PROJECT OUTCOMES
The original goal of this project was to demonstrate that EMG signals can be used as real-time inputs for VR systems in the context of biotherapy by developing a complete visualization tool. While we were unable to work with live EMG data streams nor were we able to develop a fully-functioning GUI due to time and resource constraints, we believe that the outcomes of our project were significant first steps in realizing our original goal, and that they effectively demonstrated the use of EMG signals to drive VR systems. The following is a list of the project outcomes, all of which are scalable to eventually accept live data streams:
A classifier that can distinguish between “normal” and “myopathy” EMG data to monitor the conditions of patients during therapy sessions, as well as to track patient progress with the various features extracted
An automated feature extraction tool that can take in a variety of file formats and retrieves the desired features from a given EMG signal
A visual rendering of the human arm that can produce motions given the following parameters and inputs:
Volumetric measurements of the patient’s arm such as the length, radius, etc.
Physiological measurements of the patient’s arm such as density, stiffness, and damping coefficient
Input torque signal converted from the original EMG voltage signal via non-linear transforms
POTENTIAL NEXT STEPS
Integrate the classification result and visual rendering of the EMG signal onto a single user-friendly GUI so that the patient can see the rendering while monitoring their status
Customize the arm CAD and arm parameters for a given patient to reflect accurate renderings based on the patient’s physiology
Scale up this visualization tool to accept live data streams, because currently our scripts are only processing pre-recorded data.
Scale up the visual rendering to include other rotations of the elbow joint as well as other joints of the human body
Scale up the classifier to detect muscle fatigue or other types of muscle disorders