TU-TESL

• 22

The picture comes from "specific authorization" at, L3 has "no smell". Although on March 19, in response to the yiou automobile, it said that the research and development project of level 3 automatic driving technology was not cancelled, but it was upgraded to the group level and integrated into the car software Vehicle Development Department of car.software. But it's clear that the L3 technology route that once dominated is being marginalized. I still remember that great afternoon in July 11, 2017, great in strength and impetus was unveiled by Audi motor in Barcelona, Spain. The new A8 was unveiled, and this model was famous for its L3 class automatic driving system. This model is also defined as the first mass production model developed for highly automated driving, which was once considered as the "key card" of Audi. Now, two and a half years have passed. The hardware promised by Audi officials to support L3 level automatic driving will be opened to the public in the future, but this commitment has never been fulfilled. This is also an important reason why Audi is rumored to give up L3 level automatic driving. Although the official response has not been cancelled, Audi confirmed in its response on March 17 that it is strengthening the research and development of L4 level and above automatic driving technology. Not only in Audi, but also in the whole automobile industry, the research and development of L3 level automatic driving is gradually being marginalized. The reason behind this may be simple: it's hard to land. In fact, L3 level automatic driving technology has been a controversial issue in the industry. Many enterprises have already abandoned this route. "We abandoned the transition phase of driver assistance technology and decided to provide L4 level fully autonomous vehicles directly. The reason is that I don't know how to switch from mechanical control to manual control at L3 level. " Raj Nair, then president of North America, said at a conference in August 2016. Ford, waymo and waymo have said that due to safety risks, it is not meaningful to develop L3, and they will skip L3 level automatic driving technology. Gu Jianmin, then Volvo's R & D director in China, said straightforwardly, "under L3 level automatic driving, although the driver can give up at any time, he must be ready to take over at any time. Such a state of readiness violates the original intention of automatic driving." Tabulator / business analyst Zhang Nan of yiou automobile is divided into six levels of l0-5 according to the classification standard of automatic driving formulated by SAE. Among them, L1 and L2 have been widely used, and even many self owned brand models with prices less than 100000 yuan have L2 level driving assistance functions. Specifically, L2 can be regarded as auxiliary driving, which requires the driver to hold the steering wheel, look forward and maintain attention as usual. In general, it includes lane departure warning, parallel assist, active safety / active braking system and adaptive cruise. And L3 level automatic driving is called automatic driving under certain conditions. Most of the driving operations are completed by vehicles. Human drivers need to pay attention to deal with emergency and dangerous situations. L3 autopilot has always been seen as a technology to transition to L4 and L5. At the same time, L3 is divided into TJP (congestion automatic driving) and HWP (high-speed automatic driving). It is easy to understand that L3 can achieve "off hand", "off foot", "off eye" and "liberate" the driver's "conditional automatic driving" under certain conditions. At present, propagandizing that they are L3 self driving companies, they will take HWP, TJP and AVP as typical use scenarios of L3, and propagandize them as selling points. Compared with other levels of automatic driving, L3 level automatic driving is quite ambiguous, and there is a division of labor between people and machines. Sandwiched between assisted driving and L4 / L5, which has a relatively high degree of automation, the positioning is also "awkward". At the same time, L3 level automatic driving is also widely misunderstood to the outside world. The car owner once mistakenly thought that L3 level automatic driving can completely get rid of human operation, and eventually lead to traffic accidents. This proves that L3 level automatic driving is full of unexpected dangers from the side. In terms of responsibility division, L2 and L3 are obviously different. It is simple to understand that under L3 level automatic driving, if the vehicle has an accident, the responsibility will be shared by the technology provider, i.e. the main engine factory and the driver, while L2 level will be all borne by the human driver. This also makes many car enterprises bear great pressure. On the other hand, in order to flaunt their own strength in automatic driving, car companies began to say l2.5, which is closely related to the division of responsibility. Mass production and landing - it is also a gap that automatic driving enterprises cannot cross. There is a standard for mass production vehicles in the world. Only 50 vehicles can be called mass production. This means that when the vehicle really moves from the laboratory to the user, not only a lot of costs need to be invested, but also the technical stability and consistency need to be guaranteed. In terms of the domestic market, L3 level automatic driving is particularly attractive to Chinese car companies. In March 10th, its own brand car officially released its latest model UNI-T, which will be equipped with L3 automatic driving system. But the system limits its use to structured roads - no sudden semi closed scenes of pedestrians and vehicles. From the technical point of view, Chang'an uni-t / enterprise officials analyze that Chang'an L3 level automatic driving technology is composed of 5 millimeter wave radars, 6 cameras, 12 ultrasonic radars and ADAS maps. It can sense the environmental changes outside the vehicle body. The maximum measuring distance is 200m, and the accuracy can be reduced to 10cm. The official statement is that it can accurately identify pedestrians, obstacles, etc And so on. In Chang'an Automobile plan, L3 level intelligent open platform for automatic driving will be built by 2020, and L4 level intelligent driving products will be launched in 2025. The L3 products of other