NIO and Mobileye to cooperate in developing L4 autonomous driving models

With its stock price rising recently, NIO has clearly crossed the financial red line. And Li Bin, who seemed to have just taken a breath, immediately started the next stage of autonomous driving research and development.

On August 28, NIO disclosed a plan to issue 75 million shares to be used for the research and development of autonomous driving technology and global market expansion.

On August 15, Ren Shaoqing, the former R&D director of Momenta, joined NIO to oversee autonomous driving R&D business. At the same time, NIO announced that the autonomous driving R&D team will be mainly based in China.

On August 11, Li Bin said that the current R&D funding is about 600 million yuan per quarter, and the investment in autonomous driving technology R&D will definitely increase in the future...

In just over ten days, with increased investment and the joining of bigwigs, NIO's self-developed autonomous driving has accelerated its pace again.

Mobileye and self-developed

"In the long run, ADAS will enhance its capabilities in four areas: chips, systems, algorithms and data. We must now establish full-stack capabilities." This was Li Bin's attitude when he talked about NIO's autonomous driving R&D strategy and direction at the Chengdu Auto Show not long ago.

It is very clear that "only children make choices". NIO, which no longer has to worry about life and death, wants both software and hardware.

A year ago, Li Bin, who was still shrouded in the shadow of "misery", did not have such a heroic spirit. Due to the cold market and lack of funds, NIO started a global layoff operation. Among them, the R&D center in San Jose, North America was not spared. As of the end of 2019, more than 100 people had been laid off.

Especially when engineers responsible for the development of autonomous driving technology were also listed on the layoff list, the road of NIO's autonomous driving research and development was put into question. Many people even asserted that NIO had given up the research and development of autonomous driving.

Such remarks were not groundless at the time. With the strategic cooperation with Mobileye to jointly develop L4 autonomous driving models, the speculation that NIO would terminate its autonomous driving research and development has become more solid.

From the official document released by Mobileye at that time, we can see that in the process of strategic cooperation between the two parties, Mobileye is responsible for providing L4-level autonomous driving car kits, including EyeQ series processors, chips, high-definition maps, driving strategies and safety-related software. And NIO only has three tasks: the development and integration of the whole vehicle autonomous driving system and the large-scale mass production of models.

Obviously, in this cooperation, it is Mobileye that determines whether the L4 autonomous driving vehicle can be successful, and NIO is just one of its many partners. In other words, Mobileye wants more operational data, and whether it is "NIO" or "the past" does not seem to be that important.

In this cooperation, apart from installing the EyeQ series of autonomous driving chips developed by Mobileye in the car as usual, NIO did not come into contact with any substantial R&D technology. It just gained an additional qualification that it can brag about in the future.

But obviously, at a time when the industrial structure is undergoing major adjustments, core technology is the key to success. Especially for NIO, which calls itself a "technology company", autonomous driving technology is the key to occupying the market. Therefore, insisting on self-research is still the main theme of NIO's development.

Therefore, NIO recruited Ren Shaoqing. Whether it is his experience at Momenta or his personal expertise, after joining NIO, Ren Shaoqing will undoubtedly take on the task of developing visual perception algorithms. And the visual perception system is exactly the positioning of Mobileye, a global unicorn company.

The service products provided by Momenta mainly include environmental perception, high-precision maps and driving decision-making algorithms. It can be guessed that Ren Shaoqing, who once served as its R&D director, will also provide a lot of help to NIO in these aspects.

In addition to introducing talents in software, there are reports that NIO may also cooperate with Black Sesame Intelligent Technology to start the process of self-developed chips. In June, Black Sesame Intelligent Technology officially released the Huashan II series chips, which is known as the strongest domestically produced intelligent driving perception chip. It is currently the only domestically produced chip that supports L3 level and above. The interconnection of dual A1000 chips can reach 140TOPS computing power, which is basically the same as the current 144TOPS computing power of Tesla FSD.

If the data of this chip is true, then NIO, which can carry out in-depth cooperation with Heizhima, will inevitably achieve leapfrog results in its own chip research and development.

NIO Smart Car R&D Jigsaw Puzzle

From a specific and realistic perspective, the functions and systems of smart cars mainly include smart driving systems, life service systems, safety protection systems, location service systems, and car service systems. Focusing on the car itself, it can be roughly divided into three aspects: smart driving, smart gateway, and smart driving cockpit.

With the announcement of a series of recent actions, NIO’s research and development puzzle around smart cars has gradually become clear.

First, in terms of intelligent driving, NIO has been equipped with EyeQ4 chips since ES8. In other words, in terms of visual perception, NIO currently chooses Mobileye. And because Mobileye provides a complete set of visual solutions, the vehicle manufacturer cannot make any modifications. Therefore, in this regard, NIO adopts a semi-self-developed "take-it-as-it-is" approach. After installing the chip in the car, it develops subsequent path planning and decision-making by itself.

It is reported that the EyeQ5 chip provided by Mobileye in the future can support blank software refresh, which means that even if NIO does not develop its own chips, it can write related programs by itself in the future, and simply "taking" will become pure self-development. It is believed that with the joining of Ren Shaoqing, NIO will soon become a company that develops its own visual solutions after Tesla and Xiaopeng.

In addition, NIO has also independently developed a three-eye camera in the perception system. Different from the monocular and binocular cameras that were more commonly used before, NIO's three-eye camera can cover the perspectives of road conditions, pedestrians, and traffic lights, which is more advanced and safer than the human eye.

In terms of smart gateway, NIO uses the self-developed LION system, which has data exchange and FOTA remote upgrade capabilities. In the Internet of Vehicles, the smart gateway is equivalent to the on-board terminal installed in the car, which can obtain vehicle information in real time and report vehicle parameters, health status, etc. to the owner. It plays a vital role in data collection, safety prevention and control in the field of intelligent driving.

Finally, in terms of the smart cockpit that represents emotions and pleasant experience, NIO believes that the most direct manifestation is its in-car voice assistant NOMI. It is reported that NOMI can learn and understand various complex commands issued by users: adjusting temperature, opening and closing windows, navigation positioning, monitoring vehicle conditions, etc. Car owners will not need to do it themselves during driving, they only need to command NOMI to help them complete the operation.

Although the in-car voice system has almost become a standard feature of electric vehicles, as information becomes increasingly abundant, NOMI now has the ability to connect context.

This shows that NIO, with money in hand, is sticking to its original promise and making every effort to move from semi-independent research and development to full independent research and development.