The crowd is chasing the deer, and the smoke of the autonomous driving battlefield is gradually rising!

Author 丨 Zhang Jie

An article to learn about the frontier developments and challenges of the autonomous driving industry in 2021.

Compared with 2020, the autonomous driving industry in 2021 has a more exciting "story".

Although the real intelligent networked car is far from us, the ideal sense of automatic driving is still out of reach, in 2021, automatic driving is still presented with its own gifts and smoke for this era.

The automotive industry in the general environment: two major external factors have a far-reaching impact

Looking back at 2021, counting the many events affecting the automotive industry, "carbon neutrality" and "chip famine" will inevitably be on the list.

Achieve carbon peaking by 2030 and carbon neutrality by 2060. Carbon neutrality, as a goal and action to combat climate change, has been included in the outline of China's five-year plan for national economic and social development, and has also become a hot topic of public discussion, affecting the future of countless industries. Driven by policies and markets, the automobile industry is constantly spawning new changes in the three directions of greening, intelligence and Internet.

With the maturity of autonomous driving technology, the iterative landing of on-board chips and vehicle operating systems, and the transformation of the traffic environment of vehicle-road coordination will greatly reduce the probability of traffic accidents and traffic congestion, improve the utilization rate of resources and energy and transportation efficiency, and then comprehensively reduce carbon emissions. This has been proven on the country's first open 5G commercial vehicle road collaborative demonstration road in Shunyi, Beijing, in 2019.

According to the data released by the government of Beixiaoying Town in Shunyi District in June 2020, the safety of intersections in this section has increased by 60%, and the traffic efficiency has increased by 20%, which is equivalent to two lanes to achieve the traffic efficiency of three to four passages. In March 2021, the Hengyang Municipal People's Government signed a strategic cooperation agreement with Mushroom Car Union, in which the two sides carried out in-depth cooperation in the fields of intelligent terminals, vehicle-road collaboration, automatic driving and intelligent transportation, and jointly built a smart transportation innovation demonstration city.

If the "carbon neutrality" policy has a certain role in promoting the development of the autonomous driving industry, then the "chip shortage" is undoubtedly a major constraint.

Chip shortage is the result of a series of factors, including the spread of the new crown pneumonia epidemic, the sudden fire of chip manufacturers, the strike of chip workers in chip companies, and the mismatch between the demand of the automotive industry and the cycle of the chip industry. According to the latest data from AutoForecast Solutions, as of October 10, 2021, the global automotive market lost 9.345 million units and the Chinese automotive market by 1.827 million units due to chip shortages.

The lack of production capacity directly affects the sales volume of various car companies, especially small and medium-sized car companies, and the blow caused by the suspension of production and work will be more serious. Many car companies struggling in the "chip famine" have realized that they cannot rely too much on imports in the supply chain, and they should also find suitable chip suppliers in the local area, and even consider self-developed chips when conditions permit. It is foreseeable that chip controllability will play an important role in the smart car track.

In 2021, more car companies have accelerated their layout in the chip field. Geely Holdings' Yijiatong Technology announced that it has reached a cooperation with Arm China to establish a joint venture company, Core Engine Technology, focusing on long-term research and development and mass production plans in the fields of chips such as autonomous driving, microcontrollers and smart cockpits, and launched the first 7nm vehicle-grade chip in China at the end of the year. Great Wall Motor, BYD and Dongfeng Motor chose to strategically invest in chip company Horizon and enter the chip industry. It is believed that in the future, whether for safety considerations or technical considerations, more and more car companies will choose to independently develop chips.

Developments and Challenges: Autonomous Driving Enters the "Warring States Era"

Compared with previous years, the autonomous driving industry in 2021 is more "lively". First, more familiar technology companies have joined the track, and the other is to find more feasible paths in the exploration of commercial models. The autonomous driving industry is entering a Warring States era of "rising together and chasing deer in the Central Plains". And this "era" presents three distinctive characteristics:

First, all kinds of red and purple Dou Fangfei: cross-border players enter the variable

With the development of the automobile industry to electrification, intelligence and networking, in addition to traditional car companies, some technology companies have also begun to actively invest in it. Some choose to cooperate with traditional manufacturers, such as Huawei; some choose to build their own cars, such as Xiaomi; some choose to create an ecological comprehensive layout, such as Baidu. Even the news that Apple has been secretly making cars has gradually increased.

Empowering car companies, representing enterprises: Huawei.

Xu Zhijun, Huawei's rotating chairman, said at the Huawei Analyst Conference held in April 2021 that Huawei's decision not to build cars and help car companies build good cars has not changed so far. Huawei has reached cooperation with a number of car companies: the Jihu αS HI (Huawei inside) version custom developed with BAIC was the first to be announced; Avita Automobile, jointly built with Changan and NINGDE Times, has also demonstrated its first product C11; and the Qijie M5, jointly developed with AITO, a high-end brand of Xilix Automobile, is the first car equipped with HarmonyOS intelligent cockpit.

