Abstract

Virtual assistants (VAs) are known to improve performance, reduce mental workload (MWL), and improve situational awareness (SA) in complex cognitive tasks. However, there is a lack of studies focusing on anomaly resolution tasks in time- and safety-critical environments such as long-duration exploration missions. This paper aims to investigate the effects of using a VA, namely Daphne-AT, on crew performance, SA, and MWL in the context of spacecraft anomaly treatment. Participants ($n=12$) were tasked to detect, diagnose, and resolve anomalies in two different but equivalent experimental sessions (five anomalies with the VA and five anomalies without the VA). In each condition, performance was quantified based on the number of anomalies correctly resolved and the total time needed to resolve the anomalies. SA and MWL were quantified using the Situational Awareness Rating Technique and the NASA Task Load index. Results indicate that participants resolved more anomalies in less time during sessions with VA. MWL reduced significantly during sessions with VA. Although total SA did not significantly change between VA conditions, the subscales attentional demand and understanding improved with VA. These findings suggest that VAs are promising decision support tools for anomaly resolution in domains such as spaceflight, aviation, or emergency response.