During the 2022 Beijing Winter Olympic Games, the Beijing self driving office organized a multi vehicle driverless demonstration operation in Shougang Park of the Winter Olympic Games. Professor Yang diange of the school of vehicles of Tsinghua University, as the expert group leader, was responsible for the organization and technical preparation of the automatic driving demonstration of intelligent networked vehicles in Beijing during the Winter Olympic Games. As the person in charge, he took the lead in the special project of Beijing Science and technology Winter Olympics automatic driving, led Baidu, Futian, Shougang, Zhixing, Traffic Management Bureau and other units to complete the research and development of all-weather unmanned driving technology for the Winter Olympics, and demonstrated China's 5g, Beidou and AI technologies to the world.
With the help of the stage of the Beijing Winter Olympic Games, China showed the world the intelligent connected vehicle technology of China's scheme. However, there are still many "uncertainties" in this developing industry. What are the development trends of smart cars in China, how should driverless cars be industrialized, and what is the relationship between smart cars and the construction of "dual smart" cities?
With the above questions, on April 11, the 21st Century Business Herald reporter interviewed Professor Yang diange exclusively.
Professor Yang diange believes that the path for the industrialization of high-level automatic driving in the future may be the strategy of "squeezing two ends", rather than a single path. On the one hand, it is to realize the industrialization from intelligent assisted driving and gradually develop to high-level automatic driving; On the other hand, L4 high-level automatic driving is based on scene driven landing, constantly expanding the application scene and scope, so that unmanned driving can be applied in a wider rangep align="center">
21st Century Business Herald (hereinafter referred to as "21st century"): what new developments and changes have taken place in the intelligent new energy vehicle industry in the past year? What are the key points in the competition of intelligent new energy vehicle industry
Yang diange: over the past two years, the development and competition of intelligent new energy vehicle industry have become more and more complex. However, if we peel off the surface phenomenon and look at the essence, we will find that the competition of intelligent new energy vehicle industry has changed from vehicle design and manufacturing technology itself to ecological competition and big data competition behind ecology. This is not only an important change in the current industrial development, but also the key point of the current automobile industry competition.
In the past year, the market value of Tesla, an American electric vehicle enterprise, has exceeded US $trillion, more than the sum of Toyota, Volkswagen, Mercedes Benz and BMW, which ranked second to tenth. Is Tesla's market value so high because of Tesla's battery, motor and electronic control technology? In fact, it is not. The value lies in the millions of cars Tesla drives on the roads all over the world. Every day, it continues to collect and send back the data with the characteristics of long tail scenes encountered by some vehicles to its data center, so that Tesla can continuously iterate and improve its automatic driving algorithm, and this iterative ability is constantly opening the leading gap between Tesla and the major car companies behind it.
The function and performance of traditional vehicles depend on the movement brought by the combustion of fossil fuels in the engine and mechanical transmission. The function and performance of vehicles completely depend on the mechanical and physical structure and its limitations. The first is the change of vehicle energy, and the second is the electrification of driving, that is, electric driving. The energy is no longer transmitted through machinery, but through wire harness, and the movement is also realized through the drive of motor. In this case, the function of the car is actually separated from the physical structure. The function of the car can be defined by software. For example, whether the car has automatic driving function may seem to have the same physical structure, but after loading the software, there will be intelligent function. Without loading the software, there will be no intelligent function. After the car has the characteristics of software definition, its function definition, function iteration and performance improvement rely on the big data iteration behind it. In the development of smart cars, whoever has mastered the complete ecology of smart cars and who can really grasp the big data generated behind smart cars can continuously iterate the functions and improve the performance of cars through data iteration. Tesla has such characteristics, so it has such value recognition.
