Or there is only one solution. Let's talk about the future direction of electronic and electrical architecture from a user's perspective.

Publisher:梦回归处Latest update time:2022-02-16 Source: BAO爱车工作室 Reading articles on mobile phones Scan QR code
Read articles on your mobile phone anytime, anywhere

At the beginning of the Year of the Tiger, I want to sit down and write a serious article about my thoughts on the future evolution of cars. Since I am not an engineer, I can only talk about some thoughts triggered by my observations in recent years from the perspective of the media, or from the perspective of a specific user. One of the biggest reflections is the soul-searching question raised by an Internet executive many years ago: Why can't we see popular applications like WeChat in cars?



At the time, I seemed to be able to think of the answer to this question without hesitation: because the usage scenarios are different. But as I get older and know more about the direction of cars transforming into mobile smart terminals, I am less confident in the answer I gave at the time, and even start to question myself: Why can't I see popular applications like WeChat in cars? In the end... after years of thinking... I finally found that I still can't answer this question in one sentence. Rome was not built in a day, and the answers to many questions are not black and white, so in this article, I want to choose an angle from many reasons to interpret it. Maybe it is not the whole answer to solve the problem, but at least it provides a perspective. This angle is the electronic and electrical architecture (hereinafter, the electronic and electrical architecture will be abbreviated as the English letters EEA).



From the concept of EEA proposed by Delphi, to the six-stage EEA architecture evolution roadmap developed by Bosch, and  the fact that Tesla MODEL 3  (  parameters  |  inquiry  ) took the lead in mass production of Zonal architecture, it is enough to see that both the traditional car industry giants and the new forces in car manufacturing have launched their own actions after "suffering from traditional architecture for a long time". Cars are becoming more and more intelligent. If we continue to build smart cars under the traditional architecture, the number of ECUs will increase, which will bring a lot of "unbearable weight" to the industry. The first is the increase in the length and weight of the wiring harness. To give two examples, the electronic system of the Mercedes-Benz S-Class sedan in 2000 already had 80 ECUs and 1,900 communication buses with a total length of 4km. The bus length of the Audi Q7 and Porsche Cayenne launched in 2007 has exceeded 6km, and the total weight exceeds 70kg, which is basically the second heaviest component of the whole car after the engine. The weight that the company has worked hard to cut off on the body materials has been added back on the wiring harness. Of course, it is not only the weight that has increased, but also the cost. The two examples above were from more than a decade ago. Today, many vehicles have more than 200 ECUs. With the addition of features such as autonomous driving and smart cockpits, the number will grow exponentially if traditional distributed EEA is used.



The increase in functions not only brings about a rapid increase in the length of the wiring harness and the cost of materials, but also brings about coordination and testing costs that are difficult for companies to bear. It is difficult to make different systems compatible with each other and maintain a certain common rhythm. The market for the development of a single vehicle used to be as long as 5-7 years, and most of the time was spent on coordinating product updates and testing from different suppliers. Now, it also needs to have more computing power and even software functions that need to be upgraded from time to time. Using traditional EEA will undoubtedly drive OEM integration engineers crazy.



In addition, communication bandwidth is also a bottleneck in traditional EEA. With the continuous evolution of ADAS, autonomous driving and other functions, the amount of data in in-vehicle communication has increased exponentially. Calculated by the data transmission volume of a single sensor, the amount of data generated by the radar and camera of the ADAS system each exceeds 100Mbyte/s. Taking a car equipped with five radar sensors and two camera sensors as an example, during the acquisition and storage period, it is necessary to manage a massive amount of data of about 1GByte/s. After the ET7 announced by NIO is equipped with a laser radar, the data throughput reaches 18G per second, and its own bandwidth still achieves 32G/second! However, the bandwidth of traditional FlexRay, LIN and CAN low-speed buses is only a few megabytes at most. This gap is no different from the generation gap between a five-inch disk (storage medium in the 1980s and 1990s) and a solid-state drive. Distributed EEA can no longer provide the required high-bandwidth communication capabilities.



Of course, the author believes that the most important point is that the traditional distributed architecture makes the software and hardware too tightly bound. Each ECU used in each part has its own exclusive system. The functions in a car require dozens of different systems to match each other before they can be used. In addition to the above-mentioned upgrades, which require a lot of effort to coordinate, the greater harm is that the barriers between various hardware cannot be broken through to achieve functional innovation. For example, the success of the iPhone is not due to the use of a touch screen or the installation of a camera, but through an excellent system, the hardware barriers between various component suppliers have been broken through, giving developers a very relaxed and efficient development environment. With the wisdom of the people, the computing power of the computer + the convenience of the previous handheld computer, as well as hardware such as cameras and lidar, have achieved a super user experience of 1+1 greater than 2. This experience has led to a positive cycle of more developers participating, and finally a healthy ecosystem was born. Under such a healthy ecosystem, the birth of a hit software has become inevitable. However, the author believes that the current EEA architecture does not yet have the conditions to breed this key force, so it is not surprising that a hit such as WeChat on the mobile phone cannot be born.


