Text / Sansheng
Producer / Node Finance
"It's like a ghost's hand on the steering wheel.".
In 1925, the American radio equipment company Houdina Radio Control designed a radio driverless car, the New York Times reported at the time.
This is one of the earliest reports on driverless driving. In the nearly one hundred years since then, human beings have continued to explore the field of unmanned driving, especially after entering the era of mobile Internet, unmanned driving research has entered the fast lane. In recent years, there have been more and more topics about unmanned driving, and the news of domestic manufacturers such as Baidu and Huawei on the road has also been frequently reported.
In this regard, there are those who expect it, those who worry about it, and no matter what kind of attitude, this scientific and technological trend related to the revolution of human mobility in the future is coming with an irresistible attitude. Looking at the world, driverless technology has also become one of the hopes of the mainland to "overtake in curves" in the automotive field.
Based on data and industry research, this article attempts to answer the following three questions:
1. Where has unmanned driving technology developed?
2. Where is the mainland in unmanned driving?
3. What is the most reliable landing scene of unmanned driving?
/ 01 /
Human driverless
Where have you come?
After the radio driverless car in 1925, with the advent of computers, in 1965, John McCarthy proposed in this article "Computer-Controlled Cars" the idea of helping vehicles navigate the road by "entering data from television cameras and using the same visual input as human drivers."
This idea is very close to the current design of autonomous driving. However, although the idea of unmanned driving has been proposed for a long time, it is not going well on the ground.
This is because autonomous driving is not a single technology, but a collection of a large number of high-end technologies, involving automobiles, computers, network communications, video detection, lidar, artificial intelligence and other aspects of modern technology, which requires long-term technology accumulation.
Even in the first decade of this century, the development of driverless cars was still very unsmooth, and in the commercial context of the time, it seemed to exist only in the nihilism of science fiction movies. However, in the past decade, with the rapid development of mobile networks, new energy vehicles, big data algorithms and other technologies, unmanned driving has gradually begun to take root.
As the integrator of the non-modern technology industry, once the unmanned driving is landed, it will also bring great changes and impacts to the entire society and the automotive industry, and the business prospects are extremely broad. This is also one of the driving forces for many high-tech companies around the world to flock to.
At present, many consumers should often see the promotion of driverless technologies such as L2, L3 or L4 when buying a new car. According to the Society of Automotive Engineers' definition of the ability to drive automatically, it is divided into 6 levels: L0 represents no automation, L1 represents assisted driving, L2 represents partial automation, L3 represents conditional automation, L4 represents high automation, and L5 represents full automation.
Specific to this stage, unmanned driving is already in the stage of commercial development of L2, where L2 can also be called advanced driver assistance system function (ADAS). Whether in the domestic or foreign automotive market, products equipped with ADAS systems have become more common.
Among them, in the field of mass production driverless, Tesla is the most well-known one among the mass production vehicles in the L2 level. However, although Musk has high hopes for FSD (fully autonomous driving) and has made a Flag that it can reach the L5 level of unmanned driving in the past two years, there is still a considerable distance.
This is because there are still great challenges in the promotion of driverless technology for bicycle intelligent routes to L3 and higher levels. Here it is necessary to mention the two technical routes that currently exist in unmanned driving, one is the current mainstream bicycle intelligence, and the other is vehicle-road collaboration.
Although bicycle intelligent driverless driving has been developed for many years, the risk of insufficient coping ability and failure in specific scenarios is still very large. At a higher level, the reliability of driverless driving and the ability to cope with challenging traffic scenarios need to be improved.
In addition, in the field of bicycle intelligence, there is also a dispute between visual routes and lidar routes, Tesla prefers the former, while Huawei focuses on the latter.
In fact, according to a 2016 RAND think tank report, an autonomous driving system needs to be tested for 11 billion miles to reach mass production conditions. Even Waymo, the head company of L4 level unmanned driving, its test mileage is currently stuck at the level of tens of millions of miles, which is still far away.
Therefore, the current dispute over driverless routes based on the field of bicycle intelligence may still be in the semi-automatic driving stage of the L2 level, and it will take time to truly realize the ideal state of "no one behind the steering wheel".
In this case, the vehicle-road coordination route other than the bicycle intelligent route has received more and more attention in the industry.
/ 02 /
Where is China?
When it comes to the two routes of unmanned driving technology, we have to mention the pattern of unmanned driving in the world, and the most eye-catching is China and the United States.
Among them, the United States, as a veteran automobile power, has a deep accumulation in driverless technology, and its chosen technical route is still dominated by bicycle intelligence. As the fastest growing country in the automobile market, China currently attaches more importance to vehicle-road collaboration technology, and actively seeks opportunities for "overtaking in curves".
From the perspective of the company's layout, the three strongest in the field of unmanned driving are Google's Waymo, GM's Cruise and Baidu Apollo.
Among them, Waymo and Cruise started earlier, especially the former, backed by Google, is a veteran company in the field of unmanned driving. According to the California Vehicle Authority (DMV), Waymo and Cruise are still at the forefront of the industry, with a total road test mileage of 20 million kilometers in 2021.
From the actual landing point of view, Waymo, Cruise, etc. have obtained the commercial license for self-driving taxis issued by the California Public Utilities Commission, and are in the head position in the field of L4 level unmanned driving.
