Autonomous driving is gradually approaching: Beijing opens commercialization pilot Huawei DJI has entered the game

● Reporter Zhang Xingwang, Jin Yidan

In early December, Shenzhen still had a strong sense of warmth, and the streets below the DJI headquarters of Weixin Software Technology Park were bustling with traffic. "In the future, autonomous driving will become the standard equipment of cars like headlights and rearview mirrors." Xie Khotan, head of DJI's vehicle market, told the China Securities News reporter that the autonomous driving industry is still a blue ocean, and more scenes are needed for commercialization.

A few days ago, the country's first pilot pilot for the commercialization of autonomous driving travel services was implemented in Beijing, and autonomous driving will usher in a trillion-yuan scale track, and the wide track will attract Huawei, Baidu, DJI, SenseTime and other technology companies to pour in.

Autonomous driving is evolving rapidly

In Beijing's Yizhuang Economic Development Zone, a reporter from China Securities News recently experienced autonomous driving travel services. The reporter clicked on the driverless service software on the mobile phone, entered the destination and placed an order, and selected the nearest station to wait. "I'm not a driver, I'm a safety officer." Master Wang, who sits in the driver's seat, said he would only intervene to take over the vehicle in an emergency.

Master Wang has been a taxi driver for 11 years. "Self-driving taxis can save the cost of drivers, and after the technology matures, even safety officers are not needed." Master Wang said, "The work of the safety officer is much easier, not as highly nervous as the driver, and the vehicle route is automatically planned." The reporter found that Robotaxi (self-driving taxi) has a stable speed during the urban driving process, and rarely actively changes lanes to overtake. When encountering the complex road conditions of pedestrians, Robotaxi's sensing system is sensitive and slows down the brakes in time.

On November 25, the "Implementation Rules for the Pilot Administration of Commercialization of Autonomous Driving Travel Services in the Beijing Intelligent Connected Vehicle Policy Pilot Zone (Trial)" was released, and Baidu and Xiaoma Zhixing became the first batch of enterprises allowed to carry out commercial pilot services. The country's first pilot for the commercialization of autonomous driving travel services was implemented in Beijing, marking the transition of domestic autonomous driving from test demonstration to commercialization pilot, and autonomous driving officially entered the "second half".

Autonomous driving is usually divided into five levels, from assisted driving, conditional automation and finally full automation.

Technological advances have driven the rapid development of the autonomous driving industry. Guosheng Securities believes that many automotive OEMs and Tier1 manufacturers (tier 1 suppliers) have accelerated their approach to the front of autonomous driving, but L4 levels are difficult to commercialize on a large scale in the short term. Startups strengthen cooperation with automotive companies, cover extreme scenarios through continuous data accumulation, train autonomous driving systems and algorithms, and promote product iteration. CICC pointed out that the new forces actively promote full-stack self-research, and traditional car companies mostly complement each other's advantages through cooperation. For the new forces that do not have the advantage of "car building", R&D capabilities are an important support for them to highlight the siege, and can independently complete the development of the whole process of automatic driving perception and decision-making algorithms, vehicle control strategies, underlying systems, and tests.

The autonomous driving track is wide and the market is large, attracting many companies to join. The China Securities News reporter learned that SAIC-GM-Wuling is equipped with intelligent driving solution products provided by DJI vehicles and will be mass-produced and listed. Volkswagen China is also working with DJI to develop advanced driver assistance features.

AutoX is an autonomous driving R&D service provider, and recently signed a contract with Pingshan District, Shenzhen to promote the large-scale landing of Robotaxi unmanned driving. Xiao Jianxiong, founder and CEO of AutoX, told China Securities News: "The company cut into this track, mainly because of the huge user group and strong demand in this market. Changes in the way you get around will have a huge impact. ”

According to the data of Tianyancha, there are currently more than 5800 enterprises in China whose business scope includes "automatic driving, intelligent driving, and unmanned driving". Since the beginning of this year, more than 1,000 new related enterprises have been registered.

According to the HIS report, the global autonomous driving market is expected to reach $162.9 billion by 2022, an increase of about 14% year-on-year. China's autonomous driving market is growing faster, growing at about 24% in 2022.

