Which kind of enterprises are expected to take the lead in the first year of self-driving production?

In2019, the topic of commercialization of autonomous driving continues to ferment. Several guests at the second Global Autonomous Driving Forum hosted by GE Automotive recently said that starting this year, autonomous driving will gradually enter the final sprint phase before mass production, with some auto companies launching their first self-driving cars as soon as the end of this year or early next year. Guangzhou Automobile, for example, is said to be ready for production in the first quarter of 2020 on its L3-class self-driving mass-produced car, and a mid-to low-end version of the car will be released in October this year. In addition, such as FAW Red Flag, Chang'an, Chery, BYD, Singularity, and so on, have also disclosed on different occasions next year's plans to mass-produce L3 self-driving cars. This leads more people to believe that 2020 will be the first year of mass production of self-driving cars, especially autonomous brand of self-driving products, hopefully from next year to start real popularization. L3 Autopilot Steps into Mass Production Sprint Stage When Will Autopilot Be Really Mass Production? The answer to this question, which has been debated in the industry for many years, is slowly becoming clear-soon. How fast is it? From the end of 2019 to the beginning of 2020, this period will be a key point for the true popularity of autonomous driving in China, according to Guo Jishun, head of the Intelligent Driving Technology Department at the Guangzhou Automobile Research Institute. Guo Jishun is right. According to incomplete statistics, in addition to multinational car companies and technology companies, at least a dozen autonomous car companies have announced that they will mass-produce L3-class self-driving cars around 2020. Bearing the brunt is Guangzhou Automobile, which is expected to do so as early as October this year Launch a mid-to low-end version of the L3 self-driving mass-production model. Once this mass production plan is implemented, it is understood that Guangzhou Automobile will become the industry's first real introduction of self-driving vehicles. Chang'an Automobile, as one of the autonomous automobile enterprises that have been involved in the research and development of self-driving technology, is also rapidly promoting the engineering development of L3 self-driving technology. It is understood that in the past few years, Chang'an Automobile has mastered more than 30 key core technologies and realized eight core functions in seven aspects, namely, in-depth systems engineering, environmental awareness, central decision-making, functional safety, human-computer interaction, test evaluation, and control and execution. Based on the existing achievements, Chang'an Automobile has entered the development stage of mass production, and is expected to officially realize the mass production of L3 technology by 2020. Chery, FAW Red Flag, BYD, Great Wall and others, with the help of Baidu's cooperation, have also set up Flag, said that next year will mass production of L3-class self-driving cars. Chery, it is understood that in 2018, some models have been equipped with the L2-level auxiliary driving system. This year, Chery plans to gradually install more advanced auxiliary driving functions in real vehicles, such as ACC, AEB, APA, etc. Finally, the L3-level conditional autonomous driving will be realized in 2020, and the L4-level highly autonomous driving will be realized in 2025. FAW Red Banner, on the other hand, is a little bit "aggressive", aiming to implement a small-batch off-line demonstration operation of L4-class autonomous passenger vehicles by 2020, and launch more urban operations in large batches by 2020 to achieve large-scale mass production. Others such as Singularity, Xiaopeng, Weilai, Zero Run New forces in car-building, such as self-driving mass-production plans, are also aiming for 2020. Singularity Vehicle, it is reported that the use of Nvidia autopilot platform to provide computing power and security, will be officially completed in 2019 L3 technology research and development, L3 autopilot volume production landing in 2020. Xiaopeng Motor is working with Desai Siwei to push L3-related technology to mass production in 2019. Zero-run cars are slightly different, and the company's previously published self-driving roadmap will use the OTA approach to boost production of L3 self-driving technology. However, despite the current L3 self-driving market situation is a good, Guo Jishun that there are several problems can not be ignored. The first is the regulatory level, when will the technology be fully recognised and accepted? A typical example is the full new generation of Audi A8, which is already capable of mass production in 2017, but is currently capable of releasing L3 autopilot functionality on only a few roads in Germany, due to legal and regulatory constraints, rather than fully utilizing all of its intelligent capabilities. Secondly, there are three main technical problems: First, the price of AI chips is too high, and most of them can not meet the demand of self-driving in performance, which is far less than the industry's previous expectations; Secondly, the redundancy of the hardware and sensors of the current self-driving platform can not fully meet the safety requirements of self-driving; Thirdly, the existing self-driving system can not meet the overall functional safety requirements. From this point of view, autonomous driving is actually in the midst of a great deal of activity. A mixed period of euphoria and anxiety is a key node in the industrialization of autonomous