Research On Cybersickness Detection In Virtual Reality Environment

Cybersickness is a kind of stimulation disease with the symptoms of nausea,headache,dizziness,fatigue and visual fatigue.It’s urgent to detect and alleviate cybersickness symptoms.Although some models have been invented,the real-time performance and model application haven’t been considered.This study launched virtual roaming experiment,collected and analyzed experiment data,constructed cybersickness feature set,and established a real-time model to detect the cybersickness severity.The research content includes four parts:1.A feature index extraction method suitable for real-time detection.According to the real-time requirement,the experiment data is divided into time segments,and a set of characteristic indexes are extracted from each time segment.The model application experiment verifies the feasibility of this method.2.Cybersickness classification dataset.We carried data analysis on the original experiment data,constructed a list of important features,and established two datasets SeverityData and IssevereData.3.Cybersickness real-time detection model.We trained a fourcategory classification model of cybersickness severity level with the F1 of 93.89%and a binary classification model of severe cybersickness detection with the F1 of 98.94%.4.Model application.Add the binary classification model of severe cybersickness detection into the VR experiment scene,and carried a model verification experiment to verify the model accuracy.