A 3D-printable Prosthetic Hand Based on a Dual-arm Operation Assistance Model

This paper proposes a 3D-printable myoelectric prosthetic hand to enable stable prosthetic control in a series of dual-arm operations. The proposed prosthesis involves a state-transition model of dual-arm tasks in daily livings. Each task is discriminated using a reccurent probabilistic neural network from posture information of both arms and can be accurately controlled by EMG and the task discrimination result based on posture information. In the experiment, prosthetic hand control in daily living was performed and it could be demonstrated that the participant voluntary control the prosthetic hand using the task model in dual-arm operations.