Commanding small satellites for simulated spacecraft inspections using augmented reality

Abstract

Free-flying teleoperated satellites have been proposed as a replacement for astronaut extravehicular activities (EVA), such as inspection or maintenance, to reduce astronaut risk. A major concern for this type of operation is ensuring the human operator has sufficient spatial and temporal knowledge of the free-flying spacecraft and environment to safely complete the task. This research evaluates Augmented Reality (AR) as a means of providing spatial and temporal information to an operator controlling a free-flying robot to perform an inspection task. Specifically, the research focuses on the effect of command input mode and environmental risk on performance of an EVA inspection task and the strategies adopted by the operator in completing the inspection. Subjects performed a simulated inspection task of a space station using an inspector small satellite and the Microsoft Hololens platform.

Subjects commanded the inspector satellite in three operation modes; satellite body (local) reference frame control, global reference frame control, and global waypointing system (placing markers for the inspector to follow in the global reference frame). Subjects were instructed to inspect the exterior of the station and identify any surface anomalies. Anomalies could occur in areas with low and high risk of contact with the station structure. Subjects performed the inspection task using each of the three fixed command modes (order randomized), with different anomaly configurations. Subjects then performed the inspection task but with free choice of when and how often they utilized the different control modes. Performance was evaluated through primary task measures including percentage of station inspected and accuracy in locating anomaly sites, number of AR interactions, and number of collisions with the station. Workload was also assessed using the NASA TLX survey.

Operation in both the global and local frame controls was found to maximise the percentage of the station that could be inspected, while the waypoint system was found to minimize the number of collisions between the inspector and the station. When operating with free choice of command mode, subjects preferred to stay in a single mode, typically either the global or local controls, likely due to the high usability of these modes for the selected inspection task. Environmental risk area was not found to have a signficant effect on either detection of anomalies or number of collisions. Findings of this paper will inform a follow-on study at NASA's HERA facility, with a full analog-crewed mission operating multiple inspection agents over several days.