BOOST Mentor Position at CSULA – Automated Muscle Threshold Calibration Algorithm Development

BOOST Mentor
May 2019 – August 2019

I was a mentor for CSUPerb 2019 scholar awardee Abigail Garcia for the undergraduate summer bridge program called BOOST at Cal State LA. During the 10 week session, I assisted her on her task to develop an automated method of setting threshold levels for the EMG activated DREAM app. I assisted her with data collection, provided assistance with mathematical concepts, and validated the results produced.

Problem Description

The application utilizes a dual threshold peak detection algorithm that functions as a Schmitt Trigger. This provides robustness against noisy signals that would otherwise characterize false peak detections. Currently, these thresholds are set manually using visual inspection.

This project involved developing an algorithm that would set these thresholds automatically using anthropometric information such as maximum voluntary contraction, age, mass, and height.

Once developed and thoroughly tested, the algorithm would be coded into the app using Java and Android Studio.

At the end of the term, Abigail was successfully able to develop an algorithm that was capable of setting upper and lower EMG thresholds for punch and muscle contraction detections using a model derived from linear regression. The algorithm incorporated anthropometric data such as height, age, mass in combination with maximum heart rate and maximum voluntary contraction (MVC). Test data was obtained by recording EMG and heart rate data from subjects performing three different exercises: punches, bicep curls, and military presses. An individual model was produced for each exercise.

The concept and test results were submitted in an abstract for the 2019 BMES Annual Conference. She successfully presented her work in Philadelphia.

Additional Information