簡易装着型の筋電インタフェース実現に向けて: 多数電極からの代表電極の選択

摘要

In this paper, we propose an EMG(electromyography) signal processing method for easy wearing EMG interfaces that can be attached or put on like clothes, supporters, etc. Although we can expect that such wearable EMG interfaces realize convenient multipoint EMG sensors, problems would arise in contacting electrodes, e.g., the positions and contacting conditions of the electrodes can easily change. To cope with this problem, we propose a method for choosing a few representative electrodes out of the may ones attached to a wearable interface. This selection makes recongnition more reliable and reduces the computation time in a training ( parameter adjusting) phase. Our experimental result on fingers dorsiflexion motions shows that the proposed method works properly for a middle finger. The result also suggests that our method is applicable to the other fingers by using more electrodes.%本稿では，筋電の簡易装着型インタフェースの実現に向けて，多数の電極から小数の代表的な電極を自動選択する手法を提案した.筋電信号が筋肉繊維に沿って伝搬し，筋肉繊維と垂直な方向には減衰するという性質を利用し，筋電の振幅と位相遅延の比較によって代表的な電極を算出する.提案手法の有効性を確認するために，9個の電極を前腕に装着し，指の背屈運動を測定したデータで検証した.