Galvanic Skin Response and AE-LSTM for Anomaly Detection in VR-Induced Motion Sickness

his study advances the understanding of motion sickness (MS) by integrating subjective measures, like the simulator sickness questionnaire (SSQ), with objective physiological metrics, particularly galvanic skin response (GSR), analysed through an Autoencoder Long Short-Term Memory (AE-LSTM) model. This model, designed for unsupervised anomaly detection, evaluates GSR data to detect differences in physiological (GSR) responses to conditions of different MS potential (here different weather scenarios in a naturalistic VR helicopter simulation). By comparing these physiological anomalies with self-reported cybersickness scores, our findings highlight the importance of combining machine learning-analysed physiological data with subjective reports, offering a comprehensive approach to assessing MS in different conditions. The transition from clear to stormy scenarios revealed marked elevations in MS scores, although the model was currently not able to reliably identify scenario-specific physiological responses that correlate with increased MS.

N/A

2024 IEEE International Systems Conference (SysCon)

N/A

979-8-3503-5880-3

N/A

N/A

7