Vehicle navigation method and system based on track topological map and obstacle avoidance, and medium

Abstract

The invention provides a vehicle navigation method and system based on a track topological map and obstacle avoidance, and a medium. The method comprises the steps of: a map constructing step: constructing an environment map, and generating a track topological map for navigation according to the constructed environment map; a global planning step: searching the global shortest path connecting a starting point and an end point on the generated track topological map by using a path searching algorithm A*; and a local obstacle avoidance step: generating a local target point in real time by using the obtained global shortest path, and generating a vehicle control signal by combining the local target point and the environmental information detected by the laser sensor in real time to control the vehicle to avoid the obstacle and reach the destination. The safe passages in all directions are detected by a ball body, and the found safe passages can be ensured to pass through a small vehicle. The controller combining the local target point of the global path and the safe passages is creatively constructed, and the small vehicle can be controlled to avoid obstacles and safely reach the destination under the condition that the track is not planned in advance.

Description

technical field [0001] The present invention relates to the technical field of robot navigation, in particular, to . In particular, it involves an autonomous navigation car system based on SLAM trajectory topology map and spherical VFH obstacle avoidance. Background technique [0002] At present, the autonomous navigation of the car is one of the basic functions of the robot. One of its purposes is to solve the problem of how to express and record a known environment, how to search for a path in a known environment, and avoid obstacles that are not recorded in the map during navigation. SLAM is currently the most popular method for robot positioning and mapping by researchers. However, because some SLAM-built map point clouds are sparse and have many error points, they often cannot be directly used for robot navigation, or the direct navigation efficiency of grid maps is low. . In terms of obstacle avoidance, the traditional VFH has a small amount of calculation, and the ...

AI Technical Summary

Problems solved by technology

SLAM is currently the most popular method for robot positioning and mapping by researchers. However, because some SLAM-built map point clouds are sparse and have many error points, they often cannot be directly used for robot navigation, or the direct navigation efficiency of grid maps is low.

In terms of obstacle avoidance, the traditional VFH has a small amount of calculation, and the relative potential field method is not easy to fall into the local minimum point, but it is sensitive to the distance threshold. If the threshold is too large, the short-distance channels will be ignored, and the threshold that is too small will make it difficult to find distant channels earlier. ; In addition, VFH is only suitable for single-line laser, it is difficult to extend to the application of 3D point cloud

Images