Vision-Guided Human-Robot Interaction for Intuitive Control of Anthropomorphic Aerial Manipulators

Achieving more natural and user-friendly interaction with the novel humanoid dual-arm aerial manipulator to realize hybrid intelligent teleoperation control remains an urgent problem. This paper proposes a vision-based intuitive control strategy for our designed novel aerial bimanual manipulator, focusing on enhancing human-robot interaction (HRI) and vision-based teleoperation. This innovative HRI control strategy approximates complicated nonlinear mapping functions between the human operator's motion intent and the manipulator's positioning. Moreover, robust full-body 3D skeleton tracking from multiple Kinect DK units is fused using the KF algorithm to serve the aerial manipulator's teleoperation control. Employing this control system, the anthropomorphic aerial manipulator can cooperate with a human to execute bimanual tasks flexibly and efficiently. Experiments show the adequacy and accuracy of the proposed vision-based skeleton tracking method for aerial manipulator teleoperation control in facilitating collaborative human-robot interaction.