Low Latency Myoelectric Controlled Interface for Finger Click Detection

Myoelectric controlled interface (MCI) has been extensively investigated as an effective tool for the control of advanced prosthetics, exoskeleton, entertainment gadgets, etc. In these human-in-the-loop applications, the time delay through the loop: from the generation of human intentions, to action of the machines in response to human intention, to the feedback the machines provide to the human operator, is important and some times critical for a successful human-machine interaction. However, in most current MCI studies, typically hundreds of milliseconds data were required for making a prediction, introducing a significant long delay in the control process, and limiting the application of MCI in those time sensitive scenarios. This project investigated a low latency MCI for predicting the movement of finger click by employing only several milliseconds data. We propose to exploit spatial information, including increasing the number of electrodes, optimize the location of electrodes, and extracting spatial features, to compensate for the information reduction in time dimension. The investigation of low latency MCI would improve the efficiency of the human-machine interaction and expand the application of MCI in the real world.

Faculty Supervisor:

Ning Jiang;John Tze Wei Yeow

Student:

Jiayuan He

Partner:

Brink Bionics Inc.

Discipline:

Engineering

Sector:

Manufacturing

University:

University of Waterloo