Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : September 09, 2018 - September 12, 2018
This paper describes the joint angle estimation method focused on the characteristics of surface electromyogram (sEMG) for increasing the degrees-of-freedom (DoF) of the myoelectric hand. A pulse density of the motor neurons obtained from sEMG increases with increasing a joint angle of the finger, however, the characteristics are also changed by joint angle statuses such as moving or stopping and moving speed. Even an online experiment, we could judge weather fingers were moving or stopping from the pulse densities of flexor and extensor muscles. Therefore, we proposed a finger angle estimation method which employs different temporal filters, which were changed the cut-off frequencies setting to 0.07 Hz, 0.15 Hz, and 0.12 Hz, for each movement status such as bending, rest and others. This method improved the responsiveness during the bending 67% and the smoothness during the stopping 21% than conventional methods. In addition, we moved the robot hand by an online estimated finger joint angle.
