L3 application will soon arrive _ Li Yanhong's ticket or history

The car is very much now, but it is still very stable and it feels very good. On July 5, 2017, when Li Yanhong rushed to the Baidu AI Developers Conference, the live host and his video connection, he described the "driving" experience in the car to the audience. The video shows that Li Yanhong does not Sitting in the driving position, but sitting in the co-driver. Let's take a look at the relevant content with the car electronics editor.

This "unmanned" news has caused people's exclamation and expectation, and the "collision" with reality has become very interesting: the relevant person in charge of the Beijing Municipal Public Security Bureau Traffic Management Bureau said that China's current traffic laws do not allow unmanned people. Driving the vehicle on the road, they have already investigated this. In the end, Li Yanhong got a ticket.

Baidu established the Intelligent Driving Business Group (IDG) in March 2017, which invested a lot of money in the development of autonomous driving technology. At the same time, various auto manufacturers are also making efforts. General Motors recently introduced super-cruise technology, and Audi also has the L3-class autopilot function on the new A8.

However, the industry's discussion of autonomous driving technology is still endless, and many of them are singers. “Automatic driving technology is really difficult to achieve automatic driving without driver driving. Autopilot technology is 'strong intelligence', and in the case of complex road conditions in China, it must be 'super smart'. Yang Ge, founding partner of John Capital, believes that compared with other countries, the relatively simple road conditions are different. Most of the roads in China are mixed vehicles, non-motor vehicles and motor vehicles. "Strong intelligence" can no longer meet such complicated road conditions.

Regarding the degree of automation of autonomous driving technology, the industry has a clear definition. Internationally accepted standards divide the degree of automation into 6 levels from L0 to L5, of which L0 is pure manual manual driving and L5 is the most automated automatic driving. At present, the auto-cruise function of some auto manufacturers belongs to the L1 and L2 categories, that is, the auto-driving technology with lower automation. Jia Xinguang, executive director of the Automobile Circulation Society, told the China Economic Weekly that the cruise functions at the L1 and L2 levels can only be identified by the speed of the previous car, and then follow the driving, while the L5-level autonomous driving technology is completely system-based. Controlling the car, all situations can be dealt with, which requires the car to identify and identify the external environment.

On February 2, 2018, the Beijing Municipal Commission of Communications jointly issued the "Beijing Autopilot Road Test Capability Assessment Contents and Methods (Trial)" (hereinafter referred to as "Method") and related technical documents, which is the first in China. A landing measure that recognizes and legalizes autonomous driving technology.

Jia Xinguang said that the "Methods" have made some auto manufacturers who are actively deploying autonomous driving technologies have technical standards that can be followed, which will promote the development of autonomous driving technology.

It is worth noting that the same test subjects as the ordinary human driver in the Method, such as side parking and curve driving, are more difficult. For example, the side parking time for storage and storage should not exceed 1.5 minutes, and the time for re-entry and storage can not exceed 3.5 minutes. Generally, the “old driver” can do it. In addition, the "Method" also added the examination content for traffic markings, signs and traffic police command gestures. Jia Xinguang particularly valued this: "The most difficult thing about automatic driving is the ability to recognize. The identification of the environment has caused problems. Adding the ability to identify the traffic police commanding gestures is an improvement in security."

Professor Li Keqiang, director of the Intelligent Network's Automotive and Transportation Research Center of Tsinghua University, told the China Economic Weekly that the current laws of the country allow for the use of L1 and L2, namely cruise and super cruise technology, and problems arise by the driver. With full responsibility, Beijing's "Methods" test object is the automatic driving technology of L3 and above.

He said that the most dangerous part of autonomous driving technology lies in the blind spot of the sensor, especially the temporary traffic control of the road, the traffic police command, etc., all of which are dangerous to drive automatically. "It is correct to set higher limits for autonomous vehicles. Automated driving technology should be safer than human drivers."

In addition, in an emergency, the human driver must quickly take over the vehicle, can not be allowed to auto-driving technology "Hu Lai", has also been written into the "method", the industry generally believes that this is necessary.

Professor Li Keqiang, director of the Intelligent Network's Automotive and Transportation Research Center at Tsinghua University, said that many companies' autopilot tests are now in the gray area of ​​the law. The Methodology helps to clarify some related issues. However, L3 can only be tested on the road under the current legal framework of the United States and China, and it will take time to actually use it.

Then, when the car reaches the L5 level, can it not only automatically drive on expressways and highways, but also freely and safely “drill alleys” in Beijing? “The situation in Hutong is indeed more complicated than highways and requires more powerful technologies, such as Automotive AI technology. The application of L3 will come soon, but the application of L4 and L5 will take some time, waiting for the development of technology and regulations.” Professor Li Keqiang, director of the Intelligent Network Automotive and Transportation Research Center of Tsinghua University.