"21st century": in the development of intelligent new energy vehicles, ICT enterprises such as Huawei, Xiaomi and Baidu have cut into the automotive industry, but they have taken different development paths. Geely, Weilai and other automobile enterprises have entered the mobile phone industry one after another. Li Bin said that "not making a mobile phone is equivalent to that the key of the home is not in his own hand, which will be very passive". What do you think of this? How to evaluate the different strategies of these enterprises
Yang diange: whether ICT enterprises such as Huawei, Xiaomi and Baidu enter the automobile industry, or Geely and Weilai make mobile phones, it seems that everyone does things in completely different ways and contents, but they come to the same goal by different paths. In essence, they all hope to explore and establish their own complete ecosystem of smart cars, and then fully grasp the ecology and the big data behind them, Use big data to better iterate their products, make their products more competitive and take the initiative in the development of the industry in the future.
We have adopted different development routes to build our own ecology. Now it is difficult to say who is right and who is wrong. There are many uncertainties in the technical development trend and route of intelligent vehicles. Different enterprises have different resources and characteristics. As long as we can establish an independent and controllable ecology, we do not necessarily have to unify the technical route.
"21st century": Huawei's expanding circle of friends has triggered the discussion of "soul and body" in the industry. How to treat the competition and cooperation between technology companies and automobile enterprises? How should car companies defend their "soul"
Yang diange: if the words are slightly replaced, such as replacing "soul" with core technology and turning "body" into OEM, you may not be able to question this matter. Any car enterprise or company hopes to master the core technology and the development lifeline of the enterprise, so as to make the enterprise most competitive and obtain the greatest benefits in the whole industrial chain. Therefore, enterprises want to master their soul, in fact, they want to master the core technology. In my opinion, we should correctly view the pursuit of "soul" by car enterprises. In fact, what they really pursue and defend is core technology and core competitiveness.
In the traditional automobile industry chain, automobile enterprises play a leading role, but in the process of establishing the ecology of new intelligent vehicles, it is not easy to see how to keep up with the downstream partners in order to maintain the leading position, competitive advantage and voice in the industry chain. For this reason, Huawei, Xiaomi, Baidu, Geely, Weilai and other enterprises adopt different modes and different technical routes to establish ecology. The development of intelligent new energy vehicle technology and the establishment of ecology are still under exploration. In the process of exploration, what enterprises should and should not do should be tolerated.
With regard to the relationship between Huawei and its "circle of friends", we can see that Huawei is establishing cooperation with more and more Chinese and foreign automobile enterprises, with greater and greater influence, but it is also accompanied by some doubts and concerns. These concerns stem from the unclear strategic intention of Huawei's own external transmission. At different times and stages, Huawei's strategic intentions for external transmission are different. Although Huawei makes it clear that it does not build cars, is Huawei positioning Bosch, Delphi or Google of the Internet? Should we be a part of the automotive enterprise ecology, or should we master the ecology and master the big data behind the ecology? As mentioned earlier, making a car without making a car is not the most critical competitive point in the automotive industry today. Ecology and big data are the key to competition. Huawei's control of the ecology and big data of smart cars and its positioning of itself and automotive enterprises are the most critical. If in the division of labor and cooperation, Huawei has done all the core technologies, mastered the ecology, mastered big data and mastered the iterative ability of products, what will car companies do? This is where car companies are concerned and worried.
Therefore, I think Huawei needs to have a correct positioning, and the transmission of external strategic intention should be unified and clear. Enterprises in the automotive industry should be assured of Huawei, be willing to cooperate with Huawei, and know that cooperation with Huawei will not lose their dominant position in the ecology. Car enterprises can master the core technology, use Huawei's products and technologies to establish their own ecology, and use the big data behind the ecology to iterate product functions and improve product performance. Only in this way can car enterprises safely cooperate with Huawei. I really hope Huawei can establish a win-win circle of friends with all auto enterprises in China.
