著者

小林 直樹 濱田 雅美

出版者

公益社団法人 日本生体医工学会

雑誌

生体医工学 (ISSN:1347443X)

巻号頁・発行日

vol.51, no.1, pp.17-23, 2013-02-10 (Released:2013-09-10)

参考文献数

13

---

Visual induced motion sickness (VIMS) due to real videos (passive videos) has been well evaluated by using a set of RR intervals (RRI) on electrocardiograms (ECG). However, in case of the estimation of VIMS for interactive videos suchas video games, it is difficult to obtain a stable VIMS index by using only RRI because of several modulation factors by respiration and personal errors among individual operations. For the evaluation of VIMS on interactive videos we propose VIMS quantification indices that use principal component analysis (PCA) of two time-variant biosignals, RRI and respiration signal (RESP), respectively. As VIMS indices we introduced frequency components of a score (rr) and a score (resp), which are time-variant scores resulting from PCA of the biosignals. The experimental results were estimated by regression analysis between the proposed indices and simulator sickness questionnaire (SSQ) or global motion vectors (GMVs) of videos. The experimental results show that a positive correlation is shown between the index of the score (rr) frequency analysis and SSQ for both passive and interactive videos. Because there are no significant difference between the indices for interactive video and for passive videos, the index does not influence operation factors, and reflects biomedical influence caused by video features such as GMVs. This means that the index calculated by score (rr) is effective for quantifying the VIMS caused by video factors, even though the dispersion of the index caused by the subject's individuality is not sufficiently small. This index also influenced the individuality of the subject's game habituation. On the other hand, there is little correlation between the index of the score (resp) frequency analysis and SSQ for each type of video. As a dependency of the operation and personal errors might be included in this index, we need to study more detail about this component.