Methods for Estimating the Effect of Angle of Incidence on Manual Space Station Inspection Performance

Abstract

Complex structures in space, such as the International Space Station (ISS), require routine and emergency external inspection to locate damage. Inspections performed through windows, stationary external cameras, or cameras mounted on a robotic arm provide limited coverage of station surfaces. Astronauts can perform inspection during extravehicular activity (EVA), but EVA is costly and risky. A free-flying robotic platform can provide unconstrained views of external station surfaces without risk to astronaut safety, but existing inspection planners do not consider human perception and situation awareness when maximizing station surface coverage for damage detection. This work aims to understand the effect of camera angle on the human supervisor’s ability to detect damage in images of a space station gathered by a robotic inspector. We conducted two pilot studies where participants identified damage on a space station through digitally rendered inspection images. We fit psychometric curves to the anomaly identification rate conditioned on camera angle and identified the threshold camera angles for each anomaly type. From these studies, we recommend methods for estimating parameters for future inspection planners to utilize based on the acceptable error rate for the inspection.