Tesla was a big advocate of this idea and was the first to put it into practice. In 2012, when Tesla launched the Model S, it released the Software Over-The-Air (SOTA) function for the first time, which enables the vehicle to remotely upgrade the vehicle entertainment system and application software, such as navigation system and human-computer interaction interface. When Model 3 was released in 2017, the Firmware Over-The-Air (FOTA) feature was further introduced to enable remote online upgrades of system firmware, once again setting a new standard for the industry.

05701-A-0351 In this trend, the first is the change of business model and profit model of enterprises. Tesla began experimenting with OTA paid upgrades in 2019, and in 2021 launched two autonomous driving options at different levels: EAP(Enhanced AutoPilot) and FSD(Full Self-Driving). According to Goldman Sachs forecasts, Tesla’s software such as FSD has reached $1 billion to $3 billion in annual revenue, with long-term gross margins as high as 90%, while its highest gross margin on vehicle sales is only 32.9%. By 2030, this figure will reach $10 billion to $75 billion. Considering that Tesla’s total revenue for 2022 is $81.5 billion, this figure is remarkable. It can be seen that in the era of “software definition”, software service revenue has begun to play an increasingly important role in the proportion of revenue and profit of enterprises, and enterprises need to re-examine their business models and product strategies.

What is clear is that the revolutionary changes in product Electronic & Electrical Architecture (E /E) triggered by “software-defined products” will also bring about changes in product development models.

In simple terms, the traditional automotive E/E architecture is often a distributed architecture with the ECU as the core, that is, the sensor, the ECU and the actuator correspond one by one to ensure the independence and anti-interference of each system. The big test brought by intelligent networking, autonomous driving and other technologies to vehicle development is the explosive data processing needs and higher computing speed. The traditional distributed E/E architecture, whether it is computing power and data processing capacity, or vehicle network transmission efficiency, can not meet the requirements of autonomous driving and intelligent network connection technology. Therefore, the automotive E/E architecture began to gradually shift from the traditional distributed architecture to the central centralized architecture based on cloud computing. Below is the roadmap for the evolution of the automotive E/E architecture given by Bosch.

05701-A-0351

According to Deloitte, the automotive E/E architecture will go through three evolution cycles: integration, domain centralization, and central centralization. Among them, as the ultimate goal of evolution, the central centralized architecture greatly realizes the decoupling of software and hardware, and the software is directly deployed on the central computing platform, which can support the OTA upgrade of the vehicle software. However, in the process of mass production development, the scale and difficulty of software development of central centralized architecture surge, facing multidimensional challenges such as technical complexity, organizational system, development mode, and development cycle. It is not only necessary to solve the software deployment of multi-core heterogeneous chips, but also involves the software development of multiple domain controllers with cross-domain integration. 05701-A-0351 Therefore, in order to reduce the difficulty of development and greatly improve the efficiency of development, it is necessary to transfer the development view from the domain controller level to the vehicle level, and provide an integrated development platform, development view and development method. At the same time, in order to improve the reusability and scalability of software functions, the vehicle software technology architecture will also change from the traditional signal-oriented distributed architecture to the layered and service-oriented architecture based on a unified platform.

Tesla took the lead in adopting a central centralized architecture in its Model3 models in 2019, and domestic new energy vehicle companies are also undergoing a transformation process from a domain centralized architecture to a central centralized architecture. Similarly, other industries have also put forward similar concepts, such as “software defined fan”, “software defined construction machinery” and so on.