Autobots' Innovative Development Strategy of Smart Cars has been published, and route competition is just around the corner.

The automobile value chain is being reconstructed, and its value core will shift from hardware to software. The level of software capability and OTA service will be easier to capture people's hearts than zero-speed acceleration. Traffic is no longer even the main function of cars, and we are in the process of profound changes in this industry.

Text/"Autobot" Huang

A few days ago, the 1 1 Ministry led by the National Development and Reform Commission and the Ministry of Industry and Information Technology issued the Notice on Printing and Distributing the Innovative Development Strategy of Smart Cars (hereinafter referred to as the Strategy). The strategy was released on February 12, and appeared in the media on February 23rd.

Among them, the "Strategy" clarifies the definition of "smart car": a new generation of cars equipped with advanced sensors and other equipment, using new technologies such as artificial intelligence, with automatic driving function, and gradually becoming an intelligent mobile space and application terminal. Obviously, smart cars should have two basic capabilities: networking and autonomous driving. AI interaction is not a necessity, but the development of AI interaction is ahead of both, reflecting the difference in technical difficulty.

Changes in short-term and medium-term goals

The draft of this strategy was published on June 2065438+. Compared with the previous version, only the "2020 goal" has been deleted.

The "2020 Goal" proposes that the proportion of new smart cars will reach more than 50%, of which the proportion of conditional automatic driving (L3) will be more than 10%; The market share of self-owned brand smart cars has reached 30%; The assembly rate of the new networked self-driving car reaches10%; The national coverage of the vehicle-mounted LTE-V wireless communication network reaches 90%, and Beidou's high-precision space-time service achieves full coverage; Build a "smart car innovation and development" platform and run in essence; Initially completed the construction of six systems (innovation, ecology, facilities, standards, supervision and safety).

Standing today, except for the "smart car ratio" link, other goals have failed. Time has proved that short-term planning overestimates the pace of industrial technology development.

In this way, the "2025 Goal" still existing in the Strategy has given up the hard indicators of the development of smart cars (the proportion of smart cars and the coverage rate of V2X) and changed to "the technological innovation, industrial ecology, infrastructure, legal standards, product supervision and network security system of China standard smart cars will basically take shape, which can realize the large-scale production of conditionally self-driving smart cars and realize the market application of highly self-driving smart cars in specific environments. By 2035, the standard smart car system in China will be fully built. From 2035 to 2050, it is also a soft goal to establish a "smart car industry power".

In short, the hardware development goal is changed to the system development goal, and the government can grasp it. The strategy seems vague, because IT has already gone beyond the scope of the traditional automobile industry and crossed too many industries (IT, communication, AI, transportation infrastructure, digital map, satellite industry, optical/microwave/millimeter wave sensor technology). At the national policy level, it is the top-level design of large-scale industrial integration that other countries have never tried. Can the combination of the two elements of "smart+car" act as a "grasping hand" and explode enough traction ability?

Who is the boss?

Only China and the United States have many industrial elements and have the ability to integrate. South Korea, where 5G deployment is leading, and Japan, where AI is leading but biased, are only part of the US-led industrial integration. Although the integration potential of the European Union is great, it is in a period of economic weakness and political disintegration, and it cannot exert its potential. Tier 1 component giant represented by Bosch pioneered the "ECU domain controller", but its stamina is weak. Now they must rely on the huge industrial clusters in China and the United States to develop. This coincides with the status of China and the United States as the two "largest single automobile markets".

Although the industrial integrity of the United States is a little poor, it has two advantages: first, the development of "computing power" with traction chips and software as elements; Second, based on geographical factors, it can integrate the production factors of all "Anglo" countries. Therefore, although its scale cost is higher than that of China, it is still in the leading position because of its excellent innovation ability.

Then the problem is very clear, and the innovation ability must be improved quickly, so that we can hope to take the lead in the biggest industrial development in the future. Another important focus of this strategy is "innovation". It is mentioned in the strategy that relying on the market, investing in software and hardware infrastructure on a large scale and providing financial support are all around "innovation", and industrial integration is only the result.

Similar tasks, different roads.

How to innovate? The strategy points out the main tasks:

First, focus on breaking through basic technologies, including new electronic and electrical architecture, multi-source sensor information fusion perception, new intelligent terminal, intelligent computing platform, vehicle-mounted wireless communication network, high-precision spatio-temporal reference service, smart car basic map, cloud control basic platform and other * * cross technologies; Second, develop testing technology; Third, define the core competitiveness, which is innovation. Include vehicle-mounted high-precision sensors, vehicle-mounted chips, intelligent operating systems, vehicle-mounted intelligent terminals, intelligent computing platforms and the like. Four, regulations and technical standards, require the establishment of China standards, seize the right to formulate rules of the game.

