Development of myoelectric prosthetic hands applying caging grasps for long-term use

我们研究了长期使用肌电假手的笼式抓取：(1)位置控制假手的笼式抓取的实验研究，以及环形笼式的视觉物体识别的实验研究；(2)肌电运动估计。在(1)中，我们提出了笼式抓取作为机器人对物体的几何约束，并对几个简单形状的物体进行了实验逼近手部的姿势和手指位置。此外，我们还开发了用于环形笼子的目标特征检测和姿态估计。在(2)中，我们可以用2个电极识别前臂和手指的4个动作。

We studied caging grasps by a myoelectric prosthetic hand for long-term use of the hand: (1) experimental study of caging grasps for position-controlled prosthetic hands, and visual object recognition for ring-type caging, (2) motion estimation with electromyography. In (1), we proposed caging grasps, as geometrical constraint of objects by robots, and experimentally approached the posture and fingers location of the hand for caging several simple-shaped objects. Also we developed feature detection and posture estimation of objects for ring-type caging. In (2), we can recognize 4 motions of the forearm and the fingers by 2 electrodes.