Last June, 165438+ 10, Weilai announced that it would cooperate with Mobileye in an advanced R&D project. As part of this project, Weilai will be responsible for producing the L4 class self-driving "kit" developed by Mobileye and installing it on its electric car, which can also be sold to self-driving taxis and individuals. At the same time, Mobileye will also develop a system for Weilai, which can automatically drive the electric car to the nearest charging station when the battery of the electric car is about to run out without any operation. It can be seen that Weilai has great trust and dependence on Mobileye in the future autonomous driving technology.
How is Mobileye known to the public?
Everyone in the industry knows that Mobileye is the dominant player in ADAS, with a global market share of 70-80%, but it is not always smooth sailing. 1999 has not received an order since its establishment eight years ago. Until the products developed were officially commercialized in 2007, in just ten years, Mobileye cooperated with 27 automobile manufacturers, including GM, BMW, Volkswagen, Ford, Nissan and Honda, and loaded the Mobileye system on/kloc-0.5 million cars worldwide, accumulating driving scenes across 43 countries.
On August 65, 1965, 438+0, in 2004, Mobileye's company listed on the New York Stock Exchange raised as much as 890 million dollars, setting a record for Israel's IPO in the US stock market. Moreover, the company's market turnover at that time had reached $5.5 billion. It can be said that through ADAS advanced assisted driving system, Mobileye gradually built its own moat and developed better and better.
But making Mobileye truly known to the public still benefits from the cooperation with Tesla Autopilot system. Success is also Xiao He's failure, Xiao He. Although Mobileye's EyeQ3 was applied to Tesla's production models in 20 14, the fatal case of Tesla's automatic driving in May 20 16 attracted great attention from all over the world, which also made Mobileye seriously questioned. Mobileye and Tesla have different opinions about this serious accident.
Mobileye thinks their products have limited functions. At that time, Autopilot did not have the ability to detect the horizontal entry of vehicles. It is Tesla's radical permission opening and publicity that leads to the owner's use exceeding the functional limit. Tesla stressed that the main cause of the accident was the recognition error of the camera and radar. In fact, it can be seen from the response to this accident that the understanding of the future development of autonomous driving between the two sides has been quite different.
So shortly after the accident, Mobileye issued a statement announcing the termination of cooperation with Tesla. Elon musk then responded that the suspension of this cooperation will not affect the progress of Tesla's autonomous driving system.
Recognize yourself and the market structure.
The most unexpected thing is that in 20 17, after the cooperation between Mobileye and Tesla ended, it was acquired by Intel at a high price of153 billion US dollars. As for why he chose to be acquired by Intel, Professor Amnon Shashua of Mobileye answered this question:
When we talk about autonomous driving, we are not talking about a product, but the whole industry. We are not talking about computer vision, artificial intelligence and chip systems, but about infrastructure, cloud computing, automobile manufacturers, regulatory authorities, 5G networks and data centers. How can a company with only 650 employees do this alone? So Mobileye needs to join an industry giant and start a new chapter.
Actually, what we lack is not money, but resources, including human resources and infrastructure resources. We need to negotiate with the relevant regulatory authorities, but on the scale of Mobileye, we can only negotiate with the regulatory authorities in Israel, not with the regulatory authorities in other countries. But on the scale of Intel, it can negotiate with regulators around the world.
Intel has a department dedicated to autonomous driving, and they reached an understanding with us in concept, so we chose to join. At the same time, Intel Israel Branch has nearly 6.5438+0 million employees, and its layout in Israel is very huge. This is why Intel attracts us.
In the past few years after being acquired, Shashua's thoughts have not changed. He believes that the self-driving car industry needs to be integrated. If each company only focuses on developing its own products (such as cameras, lidar, high-precision maps, etc.). ), this strategy is unsustainable for an inseparable end-to-end system.
In the development of autonomous driving technology, it will be better if companies can cooperate with each other instead of competing with each other to develop overlapping technologies. Because the future self-driving cars will also rely on artificial intelligence, machine learning and sensor fusion for navigation, all these systems need to be integrated to make the vehicles run automatically. However, the reality is that there are very few companies that can participate in the whole process from chip to autopilot system development, and it is even more difficult to achieve comprehensive cooperation.
Complete architecture
In order to further improve Mobileye's business map, early last month, Intel acquired Moovit, an Israeli urban travel planning startup, for $900 million. Muwitt started with itinerary planning. At present, it provides services including online car rental, public transportation inquiry, bicycle rental and two-wheeled electric car rental in 102 countries and regions around the world, with 800 million users and 50 million monthly active users. After the acquisition, Moovit's business will still operate independently, but all operational data will be sent to Mobileye. This huge user base and travel data will be of great help to the research and development of Mobileye autonomous driving technology.
Of course, as Sha Shuya has repeatedly stressed, Mobileye's all-round travel operation capability is based on L4-level automatic driving capability, so Mobileye has been ready to go on L4.
According to foreign media reports, a few days ago, Mobileye released a 40-minute video, showing how a self-driving car of the company was driving at a crowded and chaotic intersection in Jerusalem and on the streets of various delivery vehicles. You may have seen many demonstration videos of self-driving cars, which combine the overhead picture of the drone with the camera in the cockpit, but Mobileye is unique in that during the entire 40-minute demonstration, the automatic driving perception system only used 65,438+02 cameras, without any radar, lidar or other sensors.
Do you feel deja vu when you see this? Isn't the automatic driving technology based on camera instead of lidar what Elon Musk has been advocating? But unlike Tesla, Mobileye feels that camera-based autonomous driving is necessary, but at the same time it is not enough, because although they are safer than human drivers, the difference in proficiency is still not enough. Therefore, Mobileye adopts two independent redundant subsystems, one using camera vision and the other using sonar/lidar. Just like developing an IOS system and an Android system at the same time.
Meanwhile, EyeQ6, which Mobileye is developing, is scheduled to be launched in 2023. It is reported that 1 EyeQ6 chip is equivalent to 6 EyeQ5 chips, which can realize all Robotaxi operations. In addition, Shashua also revealed at this year's CES show that by 2022, the total cost of all hardware, including sensors, cameras, and several lidar and radars, will be about $654.38+$5,000. Mobileye believes that by 2025, the system cost can be reduced to less than $5,000.
Judging from the current situation, Mobileye has made some achievements in vision system, chip research and development, travel service and overall system cost control, and the whole system architecture has been set up. Whether it can succeed in the increasingly fierce field of autonomous driving in the future remains to be seen.
Figure | From the network
This article comes from car home, the author of the car manufacturer, and does not represent car home's position.