Augmented Reality Navigation Interfaces Improve Human Performance In End-Effector Controlled Telerobotics

On the International Space Station (ISS) and space shuttles, the National Aeronautics and Space Administration (NASA) has used robotic manipulators extensively to perform payload handling and maintenance tasks. Teleoperating robots require expert skills and optimal performance is crucial to mission completion and crew safety. Degradation in performance is observed when manual control is mediated through remote camera views, resulting in poor end-effector navigation quality and extended task completion times. This thesis explores the application of three-dimensional augmented reality (AR) interfaces specifically designed to improve human performance during end-effector controlled teleoperations. A modular telerobotic test bed was developed for this purpose and several experiments were conducted. In the first experiment, the effect of camera placement on end-effector manipulation performance was evaluated. Results show that increasing misalignment between the displayed end-effector and hand-controller axes (display-control misalignments) increases the time required to process a movement input. Simple AR movement cues were found to mitigate the adverse effects of camera-based teleoperation and made performance invariant to misalignment. Applying these movement cues to payload transport tasks correspondingly demonstrated improvements in free-space navigation quality over conventional end-effector control using multiple cameras. Collision-free teleoperations are also a critical requirement in space. To help the operators guide robots safely, a novel method was evaluated. Navigation plans computed by a planning agent are presented to the operator sequentially through an AR interface. The plans in combination with the interface allow the operator to guide the end-effector through collision-free regions in the remote environment safely. Experimental results show significant benefits in control performance including reduced path deviation and travel distance. Overall, the results show that AR interfaces can improve performance during manual control of remote robots and have tremendous potential in current and future teleoperated space robotic systems; as well as in contemporary military and surgical applications