Rea Gill¹, William Harwin¹, Harun Tugal², Peter Scarfe¹; ¹University of Reading, ²UK Atomic Energy Authority (UKAEA), Remote Applications in Challenging Environments (RACE), Culham Campus, Abingdon, UK

Remote operation, or teleoperation, systems are typically used in unsafe or inaccessible environments, such as in nuclear decommissioning. Here, human operators must overcome visuomotor transformations, which result in a mismatch between their movement and the visual feedback they receive. Despite this being a typical feature of teleoperation systems, we do not know which aspects of these transforms (e.g. altered translation, rotation, or scale) are most detrimental to task performance and user experience. Here, in a bimanual virtual reality task, we have examined the impact of translation, rotation, and scale mismatches between a user’s movement and the visual feedback they receive. Participants completed a “pick up and place” task where they picked up a cube with their left hand, passed it to their right hand, then placed it in a box. To equate the transforms, we use a two-part study in which we first determine the ‘just noticeable difference’ (JND) for each transform using a 2AFC paradigm. Next, using the JNDs (1, 2, and 3) to equate each transform in terms of sensitivity, we examined which (if any) is more detrimental to task performance and user experience. Overall, as expected, we find that increasing transformation magnitude systematically decreases task performance. We present a range of well-established task performance metrics (Bradshaw, Elliott, Watt, Hibbard, Davies, and Simpson, 2003) such as task completion time, smoothness of movement, and time in slow phases of the movement, as well as self-reported measures (workload, effort, ownership, agency, presence). Our results highlight the importance of considering human factors when developing teleoperation systems. Notably, we apply rigorous psychophysical methods to a more complex multi-stage movement than is typical in the field, but better represents real-world use cases.

Acknowledgements: This research was generously supported by the Southeast Network for Social Sciences Doctoral Training Partnership (SeNSS DTP) and EUROfusion.