The GB400 industrial gateway is equipped with NVIDIA Jetson AGX Xavier core module, containing 512 CUDA cores, Volta architecture GPUand NVDLA engine. It boasts 32 Tops real-time reasoning capability with 30W module power consumption so as to be compatible with all current neural network models. The whole machine features fanless and strong embedded design. It has strong adaptability to harsh industrial environments thanks to the wide temperature working environment and high seismic protection level. 

The GB400 industrial gateway integrates a variety of expansion interfaces, and supports 4G/5G/Wi-Fi communication. The standard PCIex8reserved design fully interprets its high scalability that allows the access of a variety of special sensors. It is widely used for smart city vehicleroad collaboration, assisted driving, unmanned delivery vehicles, intelligent inspections and other scenarios.

Its application in vehicle-road synergy and assisted driving enables massive terminal access management to improve operation and maintenance efficiency, reduces the difficulty of integration with the open and compatible platform, and facilitates the processing of computing tasks. For the auxiliary computing platform in high-speed autonomous vehicles, customers can carry out secondary development. By processing image data, combined with the capabilities of cameras and radars, it can complete the perception and detection of road events such as lane line recognition, traffic light recognition, pedestrian detection, and driver fatigue detection and analysis.