An Approach for Dynamic Obstacle Avoidance in Autonomous Mobile Robots Operating in Unstructured Indoor Environments

Abstract

This paper deals with mobile robot navigation and obstacle avoidance, presenting a probabilistic search algorithm, Rapidly exploring Random Tree (RRT) method, to ensure stability and flexibility of robots in the face of unexpected events. To optimize the map updating process, the improved RRT automatically builds the map by combining global and local paths. The path is optimized using the Dijkstra algorithm to increase the real-time performance. As the robot moves, the map is continuously updated to detect dynamic obstacles. When an obstacle is detected, a new optimal path is generated to guide the robot to the goal. Experiments on the robot operating system have shown that the optimization works and the robot can automatically avoid static and dynamic obstacles in the simulated environment, map quickly, and avoid small spaces. The results of the study can be extended to movements in unstructured environments and can be used at boundary nodes.