A collision avoidance path planning method and system for unmanned vehicles based on iteratively improved apf

Abstract

The present invention provides a method and system for unmanned vehicle collision avoidance path planning based on iteratively improved APF. The method includes: S1: establishing a two-dimensional workspace, marking the starting point, end point and obstacles; S2: discretizing multiple Cell, obtain the potential value corresponding to the cell; S3: sort the potential value, set the threshold, and mark the cell greater than the threshold; S4: traverse the cell, obtain two marker clusters, lower the threshold, and repeat S3 until Two marker clusters have an intersection; S5: Record the threshold and the cell corresponding to the threshold that make the two marker clusters have an intersection, find a cell that meets the requirement that only one cell connects the two marker clusters, and set it as the first midpoint; S6: Repeat S4-S5 to obtain midpoints that meet the requirements, and all midpoints are connected with the starting point and the ending point to form an obstacle-free path. The method can plan a path avoiding obstacles in a short time, and the path distance is extremely short.

Description

technical field [0001] The invention belongs to the technical field of unmanned driving, and in particular relates to a collision avoidance path planning method and system for unmanned vehicles based on iteratively improved APF. Background technique [0002] In car travel, traffic accidents are often caused by fatigue driving, drunk driving, answering and calling, distraction, and inattention. Huge pressure leads to reduced travel efficiency. [0003] In the field of unmanned driving research, traditional obstacle avoidance methods include the combination of radar and sensors. Such equipment is extremely vulnerable to damage, and the maintenance cost in the later period is also extremely high. Later, some scholars combined artificial intelligence algorithms to propose the APF (artificial potential field) obstacle avoidance method. The basic idea of ​​this method is to regard the movement of the vehicle in the surrounding environment as the movement of the vehicle in the ar...

AI Technical Summary

Problems solved by technology

[0002] In car travel, traffic accidents are often caused by fatigue driving, drunk driving, answering and calling, distraction, and inattention. Huge pressure, resulting in reduced travel efficiency

[0003] In the field of unmanned driving research, traditional obstacle avoidance methods include the combination of radar and sensors. Such equipment is easily damaged, and the maintenance cost in the later period is also extremely high.

Later, some scholars combined artificial intelligence algorithms to propose the APF (artificial potential field) obstacle avoidance method. The basic idea of ​​this method is to regard the movement of the vehicle in the surrounding environment as the movement of the vehicle in the artificially established virtual field, and the target point generates gravity. Guide the vehicle to move to the target point; obstacles generate repulsive force to avoid collision between the vehicle and obstacles, repulsive force and attractive force construct a virtual potential field, and the vehicle moves along the "potential peak" between the "potential peaks" of this potential field under the combined force of the two. The path planned by this method is generally smooth and safe, the method is concise and the real-time performance is good; however, when this method is used, it may fall into a local optimal solution, which makes it difficult for the computer to plan the path. Vehicle dynamics constraints, eventually falling into a calculation trap; and there is no channel between dense obstacles, and when obstacles exist, they will be affected by oscillations, and there are disadvantages such as relatively fixed scope of potential field, inflexibility, and excessive obstacle avoidance angle

Images