The full-vision autopilot solution uses hierarchical timing network technology, 8 cameras share the backbone network, and 64 terminal tasks are designed under the backbone network, including detection, segmentation, stitching, tracking, depth estimation, 3D detection, scene memory, and 3D reconstruction , Prediction, driving behavior analysis, dangerous scene diagnosis.
At present, the solution has been implemented through distributed deployment of multiple titan series?controllers In the future, the design will be deployed on the Areas chip to complete the domestic delivery of All in One.
● Remote hiring, automatic driving, and formation driving within the geo-fenced area (parks, scenic spots, etc.)
● 5G video streaming and remote driving
● V2X Vehicle-Road Cooperative Driving
● "End-pipe-cloud-platform" mobile travel service platform product line