自动驾驶车辆的控制方法、装置、设备、介质及程序产品
By obtaining different preset reference points based on obstacle type, determining the nearest reference point of the target obstacle and performing ST projection, the problem of inaccurate obstacle projection on curves or lanes in autonomous vehicles is solved, improving the accuracy and safety of vehicle obstacle avoidance.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- XIAN UNISOC TECH CO LTD
- Filing Date
- 2022-11-01
- Publication Date
- 2026-07-17
AI Technical Summary
In autonomous vehicles, existing technologies are inaccurate in determining the nearest reference point when the vehicle is in a curve or when there are many lane reference points, resulting in inaccurate ST projection and affecting the accuracy of the vehicle in avoiding obstacles, especially when the obstacle is small.
By determining the type of the target obstacle, preset reference points of different types are obtained. For example, for VRU type obstacles, reference points of virtual tracks divided on the cognitive lane are used to determine the nearest reference point and perform ST projection. For non-VRU type obstacles, reference points on the lane reference centerline are used to improve the accuracy of target projection.
It improves the accuracy of autonomous vehicles in avoiding obstacles on curves or lanes, enhancing the safety and reliability of autonomous driving.
Smart Images
Figure CN115535001B_ABST