With the development of artificial intelligence and in-depth learning technology, GPU-driven computing is being applied in areas such as large data, medical, translation, intelligent city, autopilot, education and so on. Especially Sensor in the field of automatic driving, NVIDIA and many car manufacturers and automatic driving technology research and development companies have established contact.
The company is open to the driver-side platform NVIDIA DRIVE. Specifically, the platform includes four levels, namely, automatic driving vehicle equipped with software operating system ASIL-D OS, depth learning and learning platform Drive PX, computer visual SDK (Driveworks SDK) and automatic driving computing platform Drive AV. It is important to note that Drive AV is a specialized application of NVIDIA for autopilot development. The application is designed to enable data fusion of sensors such as cameras, millimeter-wave radars and lasers, and to provide redundancy for L4 and L5 autopilot through Fuel Rail Pressure Sensor a variety of depth learning and machine vision algorithms. NVIDIA also launched a supporting development tool for developers. With the Driver PX chip, autopilot developers can deploy their own software systems on these computing platforms, run the depth of learning the network, and then automate the car to achieve all kinds of vehicle sensor data processing, and then let the vehicle to achieve environmental awareness , Path planning and positioning functions, and ultimately to achieve automatic driving.
NVIDIA's latest 'independent machine processor' Xavier's latest developments. This is known as the 'complex system-on-chip' processor can be used for simulation, training and automatic driving. According to the plan, Xavier samples will be in the first quarter of 2018 to the British Weida partners open, the product will be the fourth quarter of the year on the market. As a provider of artificial intelligence computing platform, NVIDIA has attached great importance to the application of AI in automatic driving scene. In his view, on the basis Temperature Sensor of carrying a supercomputer, the future car will be defined by the software, rather than by a 'black box' system. The future car will no longer carry dozens of CPU, but has a unified computing platform. Although it has to go through a lot of obstacles, but he is optimistic that in ten years, car manufacturers will be able to help with the British, in ten years in the $ 30,000 level mid-range car to achieve automatic driving.
Huang Renxun also on the billion car emphasizes the openness of the British Weida principle. NVIDIA is a platform provider, not a vertical integration company. At present, they have with the public, Audi, Toyota, Daimler and Tesla, including various types Speed Sensor of car manufacturers reached a cooperation in the future will be in Jingdong and other Chinese enterprises in the field of logistics and unmanned business development.