Navigation obstacle avoidance method and device, autonomous mobile equipment and storage medium

Abstract

The invention discloses a navigation obstacle avoidance method and device, autonomous mobile equipment and a storage medium, and the method is applied to the autonomous mobile equipment, and comprises the steps: carrying out the enhancement of an initial three-dimensional point cloud through a preset deep learning model, and generating a high-density point cloud; obstacle instance segmentation, semantic classification and obstacle movement trajectory tracking are carried out based on the high-density point cloud, geometric features, category information and predicted positions of all obstacles are obtained, and a cost map is constructed according to the geometric features, the category information and the predicted positions; generating a global path by adopting a global path planning algorithm and a local planner based on the cost map; and moving according to the global path, quickly generating an obstacle avoidance track in the neighborhood of the current position of the autonomous mobile device based on quick replanning of RRT when detecting that a preset emergency condition is met, and switching to move according to the obstacle avoidance track. Therefore, accurate obstacle perception is realized through high-density point cloud, and global path generation and local obstacle avoidance are cooperatively completed in combination with a cost map and RRT emergency re-planning.

Images