Human-Guided Autonomous Mobile Manipulator in Dynamic Environments: A Case Study Using LoCoBot

In the context of Industry 4.0, human-robot interaction emerges as one of the major topics to relieve human operators from heavy and repetitive tasks. The objective of this research is to develop a real-world human-robot interaction to help humans with tool transportation. In this study case, a mobile robot equipped with a robotic arm is tasked to autonomously follow a human operator through a predefined space while avoiding dynamic obstacles and carrying a payload. Once the human operator has reached the destination, the robot has to transfer the item to him/her safely. The platform is controlled using the ROS 2 middleware; all the subtasks required to perform the entire operation have been implemented through the adaptation of preexisting packages into a unique framework. A depth camera is incorporated to detect dynamic obstacles in the path of the mobile robot, while external cameras are used to obtain the poses of both the robot and the individual during the entire operation. These are determined using fiducial marker-based computer vision. To control the behavior of the robot, a state machine is implemented and the changes between states are performed using hand gesture recognition to enhance interactivity. The proposed framework is evaluated through simulation tools, followed by implementation in a real-world scenario.