Data collection and evaluation of muscle activity and hand kinematics during daily life activities

There are already a large number of data sets that contain EMG and hand kinematics data to enable decoding for the control of hand prostheses via neural networks. Some of these datasets have been summarized under the Ninapro database (Atzori & Muller, 2015), (Ninapro, 2024). While the datasets include a large number of people and different movement patterns, they do not reflect realistic everyday situations due to the controlled laboratory conditions and are therefore only of limited significance. In our study, however, the test subjects should be able to move freely. When performing the tasks, body movements away from the hand should also take place in order to recognize possible interference and thus make the control systems more robust for everyday use. Another innovation relates to the sensors used. In addition to the EMG sensors, force sensors are to be used to record muscle activity. Our hope is to use the additional force myography to counteract the strong noise behavior of the EMG sensors and thus make the control of hand prostheses more robust.

More information can be found here.

Kategorie: Master Thesis