Virtually There

Nuclear decommissioning is a major environmental issue facing the UK. Decommissioning sites often have incomplete plans and decommissioning begins with the spatial characterisation of hazardous buildings. One approach to spatial mapping is to use semi-autonomous teleoperated multi-robot teams, controlled remotely using VR.

Control systems for robot teleoperation present a number of challenges and as the complexity of a system increases (e.g., semi-autonomous robot teams) trust in autonomy has been shown to decrease. The key to trust and effective operation is in the way in which large quantities of data, captured by a robot-team, is relayed to an operator. The over-reliance on visualisation has stimulated research into novel methods for filtering and organising visual information. However, previous research has shown that multisensory integration of vision with sound can improve an operator’s awareness during high visual load tasks.

Sound has only begun to be explored in VR robot teleoperation where basic auditory feedback has been shown to improve user performance.

In this project, we propose a new approach to characterising hazardous spaces using VR by splitting information across multiple sensory modalities and delegating data to the auditory channel using sonification – “the use of non-speech audio to convey information”  which has been shown to improve user trust and workload by making system critical information easier to process.

An additional impact of adding sound to VR is that it tends to increase immersion and a sense of presence which has been shown to impact trust in VR. In hazardous environments such as nuclear sites, an increased sense of presence can result in undue stress for VR operators, and this has ethical implications for their employers.