somewacko / hmiees

A wireless, real-time EMG device that learns about your movements.

Human-Machine Interfacing via Epidermal Electronic Systems

A VCU Engineering Capstone Design project to measure and classify EMG signals to control an external system in realtime. Awarded Excellence in Design at the 2015 Capstone Expo for having one of the most exemplary projects in the ECE department.

Our poster describing the project in more detail can be found here.

Using the Matlab classifier:

    reference_signals = 2*rand(200, 10)-1;
    test_signal       = 2*rand(200,  1)-1;
    
    classifier = EMGClassifier(200);
    
    % Specifying features is optional
    classifier.features = [
        EMGFeature.MAV,  ...
        EMGFeature.WAMP, ...
        EMGFeature.ZC,   ...
        EMGFeature.WL    ...
    ];
    
    example_gesture = classifier.register_gesture('The Example Gesture');
    classifier.train(reference_signals, example_gesture);
    [gesture, distance] = classifier.classify(test_signal);
    
    fprintf('Classifier recognized %s with a distance of %0.2f\n', ...
        gesture.name, distance);

Directories:

Name	Description
cemg	C code for the classifier devices.
ctest	Tests for the C classifier.
data	Pre-recorded data to test and experiment with. Includes functions to access
matlab	Scripts used to develop and test the classifier.