Artificial Intelligence for Myoelectrically Controlled Cooperative Arm Prostheses

Description

Individuals with an upper limp amputation often use a prosthesis to conduct activities of daily living and participate in society. However, the upper limb is challenging to replace, as the arm and hand have to conduct very complex gross and fine motor movements, with many degrees of freedom. In a survey with 307 prosthesis users in Korea, the most described difficulties were tying shoelaces, using a bottle opener and using scissors. In these so-called cooperative hand movements, the action of one hand is supported by an appropriate counteraction of the other hand through neural coupling. For a prosthesis to work collaboratively with the still existing hand, it would need to mimic this neural coupling and respond to the actions of the remaining hand. In our research, we aim to improve the human control of existing arm prostheses during cooperative hand movements, by introducing online artificial intelligence (AI) into the myoelectrical control mechanism.In order to plan and initiate a larger investigation with our whole network, we will collect some pilot data with the funds from this NTN – Innovation Booster “Brain Date”. These data will help us to better understand the difficulties that users of myeoelectric arm prostheses face in everyday life. We are also creating an initial dataset of electromyogram (EMG) data for a proof of concept study.