Skip to content

Latest commit

 

History

History
17 lines (14 loc) · 923 Bytes

readme.md

File metadata and controls

17 lines (14 loc) · 923 Bytes
Raw

EMG-CNN

Depreciated

  1. Make sure to download Stockwell Transform for MATLAB (st.m)
  2. Import your time-series EMG signal to the MATLAB environment with each fatigue level its own vector/matrix (NOT table)
  3. Run dsp.m, it will output folders of each fatigue level containing individual windows
  4. Ensure 'high_dir' and 'low_dir' in model.py point to these folders
  5. Run model.py to split data, train the model, and evaluate the model

Setup

  1. Clone repository
  2. Place raw EMG signal .csv files in according subdirectories in /datasets/
  3. Run model.py to split data, train the model, and evaluate the model

Notes

  • I've gotten the model to be pretty accurate even with a ridiculously small window (currently 40 samples, equivalent to 1/100th of a second of data)
  • After exponential drops around 10 epochs the model seems to linearly improve until asymptoting (and likely overfitting) around 200 epochs