21st century: foreign car companies such as Volkswagen are cooperating more and more closely with Chinese technology companies such as Huawei. How do you evaluate the cooperation between foreign companies and technology companies in China
Yang diange: the cooperation between Volkswagen and Huawei has not been confirmed by authority. However, such news is not a bad thing, which shows that China's automatic driving technology is relatively advanced, competitive and valuable. At least for now, China's automatic driving technology is basically in line with that of enterprises in the United States, Japan and Germany. China has more complex traffic scenes with mixed people and vehicles, and the development of automatic driving technology is facing greater challenges. Of course, this also means greater opportunities for innovation. In December 2021, the authority's statistics on the automatic driving vehicles from all over the world tested in California, the United States ranked five in the top 10 companies, and five in China. The fully automated technology of China's autonomous vehicle is not backward.
Secondly, in recent years, there have been news about the cooperation needs between traditional automobile enterprises and ICT enterprises and the cooperation and investment with AI automatic driving companies, which also shows that in the past quite a long time, traditional automobile enterprises have been conservative in the development of intelligent automobile automatic driving technology, insufficient investment and lagged behind in the development of corresponding technologies, Now, more and more traditional large car companies have gradually begun to realize the huge revolutionary changes brought by smart cars and feel the survival crisis brought by the technological change of smart cars, just like when smart phones eliminated traditional phones and left Nokia behind. The continuous spread of new news also shows that these traditional car enterprises need transformation. There are many ways of transformation, which is a good development opportunity for everyone.
In addition, I think it is very wise for large foreign automobile enterprises to choose to cooperate with Chinese ICT enterprises, automatic driving companies or Chinese universities and scientific research institutes. Unlike the engine, manufacturing technology and other technologies of traditional automobiles, intelligent automobile technology involves data security and geographic information security. Localized development and R & D is very necessary and important. Therefore, it is very important for international auto enterprises to research and develop the localization of automatic driving and cooperate with the R & D force of localization technology.
"21st century": driverless driving is still far away, but there have been some pilot applications based on mining areas, parks and other scenes. How do you view the industrialization path of automatic driving
Yang diange: l5 level unmanned driving is really far from us, but I think in the past two years, including some high-level automatic driving demonstrations during the Winter Olympics, we can see that high-level automatic driving based on scene driving is completely possible. In addition, in the past two years, some L2 level and above assisted driving technologies have also been industrialized on a large scale.
Therefore, the development of automatic driving in the future may be the strategy of "squeezing at both ends". On the one hand, assisted driving is developing upward on the basis of mass production, and more and more intelligent assisted driving technologies are applied, such as automatic driving with certain functions and a certain range in scenes such as highways; On the other hand, L4 high-level automatic driving continues to expand the application scenarios and scope, so that unmanned driving can be applied in a wider range. These unmanned areas will expand more and more in the future. When they expand to a certain extent, the connected area will also become larger and larger.
Therefore, I think that the path of the industrialization of automatic driving in the future may not be a single path, and there may be a variety of development paths. One is to gradually develop from assisted driving to high-level automatic driving, which may be related to the mass production and manufacturing of cars and the iterative development of automobile enterprise technology. In addition, some autopilot companies and demonstration applications in some special scenarios. In the field of passenger cars, I am personally optimistic about Didi's exploration mode of future travel services with the mixed mode of manned special vehicle and unmanned automatic driving. It is still possible for Didi's robotaxi to achieve large-scale commercialization in the next 5-10 years.
Intelligent connected vehicles are the core of "dual smart" cities
Professor Yang diange is also a member of the expert group of China's "double smart" city pilot plan. The reporter also interviewed Professor Yang diange on the "double smart" pilot work that we have paid more attention to recently.
"21st century": Why did China launch the "dual smart" city pilot program? What is the relationship between it and the technical route of China's intelligent Internet connected vehicles
Yang diange: China's pilot of "dual smart" cities is related to China's technical route for the development of intelligent networked vehicles. The technical development route of intelligent vehicles in China is essentially different from that in foreign countries. We also attach importance to Internet connected intelligence on the basis of paying attention to single vehicle intelligence. We hope to improve the reliability of automatic driving, reduce the cost of single vehicle intelligence, and better promote the landing application of high-level automatic driving of intelligent vehicles through vehicle road coordination and vehicle vehicle coordination in the Internet connected environment.
