【Gasgoo Auto】Enjoy Move Tech: A New Paradigm of Centralized Computing Software Platform for Integrated Vehicle Cloud

  • 2023-10-13 15:01:17
  • author:盖世汽车
  • click0

On September 21, 2023, at the 3rd Intelligent Vehicle Domain Controller and Central Computing Platform Innovation Summit, Zhao Jianhong, Vice President of Products at Enjoy Move Tech, proposed that cloud, edge, and device computing should be integrated to achieve true intelligent equality. 

He further elaborated that the development approach for central computing platforms is evolving, encompassing value-oriented development, multi-party participation, horizontal integration, vertical integration, and iterative development. To address the software platform requirements for next-generation central computing, Enjoy Move Tech will provide a complete set of middleware components and a cross-platform toolchain for architecture development, integration, and validation. Zhao stated that Enjoy Move Tech's development kit can help customers rapidly validate their central computing platforms, reducing development cycles by 6-9 months to meet the demand for quick new model launches in the Chinese market.


"Zhao Jianhong, Vice President of Products at Enjoy Move Tech


The following is a summary of the speech:

"Today, I'd like to introduce a new paradigm for cloud-integrated central computing software platforms, and share some of our insights from developing both software and hardware platforms for central computing.


Intelligent Vehicle Equality and Multi-Domain Fusion

The reason we're considering cloud-integrated computing is rooted in the pursuit of intelligent equality. A $2,000 smartphone offers the same level of intelligence as a $6,000 one, with identical functionalities. However, a $200,000 car and a $600,000 car exhibit vastly different levels of intelligence. This discrepancy arises from the unequal distribution of intelligence across different price points.

The current state of automotive intelligence is relatively low, while the cost of achieving high levels of intelligence is substantial. Therefore, we need a "Xiaomi" equivalent in the automotive industry to drive down the cost of intelligent automotive platforms.

From an industry perspective, integrating cloud, edge, and device computing is essential for true intelligent equality. Placing a powerful chip in a vehicle significantly increases the overall cost. However, by offloading applications to the cloud, where everyone shares the same platform, costs can be significantly reduced. Even with a $1,000 or $5,000 chip in a vehicle, coupled with cloud-based applications, the overall intelligent experience can be consistent. This means that differences in vehicle configurations are driven by hardware variations, rather than software limitations.

Our Approach to Compact Central Computing 

We're currently working on compact central computing solutions that integrate multiple domains, consolidating common functionalities for end-user vehicles. This approach allows us to leverage existing assets, such as chips and software from controllers, thereby reducing costs. For us, this means lower controller costs and development expenses. It also reduces the development workload for industry practitioners as we can largely reuse existing hardware and software platforms. This micro-innovation can substantially reduce costs and impact vehicle sales."




Below is a summary of Leapmotor's July release of "Four Leaf Clover". Based on their own understanding, they integrated controllers to create a compact central computing unit. Furthermore, building upon the central computing for intelligent cockpits, they incorporated gateway and power domain control, realizing multi-domain integration. With an 8155+S32G3 configuration and body functions, it has become a rather standard configuration for vehicles priced around 150,000 yuan, finding widespread application.



The Paradigm Shift of Centralized Computing

The transformation towards centralized computing is profound. First and foremost, a central computing unit is required. Currently, chip options are limited, primarily to those from NVIDIA and Qualcomm. Additionally, the entire vehicle's electrical and electronic architecture needs to be revised, necessitating a zonal controller. This zonal controller requires numerous innovative small controllers, a gap yet to be filled by the current industry. Moreover, traditional Tier 1 suppliers or OEMs are unable to develop central computing units independently, as they involve a vast array of knowledge domains beyond the capabilities of a single company. Truly realizing centralized computing demands the integration of numerous controller IPs and software IPs onto a single chip, signifying a comprehensive transformation from the bottom layer of controllers to software. The industry is still in the exploratory phase, and it is estimated that it will take approximately three years to fully develop this technology.

Currently, only a few industry giants, such as NVIDIA, Qualcomm, and Unisoc, have the capability to develop centralized computing. It would be more prudent to invest heavily in centralized computing after these companies have achieved significant breakthroughs.

The development approach for centralized computing needs to evolve. When our central computing is developed with a focus on user value, users themselves can become developers. Especially with the advent of large language models and AI-generated content, personalized development should be entrusted to users rather than developers.

