Of course, the three protagonists we are going to talk about today are three semiconductor companies that exist like giants in the traditional consumer electronics field, namely Qualcomm, NVIDIA and Intel.
You may be familiar with Qualcomm, because your Android phone is loaded with Qualcomm's Snapdragon chip, or the recently popular smart cockpit chip "8155"; Familiar with NVIDIA, you may be a fan of console games, and excellent graphics cards such as GeForce RTX 3080Ti can let you gallop in the virtual world; Intel is no stranger to you. Everyone knows about the computer CPU i3/i5/i7, but today we will mention Mobileye, a subsidiary of Intel, which has been deeply involved in the field of intelligent driving chips for a long time.
How do these three giants enter the automotive field today? What kind of excellent products are launched? We continue to look down.
0 1 Qualcomm wedding, focusing on Snapdragon digital chassis.
At this year's CES, Qualcomm CEO An Meng showed us the layout and achievements of Qualcomm in the automotive field. The most striking thing is a partner map covering the whole PPT, including as many as 37 global leading automobile brands, such as Ford, BMW, Tesla and Honda, as well as domestic brands such as Great Wall, Weilai, Geely and BYD. It really is a wedding. Since so many OEMs have joined Qualcomm's circle of friends, Qualcomm naturally has advanced technology and comprehensive product lines for manufacturers with different needs to choose from.
This time, Qualcomm brought their overall smart car solution-Snapdragon Digital Chassis.
Snapdragon Digital Chassis consists of an open and extensible cloud connection platform, which uses a unified architecture to bring higher security and immersive digital experience, and supports the function upgrade of the next generation car in its whole life cycle. Automobile manufacturers can choose to use any or all platforms covered by Snapdragon digital chassis in their product lines, and provide highly customized experience for their products through continuous upgrading in the cloud. Simply put, this digital chassis includes four functional platforms: automatic driving, intelligent cockpit, intelligent network connection and cloud service.
1, Snapdragon bus platform
Snapdragon Ride is an open programmable platform, specially designed for Qualcomm's autopilot, which can meet L2+/L3 autopilot requirements. Qualcomm recently announced a series of cooperation progress, including helping GM build Cadillac LYRIQ and helping BMW build its autonomous driving platform. At the same time, Qualcomm announced at the exhibition to expand its technology portfolio to meet the changing needs in the field of autonomous driving.
Qualcomm first introduced the Snapdragon driving vision system, which is an open and extensible autopilot platform. The system is based on a 4-nanometer system-on-chip (SoC), aiming at optimizing the deployment of forward-looking and panoramic cameras, and supporting driver assistance system (ADAS) and automatic driving (AD). Snapdragon amusement equipment vision system integrates the dedicated high-performance Snapdragon amusement equipment SoC and the next-generation visual perception software stack of Arriver, and provides a variety of computing functions to enhance the perception of the surrounding environment of the vehicle, support the planning and execution of the vehicle, and help achieve a safer driving experience. The system is expected to be used for mass production of automobiles in 2024.
2. Snapdragon Digital Cockpit Platform
At this CES show, Qualcomm announced the cooperation with Honda and Volvo, which are chips based on the cockpit field.
At present, the digital cockpit platform in Snapdragon has developed to the fourth generation. The platform can help automobile manufacturers seize the opportunity to change the in-car experience, provide brand-new services, and create multi-monitor, multi-camera, advanced audio, video and multimedia experiences through highly customizable and always-connected SoC and virtualization software solutions, as well as a mixed-key environment that can safely meet the needs of consumers and security ecosystems at the same time.
The fourth generation platform adopts the 6th generation Qualcomm Kryo CPU, Qualcomm Hexagon processor, multi-core Qualcomm AI engine, 6th generation Qualcomm Adreno GPU and Qualcomm Spectra ISP. The new platform enhances the functions of graphics, multimedia, computer vision and AI, and aims to support the cockpit system with situational awareness optimization and adaptive ability, which can evolve according to the preferences of drivers and passengers.
3. Snapdragon Auto Zhilian Platform
The platform can help automobile manufacturers to build powerful LTE and 5G networking services, cellular car networking (C-V2X), Wi-Fi, Bluetooth and precise positioning capabilities, fully support the safe connection between automobiles and the cloud, other vehicles and the surrounding environment, and meet the growing demand for safer and more immersive driving experience.
4. Snapdragon Team Cloud Service
Through the pre-integrated software and service platform designed for the new profit model, it provides flexible function combination, performance upgrade and brand-new functions for automobile manufacturers. As mentioned earlier, automobile manufacturers can choose any or all platforms covered by Snapdragon digital chassis in their product lines, and provide highly customized experiences for their products through continuous upgrading of the cloud. In other words, automobile manufacturers can choose one of the module functions or package them all according to their own needs. Of course, you can upgrade in the cloud later. Such high flexibility has attracted the favor of a large number of manufacturers.
