Author: Mr.Yu

“Lenovo does not have a car-making plan. What we want to do is car computing,” said Dr. Ru Yong, Senior Vice President and CTO of Lenovo Group, during his speech on November 9 at the second day of the 2022 Lenovo Innovation and Technology Conference.

This is Lenovo, which does not make cars, but has won several champion projects in two of the world’s three major computer vision academic conferences, CVPR and ECCV, this year.

Among them, in the Multi-Object Tracking and Segmentation Challenge (MOTChallenge) of the CVPR, Lenovo Research Institute won the championship of the average multi-object tracking accuracy (mMOTA), competing against more than 20 top teams from all over the world, including the Swiss Federal Institute of Technology Zurich, the University of California, Berkeley, Carnegie Mellon University, and Udacity.

Compared with Huawei, which competes directly with Lenovo in many fields such as consumer electronics and high-performance computing, Lenovo, also emphasizing “not making cars,” has a different game plan.

What Is Car Computing?

“Car computing” can be seen as a general concept.

In simple terms, Lenovo’s “car computing” can be seen as the integration of main domain controllers for autonomous driving, intelligent cabins, body, power, and other functions, and finally aggregated to the central computing platform.

However, this is not a simple concept. As is well known, the supply chain of automobiles is too long. Just in the car display screen we previously discussed in the article, there may be dozens or even hundreds of suppliers behind one or two screens in front of users.

Likewise, car computing needs to integrate many control units on the car, so that they can be centralized from a relatively scattered working state and achieve better iteration through OTA. Therefore, this is not a simple software or hardware concept, but a large-scale and standardized computing system behind it.This type of aggregation is not easy to do, as it involves a lot of core technologies, including heterogeneous computing, multi-system parallelism, cross-domain communication, optimization, acceleration, and more.

And these are all areas in which Lenovo excels, as the world’s leading supplier of high-performance computing among the TOP500 and one of the most important computing power companies.

Looking at a single set of data, as of 2022, Lenovo has deployed 161 supercomputers in China, accounting for 32.2% of the global market share.

Enterprise computing is always online. According to Lenovo, automotive computing is a natural extension of its computing business, and it makes sense to invest in this area as a new direction for computing in specific scenarios.

As one direction of its computing power, Lenovo has presented a new architecture of Edge-Cloud-Net-Intelligence for automotive computing, which also includes clear scenes and corresponding products.

Lenovo, in the game for a long time

As early as 2014, we reported on Lenovo’s entry into the car networking industry with the title “Lenovo + Car Networking: Another big news, and this time the entry point is still OBD”. It has been seven years since then, and the concept of “Car Networking” has been subdivided over time with industry development.

Let’s jump to the present. A few months ago, Lenovo Research Institute announced a generous offer to recruit talent in the automotive industry.

From the public job postings, the recruitment includes talent for positions such as head of in-car information and entertainment systems, director of autonomous driving engineering, and senior software/hardware managers. Clearly, key areas in intelligent vehicles, such as autonomous driving, intelligent cockpit, and central computing platform, have become key areas of Lenovo’s collaboration with the automobile industry.

At the Automotive Computing Forum on November 9th, Lenovo Vice President and Head of Automotive Computing, Tang Xinyue, announced Lenovo’s car computing product layout and roadmap for the first time. Specifically tailored to different application scenarios, Lenovo provides computing platforms with varying levels of partnering hardware and software. On this basis, Lenovo provides operating system software, middleware, and application development environments.

Among them are three major product lines: smart cockpit, autonomous driving and intelligent display screen.

Take the display screen as an example. Based on the domestically produced chips of Chixi Technology, Lenovo has launched an intelligent cockpit domain controller with one machine and multiple screens, supporting instrument panel, central control screen, co-pilot screen, entertainment screen, and realizing functions such as voice control, 3D navigation, and interconnection with mobile phones.

Statistics show that 30% of global car accidents are caused by A-pillar blind spots. Especially in the process of turning, there is a high probability that personnel and obstacles in the A-pillar blind spot will collide.

Based on insight into this situation, Lenovo’s technology accumulation in the field of flexible screens will also be applied to the automotive field, such as the newly announced curved transparent A-pillar system. The system includes a flexible curved screen and corresponding scene fusion algorithm.

Through the combination of software and hardware, the A-pillar with a flexible curved screen will achieve digital perspective. Users can see the scenery in the A-pillar blind spot through the screen, and the curve correction algorithm can also solve the problem of the actual position of objects in the field of view to a certain extent.

