Hand motion estimation by EMG signals using linear multiple regression models.

The purpose of this research is to construct an intelligent upper limb prosthesis control system that uses electromyogram (EMG) signals. The signal processing of EMG signals is performed using a linear multiple regression model that can learn parameters in a short time. Using this model, joint angles are predicted, and the motion pattern discrimination is conducted. Discriminated motions were grip, open, and chuck of a hand. Predicted joint angles were multi-finger angles corresponding to these three motions. In several experiments we proved the usefulness of processing EMG signals with a linear multiple regression model.