Wholly-owned models, representing the company: Xiaomi.

On March 30, 2021, Xiaomi announced the establishment of a wholly-owned subsidiary to enter the high-quality intelligent electric vehicle business. Lei Jun announced at the press conference that he plans to invest $10 billion in 10 years, calling it "betting on the entire reputation of his life to fight for Xiaomi Cars" Wholly owning a car means a higher difficulty factor and greater autonomy.

On the one hand, Xiaomi's investment in batteries, chips, smart parts and autonomous driving technology has further increased throughout the year, and constantly improved its automotive industry chain; on the other hand, in terms of car manufacturing progress, Xiaomi Car Manufacturing Company was established in September, injecting 10 billion yuan, and Lei Jun personally took command. According to the latest news, smart electric vehicles are expected to be officially mass-produced in the first half of 2024.

Multi-line and multi-line, representing the enterprise: Baidu.

Based on the report card handed over by Baidu in 2021, it is not difficult to see its efforts in the field of autonomous driving around the ecological construction of multiple lines and actively testing the water.

In the field of automotive intelligence, Baidu Apollo's AVP, ANP and Xiaodu vehicle landing scale has been further expanded, demonstrating Baidu's strength in intelligent driving and smart cabins. In the field of autonomous driving travel services (robotaxi), the Apollo Moon, a customized model with Jihu, will control the cost of each robotaxi to 480,000, entering the cost range of manned taxis, showing the possibility of large-scale application of robotaxi.

In terms of forward-looking concept, Baidu and Geely's joint venture Jidu Automobile proposes the concept of "automotive robot", which combines the comprehensive technical strength of Baidu AI and intelligent driving and other fields, and will announce the first concept car at the 2022 Beijing Auto Show, and plans to mass-produce it in 2023.

Overall, the entry of cross-border players has undoubtedly added many variables to the "battlefield" of autonomous driving. The difference in the way of admission also shows the prediction of different companies for different technical routes and business prospects.

Second, a few joys and a few sorrows: mass production landed, the effect began to appear

When many new and old players are fiercely fighting and competing in the same field, who can find a stable path for commercial landing faster will occupy the initiative.

In 2021, with the promotion of new infrastructure, smart travel and other strategies, many cities in China began to accelerate the progress of 5G, edge computing, and vehicle-road coordination, and the basic software and hardware system on which automatic driving relied was gradually improved; on the other hand, under the pressure of the epidemic, "no one contact" just needed to emerge, and unmanned vehicles as mobile transportation tools began to enter the work and life of the public, and the concept of "automatic driving" had a wider audience.

In this context, many autonomous vehicle companies and technology companies have begun to increase mass production. All kinds of L2-L4 level autonomous vehicles out of the closed road test site, delivered on the road, "mass production" seems to have become the keyword of commercial trial and error, the success of mass production has also brought a series of chain reactions.

L2 autonomous driving penetration continues to grow, and the penetration of advanced driver assistance functions (ADAS) continues to increase

The penetration rate of L2 autonomous driving is constantly increasing, and the market competition is becoming increasingly fierce. According to data from the Ministry of Industry and Information Technology, as of the third quarter of 2021, the market penetration rate of L2 passenger cars in China has reached 20%. In addition, from the equipment rate of specific intelligent driving functions, the equipment of basic ADAS functions and sensor hardware such as fixed speed cruise, front-and-rear reversing radar has exceeded the equipment rate of 20%. Moreover, the proportion of L2-class models in the insurance volume of new energy is also increasing. According to Cybercar's calculations, the insurance volume of L2-level intelligent networked vehicles in November 2021 was 151684, up 57.83% month-on-month, accounting for 41.62% of the insurance volume of new energy, much higher than the 33.05% in October.

L3 automatic driving policy has become the key, and the access regulations of Germany and Japan have been opened

Level 3 autonomous driving, regardless of the standard, requires machines and humans to drive the vehicle together. Ethical controversies about human-machine takeovers continue to exist, and most countries have not issued clear regulations on the attribution of safety responsibilities, which has led many companies to dare not risk betting on L3 mass production. However, in 2021, under the background of the opening of access regulations in Germany and Japan, two multinational automakers, Mercedes-Benz and Honda, announced the mass production of L3-level autonomous driving systems. Recently, the South Korean government also announced the "Roadmap for the Innovative Development of Autonomous Driving Regulation" at the end of the year, planning to launch L3 self-driving vehicles in 2022.

