Automatic Driving In The Second Half Of The Year?

There is good news in the field of autonomous driving. Recently, Tucson, a driverless truck technology research and development company, announced in the future that it has obtained Navistar's investment and reached a strategic cooperation with the other party. Both sides will jointly develop the L4 level driverless truck and strive for mass production in 2024.

This is the first time Tucson has reached mass production cooperation with truck manufacturers in the future. Previously, the company's business layout was mainly at the level of driverless freight service. Now, if the mass production plan can be realized, customers can directly purchase driverless trucks from Navistar's distribution agencies in the United States, Canada and other places.

As a matter of fact, Tucson has begun to cooperate with Navistar in the field of automatic driving technology for more than two years. In fact, it has begun to explore the value of unmanned driving technology with Navistar in the future.

This is a step-by-step landing technology for commercial vehicles. The trend is similar in the field of passenger cars. In addition to the auto driving mode announced by Toyota Motor Co., Ltd., more and more large-scale auto driving companies have stopped cooperating with BMW in the field of auto driving.

Although the development of autopilot technology companies is not smooth due to high R & D costs, long test cycle and other factors, some foreign head players such as nuro and zoox are seeking to sell themselves, but at the same time, some companies with sufficient resources are also ushering in more gratifying progress.

On the one hand, it is the support of funds; on the other hand, it is the trial operation and landing of related businesses. This year, domestic autopilot technology companies have accelerated the process of testing and research and development. Some time ago, Didi travel held a high-profile live broadcast of robotaxi (unmanned taxi), while enterprises such as autox and Wenyuan Zhixing also opened their driverless vehicles to the public in Gaode map.

It is worth mentioning that although the automobile industry has suffered from the significant impact of the new crown pneumonia epidemic and is still trying to recover from the downturn, the auto driving industry related to it has been unexpectedly catalysed during the epidemic. Some people who do business in the United States told the 21st century economic reporter that the epidemic did not have a great negative impact on them. Although some negotiations and cooperation were slightly delayed in the initial stage, the business volume was improved.

Grasp the operation landing

From the perspective of ordinary consumers, the sense of presence of autonomous driving is increasing this year.

In late June, Didi travel announced the opening of automatic driving service to the public in Shanghai. Users can sign up online through didi app. After passing the review, they will be able to call automatic driving vehicles for test ride experience in Shanghai's automatic driving test section for free.

In order to let more people understand the service, Didi also carried out a trial live broadcast with CCTV. Although the effect of this activity is not perfect due to bad weather and other factors, such as delay in the start-up and manual takeover in the midway, these are normal in the process of technical iteration.

In addition to Didi, other automatic driving technology companies have also accelerated the landing of driverless taxis. Also in late June, Wenyuan Zhixing, a L4 level automatic driving technology company, announced that it had reached a cooperation with Gaode map, and launched 20 robotaxi vehicles in Huangpu District and Development Zone of Guangzhou for pilotless trial operation.

In fact, Wenyuan Zhixing started to operate driverless taxis as early as 2019, but previously, it mainly relied on its own platform weride go. This access to Gaode map is equivalent to opening the driverless taxi service to a wider range of users.

The improvement of openness is the characteristic of this round of unmanned taxi landing. In addition to its own open driving service, Didi has accumulated a large number of open driving services in Shanghai.

For autopilot technology companies, the most important reason for accelerating the landing of related businesses is to promote technical iteration. The core algorithm of automatic driving needs massive user data support. If the driving route only stays in the test site, the data can not meet the requirements, which is also the advantage of Tesla as a vehicle enterprise.

On the other hand, the field of automatic driving is also entering the shuffle period. Tens of hundreds of enterprises in the industry have poured in huge amounts of capital, which is known as the "most money burning track". In recent years, they have also begun to calm down, and their investment amount and quantity have decreased significantly, and they have gradually concentrated on enterprises with more solid technology accumulation or more obvious cash flow channels. This has undoubtedly increased the pressure on all participants - either to advance or to retreat.

Compared with the influx of hot money in the previous two years, the field of automatic driving is not as lively as it used to be this year. Although several head companies have released good news of support to the outside world, including waymo's $3 billion financing, Didi's more than $500 million financing, and the new round of more than 460 million US dollars financing from smart bank, the industry as a whole is no longer as smart as before.

According to the statistics of prospective industry research institute, in the past few years, 2018 was the financing peak in the field of automatic driving. There were 107 financing cases in the global automatic driving industry, with a total financing amount of US $10.17 billion. In 2019, the total financing of the whole industry will drop to US $6.64 billion.

This year continues the trend of 2019, and the funds are more cautious. Starsky robotics, a star auto driving start-up company, has suspended its operation due to a lack of funds. Zoox, an old American self driving company established in 2014, tried to sell itself to help itself. Even cruise, an auto driving company backed by general motors, had to cut 8% of its employees to save clothes and food and cope with the cold winter.

Long tail problem to be solved

On the one hand, accelerating the trial operation and landing is on the other hand to realize the mass production and commercialization of driverless vehicles.

At the recent 2020 World Conference on artificial intelligence, Tesla CEO Elon Musk said he was confident of completing the basic functions of L5 level automatic driving this year. At present, even though the technology of L5 can be taken over by any auto driver, it is too optimistic that the technology of L5 can be taken over by any auto driver.

In other words, the L5 can't really realize the automatic driving function. Moreover, musk himself has also explained this statement. In fact, he does not mean to express that L5 level automatic driving function can be realized this year.

In his opinion, there is no underlying fundamental challenge to achieve L5 level automatic driving, but there are many details. "The challenge is to solve all these small problems and then integrate the system to continue to solve the long tail problem. When you solve most of the scene problems, there will be some strange situations from time to time, so there must be a system to solve the training and solve these strange scenes

The long tail problem is the obstacle to the large-scale landing of automatic driving technology. In contrast, 20% of the time spent in improving the driverless taxis was spent in the company's driverless taxi service, while the remaining 20% of the time spent in improving the company's driverless taxis was spent on improving the company's driverless taxi service.

The long tail problem refers to those more trivial problems that must be solved in addition to the basic driving problems. "Some problems will only be exposed in the actual operation, and they will also be the problems that may be encountered in the real road in the future, so it will take a lot of time and energy to solve them." The people said.

This is almost the consensus in the industry. Recently, Lou Tiancheng, the co-founder and chief technical officer of the company, also said in a media sharing that the automatic driving has not yet reached the L4 product level application, and the main difficulties lie in the technology, including the technical requirements for safety, complex scenes, deep integration of vehicles in complex weather and vehicle public safety. Complexity still points to unpredictable scenarios.

The discovery and data collection of these scenarios is a challenge for all autopilot companies. Objectively speaking, it is difficult for Tesla to find these rare scenes with massive driving data, and it is obviously more difficult to meet the unmanned taxis that pick up and pick up passengers in the demonstration operation area. However, these scenes can not be covered as much as possible. Theoretically, there are always potential safety hazards in automatic driving.

Li Lin, deputy chief engineer of Shanghai International Automobile City (Group) Co., Ltd., who participated in the design and construction of test scenarios in Shanghai Jiading test area, believes that the wide application of high-level automatic driving will take many years. There are two main directions for the commercialization of automatic driving in the short term. One is partial automatic driving in all regions, such as the L2 level intelligent driving assistance system Second, automatic driving in some areas, such as driverless taxis, can only be limited to a certain range. (Editor: Zhang ruosi)