A preliminary research into joint angle prediction of the upper limb using surface electromyogram for a cooperative machine

This paper presents the preliminary results of a joint angle prediction method for the upper arm and forearm motions using surface electromyogram (sEMG) signals. An artificial neural network (ANN) was used to match the relationship between sEMG and upper limb motion in the vertical plane. The sEMG signals from the four sites were fed into the ANN and captured motion data were used as references. The prediction method was tested on one subject through drawing experiments. The performance was evaluated by a root mean squared error (RMSE) and the result was comparable to previous studies (RMSE < 0.02 rad). These results imply that the method is useful for the natural interaction between a human and a cooperative machine.