The technical route of intelligent network connection involves the construction of infrastructure. In order to truly achieve intelligent network connection and vehicle road coordination, the road must be intelligent, traffic management and traffic signals need to be intelligent, and the comprehensive transportation system of the whole city also needs to make adaptive changes for intelligent vehicles. Specific to each city, what should be built first and then? What should the road be built, what should each intersection be built, what sensors should be installed, what data should be obtained, and what data should be exchanged between the road and the car? These need to be explored.
Beijing, Shanghai, Guangzhou, Wuhan, Changsha, Wuxi and other cities are selected for the pilot. On the one hand, these cities have a good foundation in the fields of intelligent vehicle related technology research and development, intelligent vehicle related testing and demonstration area construction. At the same time, these cities have made relatively large investment in the construction of intelligent transportation system in recent years.
Judging from the current demonstration scheme, the way, mode and region of intelligent connected vehicle demonstration in these cities, the construction of smart roads and smart cities are very different and have their own characteristics, which is the significance of the pilot. Through the demonstration of different technical routes, we find common laws and gradually form standards and norms. Through demonstration and full operation, accumulate enough experience, establish standards and norms, and then promote them to the whole country, which is a more reasonable development model.
"21st century": behind the construction of "double smart" city is a huge chain, which spans multiple industries. What role does intelligent Internet connected vehicle play in it? The impact of smart car network construction on smart city
Yang diange: first of all, smart car is just a point in smart transportation and smart city. Of course, it is a very key point. If there is no smart car, there will be no "dual smart" city.
Intelligent transportation and Smart City hope to turn the vehicles in the whole transportation system into a controllable intelligent node, adjust dynamically in real time, distribute the traffic load evenly and reasonably on the road, and maximize the traffic efficiency. In the traditional transportation system, the flow of vehicles can only be controlled through traffic signals, so it is difficult for the system to achieve accurate control of each vehicle. Now every intelligent vehicle is not only intelligent, but also has networking ability, and can interact with the system in real time, which means that every vehicle has the ability to become a controllable intelligent node.
For the construction of smart city and intelligent transportation system, vehicle is not only the object of traffic control, but also the goal of system service. What kind of functions and performance intelligent networked vehicles have also determines how to build intelligent transportation systems and smart cities in order to better serve vehicles.
Smart transportation and smart city are very big concepts, which need a lot of investment, but these need to be designed around the intelligent connected vehicle. The intelligent connected vehicle is not only the core of the "dual smart" city, but also its regulation object and regulation goal.
"21st century": what are the challenges facing the development of "double smart" cities? How to deal with it
Yang diange: for the development of "dual smart" cities, in addition to smart cars themselves, it is very important to intellectualize and informationize the infrastructure serving smart cars and its management policies. In addition, a major challenge facing the development of "Shuangzhi" city is that there is great uncertainty in the technical route and time roadmap of intelligent vehicles. Therefore, what plan should be adopted and how to develop the construction of Shuangzhi city and infrastructure? What technical route should be adopted, what rhythm should be adopted, and what should be built first and then? There is great uncertainty.
For a city, the investment in infrastructure is huge. Intelligent transformation of roads and infrastructure to serve high-level automatic driving requires a lot of capital investment, but the profitable business model is not particularly clear at present. In other words, the development of "double smart" cities, in addition to exploring the technical route of development, also need to explore a reasonable business model to achieve sustainable and sound development. If we can't find a reasonable market driven business model, it must be unsustainable to rely only on the investment of the government and enterprises. In addition, the development of "double smart" cities requires close cooperation between government, industry, University and Research Institute, better system design, focus on key technology research and development, and better coordinate resources in all aspects.
