Core components of unmanned vehicles: computer vision, sensor fusion, positioning, path planning, and control.
The specific realization, the fusion of computer vision and sensors, can be called [perception].
[Positioning] Contains technologies such as high-precision maps and LiDAR.
[Path planning] includes two parts, predicting other vehicles and planning yourself.
Finally, there is the [Control] part.
High-precision map: The map used by the robot contains a lot of driving assistance information, the layout of intersections, the location of road signs, speed limit sections and traffic lights.
Centimeter-level accuracy.
1. Use high-precision map to help positioning, lidar collects landmark features, and compares with known landmarks on the high-precision map (preprocessing-coordinate conversion-data fusion).
2. Use high-precision maps to help perception. High-precision maps can reduce the perception circle to a region of interest (ROI).
3. Use a high-precision map to help plan and identify the exact center line of the lane. The high-precision map can be used to judge in advance where there are crosswalks, speed bumps and low speed restrictions.