manned demonstration, Lingang intelligent bus and intelligent logistics demonstration, Fengxian automatic parking demonstration, Pudong Jinqiao Central City automatic driving and bus sanitation demonstration.
Second, Li Jun, academician of China Academy of Engineering: There are five major problems in bicycle intelligence.
Li Jun, Chairman of China Automotive Engineering Society, Professor Tsinghua University and Academician of China Academy of Engineering, said after taking office that the topic of today's speech is "China Intelligent Networked Automobile Development Needs Innovation Breakthrough", based on the current global and domestic situation.
Li Jun believes that the development of self-driving cars needs new technical routes and new technological breakthroughs.
In his view, the J30 16 standard formulated by SAE is essentially aimed at the car itself, that is, bicycle intelligence.
This standard defines four vehicle characteristics (DDT, ODD, OEDR, DDT? But it also means that it has been predicted that bicycle intelligence will "fail" at some point. If there are many faults, it cannot be said that automatic driving has been realized.
Self-driving cars need new technical routes.
At present, relevant institutions in the European Union and the United States are already formulating some new standards for the development of self-driving cars.
As far as intelligent networked cars are concerned, most enterprises still take the technical route of intelligent bicycles.
Li Jun divided these enterprises into two categories. The first category is mainly Waymo, multi-line lidar and multi-sensor sensing. He called it "saturation perception". The second category, represented by Tesla, is based on camera and visual perception.
However, the smart bicycle routes represented by these two schemes have five weaknesses.
1, unmanned driving must rely on AI, AI has a black box effect, how to overcome it?
2. Rand think tank thinks that autonomous driving needs 1 1 billion miles (about 177 billion kilometers) of road test. How to achieve it?
3. Fully automatic driving has at least millions of extreme working conditions. How to ensure and verify software design?
4, L3, L4, L5 automatic driving cost is very high, how to enter the private car?
5. How to ensure the actual driving safety of fully automatic driving cars?
In view of the above five problems, Li Jun proposed to make technological breakthroughs in eight aspects.
The first is to establish a new concept of intelligence. Secondly, we should combine the empowerment of the outside world with bicycle intelligence. The third is to establish a learning and growth system for autonomous driving software. Li Jun believes that "software-defined cars, data-developed cars and chip-made cars."
New technical concept of "new generation intelligent networked car"
Fourth, it needs infrastructure support such as cloud support and roadside support; Fifth, in terms of security, we should ensure security from three angles: offline security, growth and learning, and online security. The sixth one is 5G. In Li Jun's view, the 5G car is a complex system, which will turn the car into a digital car. The 5G car itself has three major commodity attributes.
In addition, chips and testing also need to break through. Li Jun emphasized that people now pay more attention to in-vehicle chips, but the chips outside the car are actually equally important.
In terms of safety test verification, Li Jun said that at present, China lags behind foreign countries in highly autonomous driving laboratories and test technology.
At the end of the speech, Li Jun said that the development of intelligent networked vehicles needs national system engineering, and suggested that China should formulate its own relevant standards, such as scene standards, map standards and communication standards.
Third, SAIC Zu Sijie: Intelligent Networked Cars are the "winners and losers" of independent car companies.
Zu Sijie shared SAIC's practice and thinking on intelligent networked vehicles.
In recent years, the domestic automobile market suffered a "cold winter", and at the beginning of this year, it suffered an epidemic in COVID-19. The development of automobile industry is facing many difficulties.
In Zu Sijie's view, under this situation, the road of independent brands will also face more difficulties. On the other hand, he believes that the next few years will still be a period of strategic opportunities for domestic car companies to make a difference, and intelligent networked cars are the most important "winners and losers".
Independent brands need to develop intelligent networked cars. Zu Sijie said that from the market point of view, the first-time purchase demand is the most important source of sales for independent brands. If products want to cover major consumer groups, they need to show their individuality and technology labels through intelligent networking.
From the perspective of industrial ecology, intelligent networked vehicles are becoming the intersection of many innovative fields. The landing of artificial intelligence technology, the development of industrial Internet, the iteration of 5G communication, the exploration of big data application, the application of car gauge chips, Beidou networking and smart cities all hope that intelligent networked cars can become a breakthrough.
In addition, in terms of software-defined cars, China's automobile industry is basically at the same starting line as its global counterparts, which also provides valuable opportunities for independent brands.
Zu Sijie also shared some achievements and future plans of SAIC in intelligent networked vehicles.
In addition to cooperating with Ali Group to build a zebra system, SAIC also actively carried out organizational structure and business landing.
In terms of organizational structure, SAIC has built three platforms, namely, Fan Yishang, the cloud center of SAIC, the artificial intelligence laboratory of SAIC and the software center of SAIC.
