Intuitive Gesture Control and Advanced Training for the ISS Robotic Arm
Sherrie Hall, PhD Candidate, MIT Man Vehicle Lab
Abstract:
One of the most essential tools on the International Space Station (ISS) is the Canadarm2 robotic arm, or Space Station Remote Manipulator System (SSRMS), used for all manner of external station operations such as grappling of incoming spacecraft, moving astronauts on Extra Vehicular Activities (EVAs), and station servicing tasks. Unfortunately, its operation is notoriously difficult. The current human-machine control interface uses two independent three-dimensional joysticks, with one controlling translational motion and the other rotational. Astronauts must coordinate separate inputs with each hand to use the joysticks, while visually tracking the robotic arm’s position using three display monitors showing different viewpoints from cameras placed around the ISS. One goal of this study is to develop a more intuitive system for SSRMS operations by comparing different gesture control methods. Each method maps a different set of hand/arm gestures to SSRMS commands. The mappings are compared through preliminary experiments conducted at the Massachusetts Institute of Technology’s Man Vehicle Laboratory, with an SSRMS operations simulator. After training on the simulator with each mapping, participants are asked to perform a series of tasks with the SSRMS, such as approach and capture of a vehicle outside of the ISS. Participants complete these tasks using the different mappings with the BioSleeve interface, and with the traditional joystick interface. These experiments guide determination of which mappings, and thus, which types of gestures, would be most natural and intuitive for SSRMS operators. A second goal of the study is to investigate improvement of SSRMS control through training. Using 3D virtual reality during training, subjects may be able to form a more accurate mental model of the robotic arm in space. It is hypothesized that this will enhance operator performance. Subjects will be tested on performance of SSRMS simulator tasks both with and without virtual reality training, and with both gesture control and traditional joystick interfaces.