Our project is a low-cost prosthetic finger that can be controlled by the wielder completely mentally. To do this EEG probes are placed on top of the subject’s head near the pre-frontal and primary motor cortex. This allows us to detect the signals produced by the brain when someone decides to close their hand and when the signals from the brain act to initiate the action itself. One of the challenges we faced is that EEG data is very noisy but through filtering we were able to remove some disturbances. We used data from a person gripping and relaxing their hands to train a machine learning model to recognise what certain patterns of brain waves mean. We use a raspberry pi to straighten an bend a 3D printed finger. We also connect the pi to our EEG headset and run our ML model on it. This system allows for free movement by the user as the finger is controlled only by the raspberry pi which is light enough to carry around in one's pocket and can be battery powered. Similarly the EEG is also battery powered and connects wirelessly to the raspberry pi. EEG and computer components have the potential to be shrunk down for more convenience. The affordability of 3D printing and use of lightweight hardware makes this prosthetic incredibly accessible.
Attribution and Thanks to Nicholas Brookins (dangercreations.com) whose "Knick's Prosthetic Finger v3.5.5" design we used and modified.
Log in or sign up for Devpost to join the conversation.