To develop centralized computing, horizontal integration is necessary to connect OEMs, Tier 1 suppliers, and end-user developers. Additionally, vertical integration is required to create a seamless user experience. Three key technologies are crucial for centralized computing development: virtualization, cloud-native development, and containerization. Virtualization was initially used on smaller chips to address cost concerns. It has since evolved to efficiently utilize hardware and underlying system software resources. While we initially explored Type 1 and Type 2 virtualization, we found Type 1.5 to be more suitable for automotive applications.

Cloud-native development is not user-centric but developer-centric. Developers seek faster delivery and the ability to meet customer demands, making cloud-native development an ideal approach. Numerous organizations are currently building their own cloud-native ecosystems.

As experienced application developers know, creating cross-platform applications using generic middleware like AUTOSAR AP and CP is challenging. Developing a comprehensive set of components, including AUTOSAR AP and CP frameworks, that serve both cockpit and body domains and are not tailored to specific scenarios is difficult. Containerization is a more suitable approach for vertical integration and application innovation, as it allows for the integration of various components, middleware, and more.




Many automakers are developing their own operating systems during this process, such as Mercedes' OS and Toyota's OS. They are providing broader development platforms. This is essential for central computing and software-defined value, as such platforms are necessary for the migration of various applications and configurations. Opening up these platforms to Tier 1 suppliers and developers requires a significant amount of work. European automakers typically allocate around three years for this, while domestic manufacturers usually set a timeline of one to two years.


Central Computing and Software Platform Iteration and Layout

Our company is adopting an iterative approach to developing central computing and software platforms, rather than a one-time, all-in approach. We recognize that mid-to-low-end products, especially software development, offer greater profit margins. In 2022, we successfully mass-produced a platform integrating small intelligent driving functions and a cockpit-parking integrated software platform, achieving profitability and meeting market demand.

Several automakers have already started laying the groundwork for the next generation of central computing software platforms, with some planning for mass production in 2026. We will also provide a complete set of middleware components, as well as a cross-platform toolchain for architecture, development, integration, debugging, and validation. While the frameworks in the industry are quite similar, this indicates a consensus on the correct path forward. The ultimate goal is to be the first to achieve profitability and mass production.




"We have made significant strides in developing our central computing platform and have even introduced our own development kit. The image above showcases a motherboard we developed last year, which can accommodate various chip modules. This includes modules equipped with Horizon AI chips, J5 chips, Qualcomm chips, and more.

As for the underlying gateway, we are actively exploring suitable interconnect technologies for SOC-to-SOC communication. Initially, low-power chips were sufficient with a 100 Mbps Ethernet connection. However, as requirements evolved, we moved to 1G, 2.5G, and even 10G Ethernet. Subsequently, we adopted platforms like PCIe 4.0, which is supported by Qualcomm platforms, to address the increasing bandwidth demands. While platforms like J5 and NXP only support PCIe 3.0, their speeds are still insufficient to meet current needs.

With this flexible development kit, we can easily create various central computing architectures. For instance, we can fully customize the platform with domestic chips to meet the requirements of Chinese automakers. Alternatively, for vehicles exported to European or Japanese markets, we can combine chips from these regions, providing more options. Such customization is complex for Tier 1 suppliers and automakers, involving high costs for both hardware and platform software. Moreover, while many automakers focus on application innovation and cost reduction, they may not prioritize hardware innovation.

We can offer this solution to our customers to accelerate the validation of their central computing platforms. Through our close partnerships, we have established a small ecosystem that enables us to quickly verify whether a system meets customer requirements, reducing development time by 6-9 months. This is crucial for rapidly launching new vehicle models in the highly competitive Chinese market."


( The above content is from a keynote speech titled "A New Paradigm for Cloud-Integrated Central Computing Software Platforms" delivered by Zhao Jianhong, Vice President of Products at Enjoy Move Tech, at the 3rd Intelligent Vehicle Domain Controller and Central Computing Platform Innovation Summit held on September 21-22, 2023.)

Disclaimer

Part of the content of this website comes from content authors or automatic crawling from the Internet. The content of the relevant text only represents the views of the author or publisher of this article and does not represent the views or positions of Enjoy Move Tech. Enjoy Move Tech strives for the objectivity and impartiality of the content and opinions described in this information but does not guarantee the accuracy and completeness of the content, nor does it guarantee that the content will not be changed in the future. If the authors and editors of the content displayed on this website think that their works are not suitable for browsing on the Internet or should not be used free of charge, please inform us by email or telephone in time. Enjoy Move Tech will take reasonable measures in timely to avoid unnecessary economic losses to both parties.

Email us: contact@enjoymove.com

Communication & Cooperation

Learn more about our products and find new opportunities!

Message

Please enter the information correctly. We will contact you in time!