Yuanrong Qixing, a Shenzhen self-driving startup established in 2019, completed a $300 million Series B financing led by Alibaba in September this year. Zhou Guang, CEO of Yuanrong Qixing, told China Securities News: "The threshold of autonomous driving technology is high, and top talents are needed. At present, the industry has limited talents and it is difficult to support the development of a large number of companies. ”

Tech giants have entered the market

Since 2021, mainstream car companies have accelerated the research and development and landing progress of high-level intelligent driving, and technology giants have entered the market. WEILAI, Xiaopeng, Ideal, GAC and Great Wall have launched intelligent models with pilot assistance systems (with L4 functions), equipped with various sensors, high-computing chips and upgradable computing platforms, which can realize automatic driving in urban and high-speed sections. There are signs that Apple is working on the Apple Car, a self-driving car. Deloitte pointed out that the degree of intelligence has become the core indicator for consumers to judge the attractiveness of new energy vehicles.

Huawei accelerates the development of its smart automotive components business. Xu Zhijun, Huawei's rotating chairman, said in May this year: "For intelligent driving, Huawei and its partners have two cooperation models, one is Huawei Inside mode, which provides intelligent driving full-stack solutions including intelligent driving application software, computing platforms, and sensors. The other is the platform model, Huawei provides MDC intelligent driving computing platform, which mainly includes hardware based on Ascend SoC, autopilot operating system AOS and vehicle control operating system VOS, as well as AutoSAR middleware. The MDC platform connects the sensors and actuators of the partner and enables the partner to develop intelligent driving software. ”

At present, Huawei's intelligent automotive components business has more than 5,000 R&D teams, including more than 2,000 autonomous driving R&D personnel. It is understood that Huawei, BAIC, Changan, and GAC chose huawei inside cooperation model. In addition, Huawei is also helping Xiaokang Xilis build and sell cars.

In the field of autonomous driving, many technology companies are positioned differently. Xie Khotan told a reporter from China Securities News: "DJI's positioning is tier1, which is a supplier of software and hardware. To achieve 'good experience' of autonomous driving, it is necessary to overcome the 'four mountains', which are sensors, algorithms, computing power, and data. DJI will consider these factors to provide car companies with practical and affordable autonomous driving solutions. ”

Shi Jianping, vice president of intelligent driving of SenseTime Mobile Intelligent Business Group, told China Securities News: "There are many automobile manufacturers and more brands, and some car companies may be reluctant to choose full-stack solutions. SenseTime's strength is INI algorithms, which include perception algorithms for autonomous driving systems. We connect directly with the car manufacturer and integrate with the hardware tier1 as a software tier1. ”

"Promoting the development of autonomous driving is the protagonist of car companies, but the idea of how to land is different. If technology companies empower car companies, I think there may be more cases of 'holding thighs'. DJI wants to work deeply with like-minded customers and grow together. Khotan said bluntly.

How far we still have to go for commercialization

Intelligent passenger cars are the main way to commercialize autonomous driving. The pace of commercialization of Xiaopeng Automobile's driver assistance system continues to accelerate. In the prospectus, Xiaopeng Motors mentioned that it provides customers with self-developed advanced automatic driving system XPILOT 3.0 software in the form of paid services, and as of March 31, 2021, more than 20% of P7 car owners have purchased and activated this function.

Autonomous mobility services are currently only entering the commercial pilot phase. The relevant person in charge of Xiaoma Zhixing said that the issuance of the pilot qualification for the commercialization of autonomous driving travel services in Beijing is of landmark significance, which means that the autonomous driving service has moved from test demonstration to commercial demonstration.

Zhou Guang told a reporter from China Securities News: "Autonomous driving has not been commercialized on a large scale, in addition to technical factors, another reason is that the cost is too high." ”

According to IHS Markit data, the market size of Robotaxi is expected to exceed 1.3 trillion yuan in 2030. According to a report released by The International Data (Asia) Group, China has the potential to become the world's largest market for autonomous vehicles, with the cost of autonomous driving systems including sensors, computing platforms and software falling rapidly after 2023, with Robotaxi expected to reach a commercial inflection point between 2025 and 2027.

A number of people interviewed by China Securities News believe that the large-scale commercialization of automatic driving also involves issues of laws, regulations and ethical norms. Autonomous driving requires the verification of more miles traveled over a long period of time to achieve reliability that far exceeds that of human drivers. "The current autonomous driving model is still 'man-machine co-driving', only a small-scale commercial pilot, and there is still a long way to go for large-scale commercialization of L4-level autonomous driving." Shi Jianping thinks.