It is understood that the product is still in the early research and development stage, but Lenovo has started to contact domestic and foreign manufacturers for cooperation.

In fact, whether it is a one-machine-multi-screen or a novel curved transparent A-pillar system, they are both parts of the concept smart cockpit system jointly created by Lenovo and Chery.

In addition to these, this system also includes related intelligent cockpit domain controllers, driver health monitoring systems, multi-device wireless charging products, as well as flexible curved central control screens, dual-screen car screens, and corresponding user interfaces.

Apart from intelligent cockpits and screens, Lenovo has also announced the roadmap for its domain controller product line by 2025, covering different scenes from intelligent cockpits to cockpit and cabin fusion with various computing powers.

In terms of autonomous driving domain controllers, Lenovo has launched a 550TOPS computing power autonomous driving domain controller product which can be used for high-performance autonomous driving on non-regulated roads;

In 2023, Lenovo will release a high-performance and cost-effective domain controller solution for integrated driving and parking with 32TOPS computing power.

According to the plan, Lenovo will also release a cockpit and cabin fusion domain controller product around the year 2024-2025, with the highest computing power of up to 2000TOPS. In terms of computing power alone, it will be comparable to NVIDIA’s Orin successor and Thunderbolt (Thor) chip level announced at GTC in the autumn of 2022.

With the surge of in-vehicle computing power, Lenovo plans to provide more abundant vehicle computing solutions such as cloud computing products, cloud storage products, roadside MEC products, V2X products, and more, beyond the single-class in-vehicle computing products.

Among the known cases, Lenovo has successfully implemented a smart traffic solution in the challenging “mountain city” of Chongqing, China. With the 5G+MEC+C-V2X+central cloud system, Lenovo has designed a vehicle-road coordinated solution that integrates people, cars, roads, networks, and clouds with the aim of reducing the cost of autonomous driving and improving the efficiency of traffic management.“`markdown

In many occasions, Lenovo’s leader Yang Yuanqing has emphasized the importance of “developing Lenovo’s business from ‘computers’ to ‘computing power'” and constructing computing power around different scenarios.

Despite the series of difficulties brought about by the COVID-19 pandemic and the far-reaching uncertainty caused by the upgrading of international geopolitics, the automotive industry, which demonstrates strong vitality, is regarded as a promising investment in the eyes of many industry giants, including Lenovo.

As part of Lenovo’s plan, the total R&D investment in the next five years will exceed CNY 100 billion, and a large number of R&D talents will continue to be recruited.

If everything goes well, as the next focus of development, car computing will receive a lot of resources and talent support from Lenovo Group to ensure successful implementation of the plan.

Conclusion

Let’s go back to the beginning. As long as we are slightly familiar with Lenovo, we will know that “selling computers” is not the only business card of Lenovo.

In addition to decades of accumulation in personal computers, supercomputers, and network security, Lenovo has also expressed its own admission criteria in a slightly Versailles manner to achieve ideal car computing, which are:

Supply chain and manufacturing capabilities – maintaining good long-term cooperation with global supply chain leaders such as Intel, NVIDIA, and Samsung;

Qualification, accumulation, and cross-border integration – After years of deep cultivation, Lenovo has trained a large number of talents in server, computer, edge computing, display, and other fields;

Friends in the automotive industry – Investment partners include not only car ecosystem partners such as NIO and CATL, but also chip and algorithm enterprises such as Cambricon, Horizon Robotics, BYD Semiconductor, and Qianzhou Intelligent Navigation.

“`Many times, the cross-industry entry of top companies in the industry outside the circle may cause a certain probability of industry backlash, or even arouse some degree of hostility.

From this perspective, Lenovo’s statement of “not building cars” has maintained a relatively low-key attitude.

However, low-key does not mean humble. Lenovo is also expressing to the new and old players in the automobile circle in its own way its cross-industry background and deep qualifications, which have weight and value for the increasingly anxious competitors.

Compared with this, whether Lenovo itself produces cars doesn’t seem that important.

“Lenovo has no plans to build cars. Car computing is a new direction for Lenovo’s scenario-based computing. Lenovo will use its computing advantage to empower the new four modernizations of automobiles.”

Throughout the entire process of the online sub-forum, this sentence was repeatedly mentioned by Lenovo’s senior management.

This article is a translation by ChatGPT of a Chinese report from 42HOW. If you have any questions about it, please email bd@42how.com.