A Dynamic Obstacle Avoidance Method for Mobile Robots Based on Stochastic Reachable Sets
Chang-An Yao, Sung-Hua Chen · 2023
This paper utilizes the Artificial Potential Field (APF) method as the obstacle avoidance strategy for robots, and integrate it with Stochastic Reachable (SR) sets to calculate the collision probability distribution between the robot and obstacles based on their relative distances. The collision estimation with SR Sets can effectively improve the success rate of avoiding dynamic obstacles while reaching the target point. The APF method generates a virtual repulsive force field around obstacles and an attractive force field around the target point. The robot is then subjected to a combined force of repulsion from obstacles and attraction towards the target point, allowing it to avoid obstacles and reach the target point. The collision probability distribution between the robot and obstacles can be estimated by using the Stochastic Reachable Sets based on their relative distances. The estimated probability helps to modify the repulsive force generated by APF. The newly proposed repulsive force with various integration methods for are tested in the simulation results, which reveal significant improvements in collision rate and time consumption.