精神状態推定のためのボリンジャーバンド法による筋電信号処理方法

カテゴリ: 全国大会

論文No: 3-114

グループ名: 【全国大会】平成30年電気学会全国大会論文集

発行日: 2018/03/05

タイトル(英語): The signal processing method of electromyography using Bollinger band for mental condition prediction

著者名: 黄 晨暉(NEC),盧 驚文(NEC),梶谷 浩司(NEC)

著者名(英語): Chenhui Huang(NEC),Jingwen Lu(NEC),Hiroshi Kajitani(NEC)

キーワード: 精神状態,筋電,信号処理,サイバーネティックス

要約(日本語): Mental health care and mental state assessment have been increasing paid attention recently. Nowadays mental state estimation by biosignals of automatic nerves has come out. Electromyography (EMG) were used and integrated with traditional biosignals to assess and predict mental states effectively by wearable sensors. However, EMG is a kind of instant signal that has to be detect at a high sampling rate, therefore a feature precision decreasing will be induced by low sampling rate due to some reason for example saving energy for wearable device. In this report, we propose a new data processing method for EMG that precision of feature will be maintained although sampling rate is low. In order to evaluate the effect of this method, first, we use Stroop Test and some relax video to stimulus the mental state and then calculate the EMG feature by the new method, and finally we use machine learning method to evaluate the precision of mental status prediction and compare the results of prediction which use traditional method. For the conclusion, the results using new method is better than that the effect of the traditional method is confirmed.

原稿種別: 日本語

PDFファイルサイズ: 313 Kバイト