driving, which determines whether the industry can truly achieve leapfrog development in which vehicles are responsible for traffic results. " Guo Jishun pointed out. In addition to the current state of L3 autopilot technology, Guo Jishun also highlighted the commercialization of higher-level autopilot technology at the forum. Guo Jishun pointed out that the development of L4 technology still faces many problems, especially in the chip, redundancy and executive components, there is still room for further breakthroughs. So the L4 business model may start in a small area and be commercialized quickly through, for example, point-to-point self-driving ride-hailing services within the area. Then gradually expand the operation area, so that the various regions connected to each other, and eventually form a whole. "By the end of this year, a lot of companies will be rolling out self-driving cars in China, and what we know now is that at least 10 companies have such plans." Another L4 commercialization path is self-driving in the park, which is aimed at commercial areas, including warehouse-to-warehouse driverlessness, and is also seen by Guo as a direction for future development. "In fact, from 2015, self-driving logistics began to heat up, analysis of the main reasons behind the following points: First, self-driving cars in the park on the road almost no laws and regulations; Second, the scene is relatively simple; Third, the self-driving functions in the campus are relatively simple and can better meet the safety requirements, "said Gu Jianmin, CTO of Valeo China Dot is similar to it. He points out that there are three main types of self-driving business models: the first is a privately owned self-driving car, which will be optional in the future, but also because the car is private to the consumer, so the cost needs to be borne by the user, and the sensitivity is high. And because many private cars are generally used for a short time, the cost and time of using self-driving vehicles in this case will be disproportionate. The second is self-driving vehicles owned by carriers, such as travel service operators, shared cars, and driverless taxis, which are theoretically self-driving 24 hours a day because they are operated by carriers and whose operating costs become lower as they are used longer. In addition, autonomous driving can save drivers, so this mode of cost increase is relatively easy to accept, less sensitive, and is expected to be accepted by more users. The third is a driverless van or dining truck, which means that the vehicle is not carrying people, but goods, or take-out of our order. This requires the vehicle to be L5, that is, driverless in any scenario, because there is no driver. But because of this, this mode of vehicle passive safety requirements are relatively less stringent, after all, if there is a loss of goods can be made up with a certain amount of money, unlike personal safety. The latter two of these technical routes may be relatively easy to land on and are expected to land sooner. Tang Rui, chief executive of Longitudinal Technology, expressed his views on the commercialization of autonomous driving from another angle. Tang Rui believes that with the gradual opening of the geographical fence and the gradual spread of the scene, it may take more than 10 years for self-driving to achieve the leap from a closed high-speed environment to full scene coverage. That's why Zongmu chose to start with the special feature of autonomous parking in low-speed restricted scenarios. In Tang Rui 's view, autonomous parking will be a very clear, in the short term it is possible to large-scale commercial landing of L4 self-driving applications, but also the final mass production of self-driving technology must go through the road. "Why do you say that?"? Because if L4 is to be truly commercialized, I personally feel that three conditions need to be met: cost control, quality, and a sound supply chain system. Because autonomous parking is mainly a low-speed scenario, there will not be too many legal barriers to commercialization, coupled with the low-speed scenario driving, braking distance is very short, does not need to see very far, in terms of sensor configuration, there will be less of a must-have option of lidar, cost will be greatly reduced. In addition, we can see that parking is in fact strong just need, if we can achieve unmanned, will undoubtedly solve a lot of city car pain points. So overall, Tang feels that autonomous parking will be the first scene where the L4 will land in China or even around the world, and it could be commercialized on a large scale in two or three years' time. Moreover, Tang Rui emphasized that the first commercialization through autonomous parking will actually help the industry explore a better match between the social system, laws and regulations and self-driving technology, so as to open up more L4 farms in urban areas for later research and development Scene provides some experience for reference, so that the technology can eventually be landed in a broader range. "That's why we see autonomous parking being the focus of almost all current talk about autonomous driving, because it's a very important scenario and the market is expected to grow nearly 10 times in the next few years." And if you look at it in the long run, Guo Jishun believes that man-machine co-driving will become the mainstream practice in the industry for a relatively long period of time, Because mass-producible autopilots must rely on technology and complementary interactions to move forward, they will not be able to fully meet the industry's needs in a short period of time--either for L3 or high-level L4, or for full autopilot.