Document Type
Presentation
Date
Summer 2022
Academic Department
Engineering and Physics
Faculty Advisor(s)
Dr. Mark Brinton
Abstract
Electrocutaneous stimulation stimulates nerves in the skin and allows a user to feel what they are touching while using a prothesis. Without a sense of touch amputees must focus their attention on their prosthesis and without sensory feedback the user cannot adjust their grip to match the object. A Convolutional neural network (CNN) was implemented alongside of the Kalman filter to see which algorithm will reject artifact better. An experiment was designed in a virtual environment to test which of the two algorithms will reject stimulation artifact better. The experiment consists of three blocks that all take a different amount of force to move, and the electrical stimulation occurs when a force is applied to the blocks. We will compare finger position during stimulation and ability to identify blocks. We anticipate the CNN to perform better and ignore stimulation if the artifact is included in the training.
Recommended Citation
Griffin, Nathan and Brinton, Mark R., "Minimizing Stimulation Artifact in Real-Time Movement Decoding" (2022). Summer Scholarship, Creative Arts and Research Projects (SCARP). 45.
https://jayscholar.etown.edu/scarp/45
Notes
Scholarship, Creative Arts, and Research Project (SCARP)