However, although Waymo and Cruise are strong, Apollo from Baidu is not weak. According to the California Vehicle Authority, four of the top ten in 2021 are already from China, with Baidu Apollo coming in first and Wamyo in second. In addition, AutoX and Ma Xiaozhixing entered the top five, and Didi entered the top ten.
In the field of unmanned driving, Chinese companies have gradually grown into a rival to U.S. companies.
As a latecomer, the reason why domestic companies can quickly follow up, first of all, in the field of funds, compared with the US market, the domestic market's enthusiasm for investment in unmanned driving is still there, and there are enough funds to support long-term investment projects such as unmanned driving.
Secondly, compared with the big players such as Google and Tesla that already exist in the US market, although Baidu has taken the lead in the domestic driverless field, it has not yet reached a monopoly position on the market, and the opportunity for new players is still there.
Third, car consumers around the world have different levels of acceptance of driverless technology, and survey data shows that domestic consumers' acceptance of driverless cars ranks among the best in the world's major auto markets, which gives related companies more room to play.
Finally, domestic players not only catch up with foreign opponents in terms of test mileage, but more importantly, in technological innovation, they have gone out of their own way.
Let's go back to the road-to-road coordination route mentioned earlier. In fact, Waymo, which is based on Google's high-precision map, is also using vehicle-road collaboration technology to some extent, but in contrast, Chinese companies such as Baidu Apollo have come to the forefront in this regard.
Vehicle-road collaboration is actually based on bicycle intelligence, through the Internet of Vehicles to organically combine traffic elements such as "people-vehicle-road-cloud", and systematically upgrade the environmental perception, calculation decision-making and control execution capabilities of driverless vehicles to accelerate the maturity of unmanned driving technology.
Therefore, as an upgraded version of the bicycle intelligent route, the application scope of vehicle-road collaboration is more extensive, which can also be confirmed from Baidu Apollo's business.
At present, in addition to the driverless direction targeting RoboTaxi (self-driving taxi), Baidu Apollo is also helping mass production vehicles achieve avP and ANP (benchmark Tesla FSD) for intelligent driving. In addition, Baidu Apollo also tries to integrate Baidu's innovation in autonomous driving through Jidu Automobile to bring driverless technology to the market for the first time.
Of course, in addition to Baidu Apollo, well-known players in the domestic driverless field also include Huawei. Although the latter entered the market late, with Huawei's technical strength and corporate culture, no opponent dares to ignore its existence.
So, in today's unmanned driving development, in which scenarios can the real "no one behind the steering wheel" be realized?
/ 03 /
Driverless
What is the most reliable landing scene?
A recent document released by Beijing gives the public the opportunity to experience this L4-level driverless experience in the Yizhuang area.
On April 28, the "Implementation Rules for the Management of Unmanned Road Testing and Demonstration Applications of Passenger Cars in the Pilot Zone of Beijing Intelligent Connected Vehicle Policy" was officially released, and Baidu's Radish Run and Xiaoma Zhixing became the first batch of enterprises approved to carry out unmanned manned demonstration applications, and Beijing also became the first city in China to open a pilot project for the operation of passenger cars "unmanned behind the wheel".
This means that in Yizhuang, Beijing, the public will have the opportunity to hit a driverless car with "no safety officer in the main driver's seat and safety officer in the co-driver' seat" through the Radish Run App or the PonyPilot+ App.
Users can use the APP to call a car at a fixed station, scan the code to open the door after waiting for the car to come, start the vehicle on the display screen in the car, and then the vehicle will be unmanned throughout the whole process, realizing a series of operations such as starting, accelerating, braking, turning, and stopping.
Robotaxi is currently the highest stage of shared mobility development and one of the most expansive scenarios for high-level driverless vehicles in the future. In this field, China has been launched on a large scale since 2020, and the main participants include Baidu Apollo, Xiaoma Zhixing, Wenyuan Zhixing, AutoX, etc., and have made certain progress.
Taking Baidu's Radish Run as an example, as of February 2022, Radish Run has opened manned test operation services in seven cities: Beijing, Shanghai, Guangzhou, Shenzhen, Chongqing, Changsha and Cangzhou, with a total fleet size of more than 500 vehicles. In addition, the AutoX RoboTaxi fleet has exceeded 1,000 units, becoming the largest fully driverless RoboTaxi fleet in China and even the world.
As the representatives of China and the United States in the field of unmanned driving, Google's Waymo and Baidu Apollo have entered the operational stage. However, for driverless driving, safety remains a top priority. Especially in large domestic cities, high-density urban core areas, commercial areas, and residential areas are mixed, and peak travel congestion has become normalized.
All this means that the realization of unmanned urban roads in China will face the challenge of almost the greatest difficulty in the world.
However, it is in this environment that the domestic driverless industry has still made obvious progress. At the current speed of development, china's realization of "curve overtaking" in the field of unmanned driving in the future is very worth looking forward to.
Of course, as mentioned earlier, unmanned driving is an aggregation of a variety of technologies, whether it is the fusion of the three major sensors, the vehicle communication module, the Internet terminal, the communication service and the maturity of the decision chip and algorithm, and the final realization of vehicle-road collaboration, it needs a long way to go.
Halfway through the car, the victory and defeat are not divided, but the future can be expected.
Node Finance Statement: The content of the article is for reference only, the information in the article or the opinions expressed do not constitute any investment advice, and Node Finance does not assume any responsibility for any action taken as a result of using this article.