With the task, how to choose the implementation path? Let's go back to the two basic capabilities of "smart cars" (car networking and autonomous driving).

At present, the United States has formulated technical standards, industrial planning, and test technology research and development guidelines for car networking and autonomous driving technology, which have been in a leading position in related industries around the world, but they have not been combined, lacking a higher-level top-level design.

An interesting phenomenon is that major technology companies and car companies in China and the United States are developing autonomous driving technology and forming various technology alliances, which is no different. As soon as they got on the bus, American companies implemented "local intelligence" almost without exception; China, on the other hand, mainly adopts V2X network scheme, and "local intelligence" is struggling to compete with the former for the "mainstream position".

The US government can't promise to deploy the infrastructure related to the Internet of Vehicles, and private companies can't expect the V2X network to land before implementing the network scheme, so they can only develop the "local intelligence" scheme.

In the vehicle networking application scenario, "one cloud at both ends" is the main body, supplemented by subgrade facilities, including intelligent networked vehicles, mobile intelligent terminals, vehicle networking service platforms and other objects, involving five communication scenarios: vehicle-cloud communication, vehicle-vehicle communication, vehicle-human communication, vehicle-road communication and vehicle-vehicle communication. When the 5G-V2X network covers the road network, the computing power can be deployed in the cloud, greatly reducing the pressure of "local computing power". The main contradiction has become communication bandwidth and real-time requirements, and 5G has the potential to solve this problem. This is the advantage that enterprises represented by Huawei have 5G solutions.

Platform, software and computing power, which is the bottleneck?

The future "smart space" and "smart terminal" need the innovation of vehicle platform. Cars will change from a hardware-based industry to a software-based industry. The amount of software code required for smart cars will reach 1 100 million lines, far exceeding any product in IT history. The architecture that needs to adapt to mass computing is coming out, among which E/E architecture is being sought after by several first-line multinational car companies.

But among the challenges faced by E/E architecture, real-time and "black hole of computing power" are the biggest technical bottlenecks.

At present, almost every car company has its own electronic/electrical architecture (not hardware platform), and some even have several at the same time, but at least the software platform will be unified as never before in the future. This will be the result of market competition. For example, the mainstream platforms of mobile phones now include Android and iOS, but there is only one winner in the end of the car platform, so as to maximize the use of cloud computing capabilities and realize high-speed data transmission.

In the past few years, we have seen that every time the level of autonomous driving increases by one level, the computing power of AI will increase by one order of magnitude. If you want to realize fully automatic driving, you need the computing power of 1000TopS, which has reached the computing power of the human brain. Due to the quantum "tunneling effect", Moore's law has long failed, and it is very difficult to improve the computing power of bicycles because of energy consumption and chip size. This is the biggest bottleneck for L3 autopilot to delay commercial landing.

The premise of Strategy 2025 plan is that 5G-V2X covers 90% of the road network. In the marginal field of 5G and 4G, it relies on "edge computing" (Huawei has given a solution, but it has not been verified on a commercial scale).

As far as V2X is concerned, the United States is implementing DSRC (Special Short Range Communication Technology) improved by Wi-Fi. Because of its own technical advantages, China prefers C-V2X, which is the application and evolution of 5G-V2X.

At least for now, C-V2X has advantages in low-latency and high-density scenes, as well as coverage and distance. China 13 automobile enterprises, led by large state-owned enterprises (FAW, Dongfeng, Changan and SAIC), released the commercial signpost of C-V2X in April 20 19, and will realize the mass production of C-V2X vehicles in 2020 and 20021year.

This is regarded by observers as a sign that "collaborative intelligence" is ahead of "local intelligence", but as far as "local intelligence" itself is concerned, enterprises represented by Waymo and Cruise are in a leading position and have never been shaken. The two major technical routes are about to face the final competition.

As pointed out in the Strategy, the goal is often clear, but the path to the goal requires fierce fighting and vertical and horizontal cooperation. As always, the victory will belong to the party with more infrastructure support and the best integration capability at a reasonable cost, rather than the "most advanced" solution.

From PC to mobile phone to robot (not necessarily tangible), each generation of smart devices is ten times larger than the previous generation. As an intelligent integrated terminal, smart car is a killer application in the AI era. When it is created, it will also change the manufacturer itself (the boundary between car companies and technology companies will become blurred). Smart cars will eventually incite much greater commercial value than they are now.

Because of this, the automobile value chain is being reconstructed, and the core of its value will shift from hardware to software. The level of software capability and OTA service will be easier to capture people's hearts than zero-speed acceleration. Traffic is no longer even the main function of cars, and we are in the process of profound changes in this industry. (Text/"Autobot" Huang, part of the picture source network) Copyright statement This article is the exclusive original manuscript of Autobot, and the copyright belongs to Autobot.

This article comes from car home, the author of the car manufacturer, and does not represent car home's position.