Artificial intelligence technology is coming and combining with all walks of life. The automobile industry has now become one of the key directions of artificial intelligence development, and major high-tech companies have entered the intelligent automobile industry. With the continuous development of artificial intelligence technology, a technological revolution has been set off in the automotive field, which also brings great challenges to automotive manufacturers. Regional control ignites new changes in the automotive industry As a new trend in the future transportation industry, automatic driving can not only liberate human labor and bring more free time to human beings, but also optimize urban traffic problems and improve driving safety through intelligent computing. It is understood that for vehicle manufacturers, the automotive electronic and electrical architecture is divided into three stages: distributed electronic and electrical architecture, (cross domain) centralized electronic and electrical architecture, and central centralized electronic and electrical architecture. With the continuous development of artificial intelligence technology, the current automotive electronic and electrical architecture is facing the transformation from the first stage to the second stage, that is, from distributed architecture to domain controller architecture. For chip manufacturers such as NXP, it is a transition from distributed to domain control and then to regional control, which also ignites new technological changes in the automotive industry. It is understood that the traditional distributed electronic and electrical architecture can not support the automatic driving ability of Ota, L3 and above of the whole vehicle, so the domain controller architecture needs to be used. With the development of automobile industry, the number of ECU (i.e. electronic control unit and controller) in automobile is rising rapidly. According to the public information, at present, the average number of ECUs used in a single model is more than 20, and there are hundreds of individual models. In the functional vehicle stage, these ECUs operate in isolation from each other and participate in the realization of each function of the vehicle. There is no need for large-scale interaction with each other. This situation also corresponds to the distributed architecture. However, in the intelligent vehicle stage, the interaction between different hardware is very frequent, and the interaction with a large number of ECUs developed by different suppliers has become more and more complex. For example, a simple algorithm command should be written into dozens of algorithms in the distributed architecture so that it can interact with ECUs from different suppliers. Zhu Xichang, director of the Institute of automotive safety technology of Tongji University, said that the original electronic and electrical architecture can achieve a certain degree of intelligence, but it is difficult to achieve L3 level automatic driving. For example, a vehicle uses Bosch's braking and ZF's steering. The ECU information of the two manufacturers often can't communicate well. In this case, the vehicle can only realize low-level intelligent driving function. The regional architecture enables efficient power and data distribution around the vehicle, while saving wiring costs and improving weight and manufacturing. A key component of this architecture is the area controller, which is responsible for connecting a large number of actuators and sensors to the central computing ECU, and can play an important role in the strategy in the area according to the application distribution. As the number of car services /ECU grows, original equipment manufacturers are looking for more scalable and cost-effective solutions to improve the E/E architecture and meet the future needs of networked, electric, and self driving vehicles. This evolution can be realized by distributing the functional logic to fewer kinds of software / hardware platforms and by making physical changes to the area based network. Liu Fang, general manager of NXP's automotive electronics business in Greater China, told China Electronics News that the regional controller architecture can enable the vehicle to have a higher speed response in different domains and directions, achieve low power consumption and reduce the cost of the whole vehicle harness. NXP started to promote the upgrading of automotive network architecture towards regional control two or three years ago, which brings more possibilities for the future intelligence and networking of the industry and adapts to the direction and trend of intelligent networking in the future. Adopting advanced manufacturing process is the inevitable trend of intelligent vehicle development In addition to architecture, with the continuous development of artificial intelligence, automotive chip has also broken people's traditional cognition and experienced the evolution from mature process to advanced process. Gartner data show that the global automotive semiconductor market is expected to reach US $65.1 billion in 2022, accounting for 12% of the global semiconductor market, and become the fastest growing part in the semiconductor segment. This also makes many automobile chip manufacturers compete for the technical priority in the field of automobile chip. Qualcomm, NXP, Geely and other auto manufacturers have recently developed 5 nm and other advanced process Auto chips, which has become one of the development paths of intelligent vehicles. In this regard, Liu Fang told China Electronics News: "Advanced process is the inevitable trend of technology development, but from the perspective of the industry, whether to adopt advanced process depends on the requirements of specific applications. Now, whether from the perspective of host plant or application scenario, we all look forward to the characteristics of high computing power and low power consumption to meet the needs of automatic driving, intelligent networking, network real-time response, more complex algorithm support, etc However, in practical applications, the best performance of performance and power consumption may be achieved through the combination of a variety of technologies, and the most appropriate path needs to be verified by mass-produced products in the future. " All technologies serve the market and demand. For automotive chips, whether to adopt advanced processes and advanced processes depends on whether the existing processes and processes have encountered bottlenecks in technology research and development, and whether they can no longer meet the needs of the market and products. A series of research and development of technologies and products must be considered based on technical feasibility. "Compared with portable devices, cars do not have more urgent requirements for chip size and power consumption, but they have higher requirements for robustness and reliability suitable for a variety of working conditions, and NXP's 5nm technology is developed based on these two requirements," Liu Fang said. Create differentiated design to achieve sustainable development Although artificial intelligence is driving various changes in the automotive industry, with the increasing technical difficulties, challenges also follow. It is understood that the gap between the market's expectations for intelligent driving and the current situation of technological innovation has become one of the biggest challenges for the development of autonomous driving enterprises. Therefore, it is necessary to have a sustainable technological development path to quickly meet the needs of market change Liu Fang believes that the key to building sustainable development lies in the differentiation of chip design and the differentiated development model for different markets. Firstly, the architecture of general computing power is used to form the differentiation of chip design, so that the chip itself forms a resource pool to jointly support software and hardware capabilities, realize multi-core linkage and simultaneous operation of multiple applications, and on these basis, it can also ensure the safety and reliability of the whole system, so as to create a sustainable development strategy. "The development trend of software defining automobile or the combination of software and hardware will become the main differentiation factor in the market, but the subsequent bottleneck is how to make these software engineers adapt to the rapid technological change and realize the high-speed reuse software capability in the upgrading of automobile electronic architecture in the future. The development platform, software support Hardware capabilities can be flexibly invoked to maximize the realization of these technologies. This sustainable ability is also an important support for more and more industries to integrate into the whole IOT service field. "Liu Fang told China Electronics News. Secondly, for different markets, we also need to create differentiated market development strategies according to local conditions. In addition to local technical support, R & D and design, production and manufacturing, we also need to cultivate and develop local partners in different links, including cloud service providers, data processing service providers, system and solution partners, AI partners, etc. Artificial intelligence has a huge driving force for the development of the automotive field. For major automobile manufacturers, if they want to effectively fill the gap between market expectations and technological innovation, they need to create sustainable innovation through differentiation, calmly deal with existing problems, and grasp development opportunities, so as to quickly occupy the market and realize rapid transformation and upgrading under the wave of artificial intelligence. (Xinhua News Agency)