In a word, Snapdragon Digital Chassis is a brand-new platform to meet the needs of consumers and the automobile industry, which provides a brand-new idea for the evolution of automobile technology in the future.
NVIDIA plowed the autopilot platform, and Hyperion 8 appeared again.
If Qualcomm is the leader in the field of intelligent cockpit chips, then NVIDIA has a higher position in the field of autonomous driving chips.
At this CES show, NVIDIA announced that more companies will adopt its open DRIVE Hyperion platform, which includes high-performance computer and sensor architecture to meet the safety requirements of fully self-driving cars.
The latest generation of DRIVE Hyperion 8 consists of a redundant NVIDIA DRIVE Orin system on chip, 12 surround cameras, 9 radars, 12 ultrasonic modules, 1 front laser radars and 3 internal induction cameras.
The system has strong security redundancy. Even if the computer or sensor fails, the backup device can ensure that the self-driving car can safely send passengers to the destination.
At present, the system has cooperated with many manufacturers, such as Volvo's high-end brands Polaris, Weilai, Li, R cars, Zhiji cars, etc., all of which adopt DRIVE Hyperion. In addition, driverless taxi services such as Cruise, Zoox and Didi, and truck services such as Volvo, Navistar and Plus are also using DRIVE Hyperion.
It is worth mentioning that two companies in China have also announced that they will carry NVIDIA's Orin chips during CES. One of them is that Du Ji's first production model will be equipped with NVIDIA DRIVE Orin chips. Du Ji's mass production model is expected to be launched in 2023, when it will become an automobile robot with L4-class autonomous driving capability. At the same time, the concept car of this product is expected to be officially unveiled at the Beijing Auto Show in April this year.
The other is Tucson, a self-driving truck company, which announced at CES that it will build a new platform based on NVIDIA DRIVE Orin chips. The single operation capability of Orin chip can reach 254TOPS, and the process is 7nm. At present, the performance is still outstanding.
Intel's Mobileye blew the horn of counterattack with three chips in succession.
Previously, there was a saying in the industry that Mobileye was no longer the dominant position, because in recent years, many car companies stopped cooperating with it and turned to companies like Qualcomm, NVIDIA and Huawei. Since 20 16, Tesla has taken the lead in giving up Mobileye and embarking on the road of chip self-research. The familiar FSD chip is the result of Tesla; Later, Li ONE changed from Mobileye EyeQ4 chip to domestic Horizon Journey 3 chip, which dealt a greater blow to Mobileye. Even BMW, a long-time partner, will invest in Xiaolong riding platform in Qualcomm. A series of blows made the industry look down on Mobileye. However, at this CES show, Mobileye seems to have blown the horn of counterattack, and successively launched three new chips, namely EyeQ Ultra and EyeQ? 6L and EyeQ? Six hours.
EyeQ Ultra is a single-package self-driving automobile integrated chip supercomputer platform specially designed for end-to-end autonomous driving. The platform is based on the validated Mobileye EyeQ architecture, and achieves the excellent performance of 176 TOPS by optimizing the computing power and efficiency. EyeQ tablet equivalent to 10? 5, which means EyeQ? A leap of series system integrated chips. With the help of 5 nm process technology, EyeQ? Ultra can meet all the needs and application scenarios of L4 autonomous driving, while avoiding the extra energy consumption and cost caused by the combination of multiple system integrated chips. This chip is expected to be supplied by the end of 2023 and fully mass-produced in 2025.
EyeQ？ 6L and EyeQ? 6H is the next generation EyeQ system integrated chip of ADAS.
First of all, EyeQ? 6L is EyeQ? 4, its packaging size is only EyeQ? Fifty-five quarters. This means that higher computing power can be achieved with lower energy consumption. This chip focuses on L2-level assisted driving system. Samples have been provided since last year, and mass production is expected in mid-2023.
Another EyeQ 6H, through the configuration of panoramic camera, can realize high-end ADAS and some automatic driving functions. Even in terms of strength, it is equivalent to two EyeQ? 5。 But more importantly, EyeQ? 6H supports visualization, L2+ ADAS function, multi-camera processing ability including parking camera, and third-party applications such as visual parking and driver monitoring. This eye? The most advanced ADAS system integrated chip in this series will begin to provide samples this year and is expected to be mass-produced by the end of 2024.
After talking about these new products, we might as well review the evolutionary history of Mobileye's chips:
-In September 2003, Mobileye released EyeQ 1 chip.
The chip supports front collision warning (FCW), lane departure warning (LDW) and traffic sign recognition.
-In 2008, Mobileye released its second chip: EyeQ 2.
The chip adds ACC and pedestrian detection emergency braking functions. & lt/