According to Zu Sijie, Fan Yi is still building a hybrid cloud architecture to build a big data storage and application platform for intelligent networked vehicles and provide cloud computing capabilities for intelligent networked vehicles.
At present, SAIC Artificial Intelligence Lab gathers more than 120 artificial intelligence talents, focusing on AI technologies such as machine vision and multimodal interaction, and supporting the iterative growth of products such as intelligent driving and intelligent cockpit.
Last month, SAIC and Horizon jointly established the artificial intelligence joint laboratory.
SAIC software center zero beam technology undertakes the research and development tasks of new generation electronic architecture, SOA software platform, OS and AI computing chips. At the same time, it is also an important institution to attract software talents and establish a circle of friends in software and chips.
In addition, SAIC also focuses on the industrialization of the core 4i technology of intelligent networked vehicles through Lianchuang Electronics Company.
Development of SAIC Intelligent Networked Vehicles
In terms of business landing, Zu Sijie said that the 5G smart car demonstration experience zone project was delivered this morning to realize the marketization of automatic driving scenes such as high speed, urban congestion and parking.
According to the plan, SAIC will officially put into production the L3-class 5G smart car MARVEL at the end of the year. r .
At the same time, SAIC also plans to complete the transportation task of 20 thousand TEUs this year. "We have started small-scale commercial operations, and strive to complete 200,000 standard containers next year, and operate according to actual business scenarios." Zu Sijie said at the scene.
In addition, SAIC is still laying out the electronic and electrical architecture, software platform and data platform to build the basic capabilities of intelligent networked vehicles.
Fourth, Didi Mengxing: The number of people signing up to experience autonomous driving is close to 40,000.
Meng Xing, chief operating officer of Didi Autopilot Company, shared some work and thoughts of Didi Autopilot in the past four years.
According to Meng Xing, Didi Autopilot Company was established in Beijing four years ago and now has offices in California, Beijing, Shanghai and Suzhou. Shanghai is the most important trial operation scene at present.
At the end of June this year, Didi launched the Didi autopilot experience in Jiading Automobile City, Shanghai. As of today, the number of applicants is close to 40,000.
So, how does Didi do autonomous driving?
Meng Xing said that Didi, as a travel company, considered how to integrate traditional travel with go on road trip when doing go on road trip. This is also the reason why Didi does "mixed scheduling".
In addition, Didi believes that doing autonomous driving requires three aspects. One is AI, the intelligent ability in the car. One is to need enough data as the fuel of AI. Secondly, there must be an ecological network for travel.
In terms of technical ability, we must first have the ability of full stack. This is because "the upstream and downstream relationship between each module is very close. Only when each module has its own research and development ability can you iterate fast enough." What we are concerned about today is not how high the level of each module or the overall system is, but more importantly, how fast the iteration speed is. "Meng Xing said in his speech.
To sum up, the business of Didi autopilot involves three aspects, namely AI, data and network. In Didi's view, only these three aspects are done, and the self-driving travel service is considered to be done.
Verb (abbreviation of verb) Horizon Yu Kai: In 2030, smart car chips will become the largest vertical market for semiconductors.
Yu Kai, founder and CEO of Horizon, shared Horizon's thoughts on automotive artificial intelligence chips.
In Yu Kai's view, if in the era of traditional fuel vehicles, the engine is the core competitiveness and the core component of the car. Then in the era of software-defined cars, in the era of big data and artificial intelligence, data is oil and energy; The chip is the engine.
At present, the value chain of the whole automobile industry is undergoing profound changes and changes. In the future, cars will become supercomputers on four wheels. In this process, the chip and operating system are very important. "It may be subversive innovation, not incremental innovation". Yu Kai said at the scene.
Yu Kai further emphasized the importance of data and data processing ability.
In his view, compared with the amount of data collected by Waymo and Tesla, Waymo can be regarded as an innovation of research mode. In contrast, Tesla has created a connected AI body through shadow mode, which has complete computing power, algorithms and data and can evolve rapidly. Although Tesla's autonomous driving ability is not perfect, what is really terrible is its evolutionary speed.
Yu Kai said in his speech, "If it is difficult for other car companies to keep up with its evolution speed, it will become the definer of the smart car era and the definer of the platform, and its control over this industry will surpass that of Microsoft, Intel and Apple."
Regarding the future, Yu Kai thinks that there is still a big gap between today's most automatic driving and future demand. He predicted that by 2030, smart car chips will become the largest market in the entire semiconductor field. At the same time, in this field, a giant computing platform enterprise is likely to be born, and such a giant enterprise is likely to be born in China.
This article comes from car home, the author of the car manufacturer, and does not represent car home's position.