First-tier OEM companies naturally discovered this problem! In recent years, more and more companies have begun to improve EEA. For example, domestic first-tier brands such as Great Wall, Changan, and Geely have begun mass production of "Domain Centralized EEA" in their current products. Compared with the previous distributed architecture, this type of EEA further improves the degree of modularity and ECU functional integration, and the concept of "Function Domain" has emerged. From the perspective of software and hardware architecture, the most intuitive manifestation of domain-centralized EEA is the "Domain Control Unit (DCU)" as the core of the entire functional domain. The so-called domain controller refers to the general term for the entire system composed of domain master hardware, operating system, algorithm, and application software. The domain host processor (Domain Host Processor) is the brain core of the domain controller, usually served by a processor with higher integration and stronger performance.



In short, the management system in the "domain-centralized EEA" car is no longer a one-to-many leader commanding all employees, but instead has department managers who will help the top executives manage the grassroots. These department managers usually determine their titles according to the functions of the vehicle. For example, Bosch once divided the body control domain into the power domain (Power Train), chassis domain (Chassis), body domain (Body/Comfort), cockpit domain (Cockpit/Infotainment), and autonomous driving domain (ADAS). The Volkswagen ID. series based on the MEB pure electric platform adopts a new generation of vehicle electronic and electrical architecture, in which there are three main computing units, namely the ICAS1 body control domain, the ICAS2 advanced autonomous driving domain, and the ICAS3 intelligent cockpit domain (currently only ICAS1 and ICAS3 are equipped in mass-produced products), of which ICAS1 is produced and provided by Continental AG of Germany. Huawei's MDC solution is similar to Volkswagen's idea.


There are three advantages to this approach: first, centralized management reduces management costs; second, high integration reduces the number of ECUs, and the platform is more scalable; third, more complex sensor fusion algorithms can be run on domain master processors with more powerful computing power, making it possible to implement complex functions across sensors. In this context, many practical functions have been born, such as supporting the download and installation of vehicle software, and online updates and upgrades of safety functions, giving vehicles the ability to "learn throughout their lives."



However, such a solution is obviously excessive. Tesla, a "catfish" with no historical burden, uses a more radical EEA form on its mass-produced product MODEL 3. This system does not divide the domain according to function, but directly divides the main management work of the vehicle into three ways, which become the left body control module, the central computing unit, and the right body control module. The industry calls this form a cross-domain/centralized EEA. Centralized EEA can not only concentrate computing resources and facilitate the separation of software and hardware; it also brings a lot of imagination to the power management of each controller of the whole vehicle. Taking the cross-domain/centralized EEA adopted by Tesla MODEL 3 as an example, it greatly reduces the use of wiring harnesses and greatly reduces the cost of vehicles. At the same time, it provides a foundation for Tesla to throw away the constraints of traditional parts suppliers and vigorously develop its own full-stack software and hardware capabilities. Of course, there are also problems with Tesla's approach, that is, it has not been able to physically separate the IVI system from the ADAS control system, resulting in the possibility of being hacked from the outside to some extent. Therefore, when developing cross-domain/centralized EEA, companies including BMW and Toyota have considered the idea of ​​physically isolating core safety content such as autonomous driving and vehicle driving from the IVI system. However, despite the follow-up of the above-mentioned major OEMs and first-tier suppliers including Continental and Aptiv, in the face of such a huge change, the follow-up speed of traditional companies in the automotive industry is still relatively slow. According to the predictions of some corporate research teams, the real popularization of cross-domain/centralized EEA may not be until 2025, and 2022 is still the year when "domain-centralized EEA" begins to become popular. However, even so, it can be seen in the market that vehicle brands have enjoyed the "bonus" of electronic and electrical architecture. For example, the "full-stack R&D capabilities" emphasized by some brands are mostly based on breaking the software and hardware barriers in traditional electronic and electrical architecture, which means that it is possible to develop functions that require the coordination of multiple sensors of different brands.

[1] [2]
Reference address:Or there is only one solution. Let's talk about the future direction of electronic and electrical architecture from a user's perspective.

Previous article:New observations on automotive electronics: Testing is becoming increasingly important
Next article:Global electric vehicle sales in 2021 will reach 6.5 million, with China accounting for half

Latest Automotive Electronics Articles
Change More Related Popular Components

EEWorld
subscription
account

EEWorld
service
account

Automotive
development
circle

About Us Customer Service Contact Information Datasheet Sitemap LatestNews


Room 1530, 15th Floor, Building B, No.18 Zhongguancun Street, Haidian District, Beijing, Postal Code: 100190 China Telephone: 008610 8235 0740

Copyright © 2005-2024 EEWORLD.com.cn, Inc. All rights reserved 京ICP证060456号 京ICP备10001474号-1 电信业务审批[2006]字第258号函 京公网安备 11010802033920号