L4 autonomous driving mass production is mainly commercial vehicles, and passenger cars are open to commercialization pilots in the Robotaxi scene

L4 autonomous driving is a high-level, leapfrog technology, and there are fewer concerns about using the road conditions and low accident rates, so it is easier for commercial vehicles to achieve L4 large-scale landing. For example, in the field of unmanned distribution, the self-driving enterprise Mimu Zhixing, which was incubated from Great Wall Motors, achieved large-scale mass production and delivery of the L4-level unmanned delivery vehicle "Little Wild Donkey" produced in cooperation with Ali Damo Academy, and received nearly 1 billion yuan of financing from the A round led by Meituan at the end of the year.

In addition, in the field of passenger cars, there will be breakthroughs in 2021. On November 25, Beijing officially opened the first robotaxi commercialization pilot in China in the form of policy permission, which means that Robotaxi operators can open charging services to passengers through a market-based pricing mechanism. Baidu Radish Run Platform and Xiaoma Zhixing became the first batch of enterprises to be approved to collect fees.

The dream of L5 automatic driving has become empty, and the reality of "landing difficulty" has not been broken

Musk, who once promised to launch L5 fully autonomous driving in 2021, has once again blown through the cowhide. Looking at the world, L5 level is still full of unknown "no man's land", and its required technical level, burning speed and time cost have once again forced some contestants who try to brave no man's land. Following Uber's sale of its self-driving division, another ride-hailing platform company, Lyft, also announced in April 2021 that it would sell its self-driving car division, Level 5, to Toyota's subsidiary, Woven Planet.

At present, the development of autonomous driving can be divided into two routes, one of which is to gradually move towards high-level automatic driving. The advantage of progressive players is that they can take the lead in mass production sales and exchange market scale for technological growth space; the other is a leapfrog player who has insisted on L4 level automatic driving from the beginning, and its advantage is that automatic driving technology is more mature, safe and reliable, and at the same time as mass production of unmanned vehicles, L4 capabilities can be reduced to L2 or L2+.

However, no matter which route is taken, many players on the autonomous driving track are aiming to accelerate mass production delivery and achieve commercial operation. The sooner mass production is commercially realized, the more obvious the accumulation of data advantages and algorithm advantages, and the "Matthew effect" of the autonomous driving industry will inevitably appear, and the industrial pattern will also begin to emerge.

Third, the tip of the needle to Mai Mang: the eternal "security" battle

With the landing of large-scale landing, whether it is passenger cars or commercial vehicles, the safety of automatic driving is the top priority for business advancement, and the public will inevitably have more and more stringent expectations for safety issues.

Without the premise of "safety" for research and development design, autonomous vehicles are "smart" is also in vain. Therefore, autonomous passenger cars generally undergo long-term closed road tests and public road tests before they are officially delivered on the road. But in fact, the safety controversy surrounding "autonomous driving" has always been endless. The 2021 WEILAI autonomous driving incident has once again pushed this problem to the outlet of public opinion.

On August 12, Lin Wenqin, founder of Shangshan Ruoshui Investment Management Company, founder of Yitong Tianxia Catering Management Company, and founder of Meiyihao Brand Management Company, drove a Weilai ES8 car in a traffic accident on the highway and unfortunately passed away at the age of 31.

After the accident, the victim's family and the car company have their own opinions, the family emphasizes that the vehicle starts the automatic driving function at the time of the accident, and the responsibility must not all belong to the victim; the car company points out that the "pilot mode" of the accident vehicle has only the function of assisting the driver, and the driver must still pay attention to the surrounding environment during the driving process of the vehicle. The difficulty of the controversy is whether the car company has fully informed the victim of the risks and limitations of the L2 level autonomous driving system.

Determining the attribution of liability after an accident in a car with autonomous driving capabilities is not an easy task. According to the relevant provisions of the Ministry of Industry and Information Technology's "Automobile Driving Automation Classification", in the automatic driving of the L0-L2 level, the driver fully grasps the exercise of the car, and the automatic driving system only plays an auxiliary role. It is generally believed that the driver's duty of care at this time is equivalent to that of driving an ordinary car, and the driver is responsible for tort liability and criminal risk.

Even so, in fact, whenever an accident occurs in a vehicle under the name of "autonomous driving", it will inevitably lead to a crisis of trust. Therefore, car companies need to be cautious when promoting the automatic driving function, and have a clearer definition of the degree of "automatic" and the functional boundary. For higher-level autonomous vehicles, in order to promote industrial development, it is necessary to further improve laws and regulations, clarify responsibilities and obligations, and reduce the occurrence of ethical disputes and moral dilemmas.

epilogue

Overall, the autonomous driving market is becoming more pragmatic and prudent, and more advocating a step-by-step approach to explore a commercial landing path. The opening of unmanned commercial space requires the main engine factory and technology companies, upstream and downstream of the industrial chain, and multi-party cooperation to form a large-scale market demand and development space. In the field of autonomous driving, who can win the throne in the unprecedented change is not yet known, but when the shuffle is nearing the end, the head aggregation effect will inevitably gradually increase.

The CTO says live trailer