Development of Autonomic Docking Algorithm for Autonomous Mobile Robots by Integrated with Behavior Tree

Abstract

This paper is to introduce the development of autonomic docking for autonomous mobile robots (AMR) using behavior trees (BT) and gazebo simulation software in a ROS2 environment. The virtual robot in this study is capable of moving around and locating the charging station by using the camera to identify the ARTag on the map generated through simultaneous localization and mapping (SLAM) data. BT is used to develop an autonomous docking algorithm and compared three different scenarios through experiments with the AMR. In the experiments, the parameters were adjusted when the AMR was positioned 2 m away from the charging station along the x-axis, with no discrepancy along the y-axis. Additionally, the parameters were tweaked when there was a y-axis discrepancy of ± 0.25 m in cases 2 and 3. The results indicated that modifying the heading tolerance can reduce docking errors and improve the efficiency of the algorithm. Moreover, using BT to design algorithms can enhance functionality and simplify the system's operation.