A visual servoing docking approach for marsupial robotic system

This paper presents a visual servoing docking approach for marsupial robotic system. A vertical V-shaped visual benchmark is designed to guide the docking motion of a child robot with the feedback of image features. By changing the angular of the camera with a rotational DOF, the child robot adjusts its heading orientation for docking. The task is modeled into six states: Blind, Adistance, Atangent, Around, Aiming, and Parking states. The transform conditions among the states are also given. To verify the approach, a simulation platform is designed, including an image-infor virtual machine, a decision-making unit and a pose refresher. The simulation results verify the presented docking approach.