Development of a Calibration Routine for Remotely Controlling a Robotic Hand

Abstract

Maintenance and de-orbiting of satellites are tasks that are yet to be solved. One of DLR’s approach to this task is to on-orbit servicing is using a telepresence system. High performance telepresence systems enable intuitive interaction with the remote environment. There are several factors influencing the overall performance like the level of immersion of the video display. This work focuses on improving the coupling between the hands, DLR/HIT Hand II, of the remotely located humanoid robot Justin and the hands of the human operator. In order to ensure improved coupling of the two, both the ability of the DLR/HIT Hand II to track given references, and the calibration and mapping of a dataglove, which measures the human hand joint angles, are improved. This is achieved by three main tasks: implementing a friction compensation module with a friction parameter identification routine, designing a controller based on the functionally related systems framework, and developing a calibration routine that handles the calibration and mapping of a dataglove simultaneously. Experiments proved the improvement of the reference tracking ability of the DLR/HIT Hand II and the performance of the calibration and mapping of Cyberglove is verified by subjective opinions of people who have previously operated the telepresence system at DLR without the new calibration.