- Circuits and PCB for 4 channel EMG data acquisition
- Bluno beetle and Android app interfacing for wireless real-time transfer of data
- Machine learning and implemention of Random Forest Classifier for 9 exercises
-
Take EMG signals from 4 channels and pass them through the PCB to remove noise and amplify the signal
-
Transfer the data over Bluetooth to the Android app which receives and updates multiline graphs in real time.
-
Save the data on a local file or online database
-
Put the data through a random forest classifier for exercise classification of 9 different exercises
- Low quality circuit boards are near impossible to work with
- Natural inconsistencies in the hardware leading to hours of debugging
- Lack of guides for using a multiplexer
- Training the random forest classifier
-
Attach the “nervous system” to the “skin”
-
Pass data through seglearn
- Option 1: get seglearn working on a phone
- Option 2: get the data off the phone and run seglearn on a computer
-
Improve machine learning and database handling
-
Test interfacing of the app and the Bluno Beetle instead of the app and Arduino for communication