Autopilot, who can laugh at the end?

American research company Navigant? Research institutions released the annual ranking of the development of autonomous driving technology on Monday. According to production, strategy, market, marketing, ability, endurance and reliability, nearly 24 related companies are ranked again.

Among them, Waymo ranks the highest in strategy and execution, and is classified as the "leader" part of Navigant on the coordinate axis. As can be seen from the figure, Google is ahead of Ford's autonomous driving business, and these two companies are ahead of Cruise of General Motors, Baidu of China and Intel -Mobileye.

The next category is "competitors", including Daimler (with Bosch), Hyundai (with Aptiv), Toyota and Volkswagen. And Zoox and May? Russian start-ups such as Mobility and Yandex, Yandex first appeared in Navigant's ranking.

The penultimate "challenger" category represents companies that have fallen behind all the above companies so far. There are also some familiar faces here, such as Volvo, Renault, Nissan and Mitsubishi Alliance. Most importantly, this is the location of Tesla.

The immediate and long-term worries of autonomous driving

Tesla fans need not be shocked by the countdown. After all, Tesla is in this position all the year round. The answer given by Navigant is: "Tesla's high-profile commitment includes having 6,543,800 vehicles with autonomous driving capability on the road by the end of 2020"? Navigant's report wrote. "However, the performance of its system is still inconsistent, and its products do not conform to its proposed mobile business model."

In fact, some well-known companies have not even entered Navigant's ranking. For example, Apple lost because Navigant said that the company did not provide "a clear view of any commercialization plan". Although Uber is trying to resume its autonomous driving business as quickly as possible, one of its cars has a fatal collision with a pedestrian. Raffles have not made much progress in recent months, and Navigant is worried about whether self-driving cars will really bring profits to taxi companies.

It is worth mentioning that this study focuses on the ongoing technological development, and the high ranking does not necessarily mean that these companies will produce feasible self-driving cars within any specific time range. Like Waymo? A limited range of pilot projects like One can still bring hope, and the public should better understand who is really making efforts in the field of autonomous driving.

More remarkable than the ranking, last month? According to the latest research released by Information, in the past few years, a team of 30 companies spent at least $654.38+06 billion on developing self-driving driverless cars. So far, their income (if any) is very small. However, billions of dollars may be needed before the mass production of this technology is on the road.

According to calculations, three of them spent half of the money, namely Waymo of Alphabet, Cruise and Uber of General Motors. Four other companies, including Apple, Baidu, Ford and Toyota, spent most of the rest. Only a few of these enterprises (the most famous ones are Alphabet, Toyota and Apple) have huge financial resources to ensure long-term development. However, Apple has withdrawn its investment.

It seems that this promising technology is also a real money-burning machine.

Then why are these companies willing to continue to overweight in the field of autonomous driving? On the recent investor day, Cruise? CEO？ Dan. Amman described the future self-driving cake, saying that its market scale can reach 8 trillion US dollars per year, including 5 trillion US dollars in the passenger transport market, 2 trillion US dollars in the freight market, and at least 500 billion US dollars in in-car experience and data mining opportunities.

According to available data, Americans spend about 50 billion hours driving every year. If the average wage is $35 per hour, the value of this part of productivity can be as high as $65438 +0.7 trillion. It is roughly estimated that by using driving time, the world can release at least 5 trillion US dollars of productivity.

After self-driving cars are on the road on a large scale, their driving skills will continue to improve over time. In 20 14 years, the data given by the expressway Safety Administration showed that the annual loss caused by traffic accidents in the United States was as high as 87 1 100 million dollars. Globally, this figure may reach 3-4 trillion US dollars. Therefore, even if the accident can be reduced by half, it can also bring huge benefits.

In California, Waymo, a sister company of Google, and Cruise, a general motors company, are obvious leaders. These two companies are far ahead of other companies in mileage and disengagement. Waymo just announced a total mileage of 20 million miles, most of which is not in California. In 20 18, Waymo traveled10.2000 miles in California, and the separation rate per 1000 miles was 0.09 times. The second place is the cruise of General Motors, about 500,000 miles, and the detachment rate per 1000 miles is 0. 19 times.

