NVIDIA released the company's latest developments in the self-driving field at CES 2018, and Danny Shapiro, senior director of the company's automotive division, noted that NVIDIA's goal is to provide the computing power and sensor support required for Level 4 or higher self-driving levels in the short term. It is expected that in the next 2-3 years, self-driving technology will replace human driving in most cases. However, it is difficult to integrate all kinds of complex computing tasks of self-driving into a single computing platform, because future applications will only become more and more complicated.
There is little chance for self-driving to develop a single computing platform
According to The Street, Shapiro said that looking at the history of the computing industry, applications will only become more and more complex, which is also true in the automotive field. Therefore, even if the built-in functions of the chip platform will gradually increase, most of them will still have multiple The demand for chip platform products.
NVIDIA currently develops proprietary computing platforms for both self-driving technology and Drive IX. It is worth noting that Shapiro pointed out that the NVIDIA Drive AR platform is not completely developed for self-driving cars, and the same is true for Drive IX.
Given NVIDIA's partnership with a number of automakers, the use of information from a car manufacturer to strengthen other solutions that the company provides to the industry has also become a focus of attention. Shapiro said that NVIDIA has always had a confidentiality agreement. In the case of building its own software stack, neural network, various basic tools and databases, it will not allow the information of a partner car manufacturer and another car manufacturer. shared.
Shapiro also mentioned that if a car manufacturer relies on sensors such as LiDAR, radar and cameras to collect data for its self-driving platform, another car manufacturer relies on a different configuration of the sensor solution. To collect information on the self-driving platform, the information of each other may not be able to communicate effectively.
Shapiro believes that Level 4 self-driving will likely drive on closed roads, airports, amusement parks or towns with specific test areas in the next 2-3 years.
Software definition helps progressive self-driving technology development
In addition, Tesla has said that the second-generation Autopilot system, which was launched at the end of 2016 and adopts NVIDIA technology, will be strong enough to support the full auto-driving technology once the software level keeps up. Shapiro believes that software-defined cars can acquire new applications ( App) and update, even if it is not a full-automobile technology car at the time of purchase, it can be used to obtain more self-driving functions through software updates.
Shapiro said that the development of NVIDIA and automakers in the development of self-driving technology software is similar to the cooperation between game developers and game developers. For example, NVIDIA does not develop its own video games, but has written all tools, all databases, etc. To allow game developers to develop their applications on this basis.
The same situation can be seen in the automotive industry. For example, NVIDIA will supply automotive OEM software stacks, developer tool execution environments and simulators. However, it is up to the car manufacturer to decide which car, specification, etc. NVIDIA solution will be imported. The same is true for the fine-tuning of the user experience.
NVIDIA is working with a number of third-party map service providers such as HERE, TomTom and Baidu, and is hoping to enhance this map service with NVIDIA's artificial intelligence (AI) advantages. . Because the map service is very localized, it is a labor-intensive business model in this case, so if it is more automated, it will be able to complete more tasks in the data center, so more AI will be involved.
Unlike Waymo and Mobileye, NVIDIA does not develop its own map service. The strategy is to help other companies to use the company's solutions for map development, including in-car hardware and software, and deep learning optimization. Server graphics chip (GPU).
Fiber Optic Distribution Box,Fiber Optic Breakout Box,Fibre Optic Breakout Box,Fibre Break Out Box
Cixi Dani Plastic Products Co.,Ltd , https://www.dani-fiber-optic.com