Our reporter Ji Wei reports from Beijing
Guide
"If you divide the scenes of autonomous driving, you may divide into thousands of scenes, and each scene should be fully tested, but the participation of many companies will promote the entire industry to develop faster."
Auto-driving seems to be getting closer.
On June 27th, Volvo announced that it has cooperated with Nvidia in the field of autonomous driving, and the two camps of autonomous driving around the world have grown stronger. At present, two camps with two major chip companies, Intel and Nvidia, have been formed. The former include BMW, Volvo, Delphi and mobileye, including Volvo, Bosch, ZF and Tesla.
Of course, the big event in the domestic autopilot field is also in the pipeline. On July 5, Baidu will release the details of the Apollo program, which claims to open an autopilot platform and try to build a fast innovation ecosystem and promote autonomous driving technology. Since the popularization plan was announced in April, it has caused great sensation and concern.
Then, in the face of the surging of self-driving at home and abroad, what is behind the rapid alliance between the two camps?
What kind of role will a technology company like Baidu play in the industry chain? And what role will traditional companies like BMW, Audi and Volvo play? What is the difference between the Internet car manufacturer Tesla in the development of autonomous driving? With 2020 as the time node, to what extent will autonomous driving develop?
With a series of questions, on June 28th, the 21st Century Business Herald exclusively interviewed Yu Kai, founder and CEO of Horizon Robotics.
"At present, there is no standardized product and service for autonomous driving, so the mode of forming the camp now is based on improving the efficiency of promotion." From the vice president of Baidu Research Institute to the horizon of the technology company that created the algorithm chip, Yu Kai as the automatic driving field Experts have their own judgments on the role of technology companies in the autopilot industry chain.
"Bao Tuan" research and development makes the development and promotion of autonomous driving more efficient
"21st Century Business Herald" (hereinafter referred to as "21st Century"): On a global scale, autonomous driving has formed two camps, Intel, BMW and moblieye camp and Bosch, NVIDIA, Mercedes-Benz camp, what is the reason behind this? What will happen to the future?
Yu Kai: I divide the enterprises in the whole industry chain into two categories. One is a relatively traditional enterprise, such as a OEM, Tier One (a Tier 1 supplier), and the other is a new force with the background of an Internet company. Car manufacturers, they are more concerned about the user experience, future travel operations and services. Of course, these two categories are also converging, such as Mercedes-Benz launched car2go.
At the same time, the development of autonomous driving technology has also appeared similar to the above camp, such as BMW's cooperation with Intel and Delphi, Mercedes-Benz cooperation with NVIDIA and Bosch.
"21st Century": On June 27th, Volvo also announced cooperation with Nvidia to develop artificial intelligence platforms in cooperation with other companies, and to develop automatic driving systems in the next 10 years. How do you comment on this?
Yu Kai: At present, the autonomous driving industry is still in the early stage. There is no general technical standard and the division of labor is not clear. The cooperation between enterprises can be more bundled to do the research and development of "husband", so that the future development of autonomous driving is faster and more efficient.
"21st Century": How to understand the research and development of "Bao Tuan"?
Yu Kai: The so-called "station side", specifically the cooperation between enterprises. The main reason for forming a group is that if the company has mature technology and mature division of labor, then the complete technical solution can be sold to many other companies; but at present, autonomous driving is still in the early stage of the industry, and no enterprise can independently provide complete and mature technology. Because of the uncertainty of the industry, the general technical standards have not been formulated, and the resources are still limited. Therefore, upstream and downstream enterprises are required to jointly promote R&D and landing, and they can improve their efficiency and reduce risks while exerting their respective advantages. Cooperation between upstream and downstream companies will be even closer.
"21st Century": In the field of autonomous driving, two camps with Intel and NVIDIA as the core have been formed. How do you evaluate the strength of these two companies?
Yu Kai: I think these two companies have their own characteristics. Intel has strong comprehensive strength and has a leading position in the field of computing platforms. At the same time, it has 5G core patent technology, and through the acquisition of Mobileye, it quickly acquires powerful artificial intelligence application modules. And human resources, and Nvidia has accumulated a wealth of software talents and has a first-mover advantage in the field of artificial intelligence computing. Both companies are highly competitive in the field of autonomous driving, and the competition between them is worthy of continued attention in the future.
Technology companies will play a key role
"21st Century": The details of the Baidu Apollo program will be released soon. Do you think that technology companies like Baidu will play a leading role in the autopilot industry chain?
