Docking to a Specific Person of an Autonomous Mobile Manipulator Using UWB Positioning and RGB-D Camera

In this paper, a novel navigation method is proposed to dock a mobile manipulator in front of a specific person by combining the ultra-wideband (UWB) positioning system with image-based motion control. The UWB is used for a longer distance human identification and positioning. As the robot approaches the target person, face recognition and face pose estimation are used for close docking control the person. The experimental results show that the average error of docking to a person is 5.2° in orientation and 8.48cm in distance. The integrated experimental results demonstrate that the autonomous mobile manipulator can navigate to a target person and handover an object to the person.