automobile enterprises are also on the way. Previously, Chinese brands such as,,, Dongfeng, and Hongqi also planned to achieve their L3 model mass production in 2020. New car manufacturers also announced that they would launch L3 level automatic driving products in mass production this year, and Weima set this time point in 2021. From the perspective of mass production time, 2020 can be regarded as the first year of L3 level automatic driving in China. In this process, many car companies think it is difficult to jump directly from L2 to L4. The L3 level can be used as a transition to display technology to industry and consumers, and to prove that it can develop high level automatic driving system. Another analyst said that skipping L3 level automatic driving directly means that the car companies will lose this part of the market share, which is more or less reluctant. But in fact, the L3 level automatic driving that the domestic automobile enterprises boast is about to land, which is deeply questioned by the industry. Some automobile enterprises have not reached the real L3 level standard in function. This time, Chang'an caused such disputes. Some people think that Chang'an didn't mention more scenes of automatic driving. The autopilot test on this display is considered as low-speed L3 and high-speed L2, which are between them, so it can not be regarded as L3 in a strict sense. In the view of yiou automobile, for car manufacturers, they are more concerned about what functions mass production models can provide and enhance their sense of technology. And L3 level functions may be redefined by domestic vehicle enterprises to pursue partial balance. It can not only attract the attention of the audience, but also ensure mass production to a certain extent. However, domestic car companies can't get around the same problem - just like Audi's rumored predicament, China's current policies and regulations don't seem to make car companies feel comfortable and bold. Zhu Huarong, President of Changan Automobile, also specifically mentioned policy restrictions on the day of release of L3 level automatic driving. He said, "with the permission of road conditions and policy factors, the L3 level automatic driving technology of Changan Automobile can achieve mass production within the fastest year." This means that as long as the regulations are not yet issued, the L3 class automatic driving vehicle will not be able to go on the road, nor can users enjoy this function. Tabulator / business analyst Zhang Nan of yiou automobile, but good news has come. On March 9, the official website of the Ministry of industry and information technology publicized the draft for approval of the recommended national standard of automobile driving automation classification, which is to be implemented on January 1, 2021. This version of the grading standard is based on the degree to which the driving automation system can perform dynamic driving tasks. According to the role assignment in the implementation of dynamic driving tasks and whether there are restrictions on the design and operation conditions, the driving automation is divided into 0-5 levels. The industry believes that for the development of automatic driving technology and the large-scale application of landing, the introduction of national standards is the key prerequisite, and a clear grading standard will help to promote the production and landing process of all kinds of automatic driving vehicles. 03. The realistic factors that are hard to cross in the road ahead make the development of automatic driving enter the bottleneck period, which makes the automatic driving industry fall into an awkward dilemma that neither up nor down. From the perspective of the two forces, it seems that the pattern of "the public says that the public has reason and the woman says that the woman has reason" cannot be broken temporarily. At this time, for many forces, it is an era of progress, all the routes are just exploration. Before the real mass production, no one dare to jump to conclusions. In recent years, with the increase of capital, the automatic driving industry has achieved certain development, the whole market is in a new stage, and the mass production landing has become the theme of this year. But at the same time, the industry has already fallen into a strange circle of development. Looking at the overall market development trend, in a long period of time, the R & D strategies of each vehicle enterprise are just like copy and paste. Under the wind, swarming in has become the normal development of the industry. But when the disaster comes, it is the time to test the enterprise's ability of risk prevention, and how to withstand the test of the market is the fundamental. For domestic and foreign car companies, how to balance technology research and development and commercial landing is very important. The auto driving industry needs to return to rationality. To impress consumers with practical value is the foundation of product research and development. The idea of making gimmicks can't last forever. In addition, all automobile enterprises can't jump out of the two propositions of safety and cost on the road of intelligent development. In 2020, the automobile industry itself has been very difficult, coupled with the impact of the epidemic, the whole industry is even worse. A large number of car companies are reducing costs through layoffs and salary cuts, so as to increase investment in intelligence. Although car companies are "rich" relative to start-ups. But the R & D of automatic driving needs a huge investment. If we go into a dead end, all our efforts will be burnt. In terms of security, no matter which route, this topic is indispensable. In the process, enterprises that develop L3 level automatic driving system should be aligned with the whole system architecture and the redundant design of the underlying system, and basically keep up with L4 in hardware and so on. In addition, by limiting the use of scenarios, we should also move downward in terms of functions to ensure safety. At the same time, enterprises that directly skip L3 level R & D and L4 level automatic driving should also consider the feasibility of technology landing, avoid investing a lot of money, and the final result is "a total loss". This is the best of times and the worst of times. Any change is possible and any opportunity will come back.