Yu Kai: How to locate the role of Baidu, personally feel that it is not clearly defined, but it will have a greater impact. At present, Baidu is recognized as the forefront of artificial intelligence technology in China. In addition, Baidu also has high-precision maps, which is also an essential part of autonomous driving, and Baidu’s cloud and big data capabilities are for traditional OEMs. The Tier One supplier is a great addition. Therefore, I think that technology companies may play a key role in the entire industry chain.
"21st Century": For Baidu's Apollo program, the outside world understands that the plan will define the roles that companies in each chain should play, and will open up some artificial intelligence technologies. The purpose is not to let other companies not Do repeat research and development?
Yu Kai: When foreign automakers promote autopilot in China, they will definitely choose service providers with high-precision maps and operators with service capabilities. As far as I know, Baidu's Apollo program will open up some code so that companies can work together on the code and the entire industry will evolve faster.
If the autopilot scene is divided, it may be divided into thousands of scenes, and each scene should be fully tested, but if the company participates together, the whole industry will develop faster. Before Baidu made the Apollo plan, it must have been planned, and then as an open ecology, the purpose is how to let everyone go together to evolve.
"21st Century": What details do you pay attention to the Baidu Apollo program?
Yu Kai: I still look forward to Baidu's AI Developer Conference on July 5th. The most concerned is the extent to which Baidu is open. For us, the more open it is, the better. Personally think that Baidu may be gradually open, and Baidu must think about the border.
"21st Century": What kind of role does the company like the Horizon play in the new industrial ecology that Baidu Apollo plans to build?
Yu Kai: Horizon is the only company in the domestic auto-driving field that focuses on algorithm chips, and the algorithm chips themselves need to be integrated and integrated with other industry chains.
For the Apollo program, my understanding is that it focuses more on the cloud, while the horizon-embedded embedded artificial intelligence solution pays more attention to local processing power, and there are opportunities for cooperation. For example, Baidu may open up a lot of data in the Apollo program. The horizon can use this data to train the model. The trained model will also accumulate its own data on the processor on the horizon. In this way, we have data on the horizon ourselves. Other companies participating in the Apollo program also have their data. Everyone puts their own data into this platform, and this platform will also feed back a lot of data.
"21st Century": Will the Horizon directly cooperate with OEMs, or will it only work with Tier One suppliers?
Yu Kai: Horizon works with Tier One suppliers as well as with OEMs. Our positioning is relatively clear, the core is the provider of artificial intelligence capabilities, is to sell solutions, can be provided to suppliers, but also to the OEM.
Automated driving requires centuries of accumulation and internet thinking
"21st Century": Many multinational companies have invested heavily in autonomous driving, such as Tesla, Volvo, and Audi. What do you think Tesla has in terms of autonomous driving research and development?
Yu Kai: Tesla is a real car thinking company. When there is no autopilot software in the car, the hardware has been laid out in advance. Every time a car is sold, it will continuously accumulate data and let Tesla’s owners and Enterprises share data; unlike Tesla, the traditional OEM will complete the entire solution ahead of time, complete the software and hardware, and then release the product, before the OEM is afraid to do online software upgrades.
"21st Century": Do you think that in the future, in the field of autonomous driving, will Tesla lead other traditional enterprises by 2020?
Yu Kai: At present, from the perspective of data accumulation, Tesla is more likely to surpass other car companies. After all, it has tens of thousands of cars running on the road and constantly accumulating data, which is different from the traditional car business model. The former has entered the mode of operating services, and the latter is still the model for selling products.
"21st Century": In the field of autonomous driving, multinational companies must meet Chinese standards in order to enter China. So what do you think you need to pay attention to regarding the issue of automatic driving standards?
Yu Kai: The standard setting for autonomous driving is not available overnight, because many things cannot be foreseen. At present, the country should actively promote the development of facilities, sites and standards, so that the car companies can do a good job in a clear framework.
"21st Century": At present, the industry talks about autopilot with 2020 as a time node. Which stage do you think autopilot will reach by 2020?
Yu Kai: The main engine has a fixed cycle for launching new cars. It usually takes three years from the time the project is completed, and then it will take three years to go public. Therefore, the car that was launched in 2020 will start research and development this year. Now it is necessary to discuss the production model after four years. What it is, this is determined by the inherent rhythm of the entire industry. At present, Tesla and other enterprises have already realized the function of Level2 assisted driving. The main application scenario of Level2 is automatic driving on the expressway. The automatic driving on the highway is relatively controllable and the environment is relatively simple. According to the classification of automatic driving, Level2 is Level3, and most of the OEMs are planning to achieve Level 3 autopilot by 2020. (Editor Zhou Kaiping)
Energy 8000,Energy Vape Pen,Energy 8000 Disposable,Energy 8000 Puff Disposable Vape
Shenzhen Zpal Technology Co.,Ltd , https://www.zpal-vape.com