The Rise of Enterprises in China

Following the release of the annual report on self-driving cars by the California Automobile Department, Beijing Mobile Intelligent Innovation Center (BICMI) released its investigation report on self-driving cars tested on local roads in 20 19. Apart from Pittsburgh, Pennsylvania, Beijing is one of the few cities in the world that requires self-driving car companies to disclose their mileage, fleet size and failures of their autonomous driving systems.

Baidu, Weilai, Beijing New Energy, Daimler, Ma Xiao Zhixing, Tencent, Didi, Audi, Chongqing Jin Kang, Siweituxin, Toyota, Beijing Sankuai, etc. 13, 77 self-driving cars of China companies, driving10.04 million kilometers, much higher than that of 8 companies in 20 18.

BICMI's report is part of a recent white paper issued by China's National Development and Reform Commission, Ministry of Industry and Information Technology and nine other "ministerial" institutions outlining China's strategy to promote the development of driverless technology. As early as 2025, China intends to formulate guidelines around infrastructure, supervision and safety, so that enterprises can "mass produce vehicles that can drive independently conditionally" and "commercialize" ... self-driving cars.

By the end of 65438+February, Beijing had allowed self-driving cars to test 503.68 kilometers on the road 15 1, and the test area in the urban area increased by about 40 square kilometers. China has stipulated five levels of automatic driving test licenses, from T 1 to T5, similar to the automation level issued by the American Society of Automotive Engineers.

In July last year, the Beijing Municipal Transportation Commission allocated the first batch of T4 independent test licenses to Baidu. At that time, the company pointed out that T4 was the highest-level license in China, and it was an open road test license, enabling it to deploy driverless cars in urban roads, tunnels, school areas and other places.

At the end of 20 19, Baidu announced that it had obtained 40 licenses to test driverless cars on designated roads in Beijing, making it one of the first companies in Beijing to do so. The company is trying to surpass its well-funded competitors, such as Tencent, Alibaba and Ma Xiao Zhixing.

In April this year, Alibaba confirmed that self-driving cars had been tested with the goal of reaching L4-level autonomous driving ability, and expressed the hope to recruit as many as 50 engineers for its AI research laboratory. In May of this year, Tencent obtained the permission of the government to start testing self-driving cars in Shenzhen. Just last week, Ma Zhixing? Raised $462 million in venture capital with a valuation of $3 billion.

Baidu and its competitors are competing for a veritable market gold mine. According to a report by McKinsey, by 2030, the value of driverless cars and travel services in China will exceed 500 billion US dollars, and there will be 8 million self-driving cars on the roads.

Recently, the California Motor Vehicle Administration released the data report on the separation of self-driving cars in 20 19, and the total test mileage of enterprises that conducted self-driving tests in California in 20 19 was about 2.88 million miles. Among the top ten companies, four are from China, and Baidu has become the number one company, while last year's "champion" Wamyo ranked second. Apple ranked at the bottom last year and ranked 13 this year. Car companies such as BMW, Toyota, Nissan, Mercedes-Benz and SAIC are also conducting autonomous driving tests in California. As in previous years, the "achievements" of these car companies are relatively backward.

Although many people in the industry believe that the reported data can be used as a positive signal for the commercialization of autonomous driving, it cannot be a myth. The reason is that even simple driving operation, in fact, the operation of automatic driving program is far more difficult than it sounds.

Especially on the basis that the data collected by self-driving cars are very expensive, some small probability events rarely really happen, so they may exceed the depth perception of the simulation environment of the test vehicle, because it is impossible to encounter such a situation in the test data. In other words, modern machine learning system has good performance when there is a large amount of data, but it has poor performance when there is only a small amount of data.

Some researchers believe that it is difficult for humans to have a wide range of self-driving cars unless the roads are significantly changed. This is not only an expensive expense, but also requires nationwide coordination, so the improvement of the traffic environment should be before the widespread application of self-driving cars, not after. Of course, before all this, and most importantly, relevant laws and regulations are needed to determine the responsibilities arising from autonomous driving.

Text/Yang Jing

-

Wechat searches for "Car Commune" and "One-sentence Comment" to follow the official account of WeChat, or visit the Daily Auto News Network for more industry information.

This article comes from car home, the author of the car manufacturer, and does not represent car home's position.